Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page i • National Registry of Exonerations • September 1, 2020
Government Misconduct and
Convicting the Innocent
Samuel R. Gross, Senior Editor, srgross@umich.edu
Maurice J. Possley, Senior Researcher
Kaitlin Jackson Roll, Research Scholar (2014-2016)
Klara Huber Stephens, Denise Foderaro Research Scholar (2016-2020)
NATIONAL REGISTRY OF EXONERATIONS
SEPTEMBER 1, 2020
The Role of Prosecutors, Police and Other Law Enforcement
National Registry of Exonerations
Newkirk Center for Science & Society • University of California Irvine • Irvine, California 92697
University of Michigan Law School • Michigan State University College of Law
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page ii • National Registry of Exonerations • September 1, 2020
For Denise Foderaro and Frank Quattrone
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page iii • National Registry of Exonerations • September 1, 2020
Preface
This is a report about the role of official misconduct in the conviction of innocent people. We
discuss cases that are listed in the National Registry of Exonerations, an ongoing online archive
that includes all known exonerations in the United States since 1989, 2,663 as of this writing.
This Report describes official misconduct in the first 2,400 exonerations in the Registry, those
posted by February 27, 2019.
In general, we classify a case as an “exoneration” if a person who was convicted of a crime is
officially and completely cleared based on new evidence of innocence. A more detailed definition
appears here.
The Report is limited to misconduct by government officials that contributed to the false
convictions of defendants who were later exonerated—misconduct that distorts the evidence
used to determine guilt or innocence. Concretely, that means misconduct that produces
unreliable, misleading or false evidence of guilt, or that conceals, distorts or undercuts true
evidence of innocence.
Three years ago, the Registry released a report on Race and Wrongful Convictions in the United
States. We found, among other patterns, that Black people who were convicted of murder were
about 50% more likely to be innocent than other convicted murderers, and that innocent Black
people were about 12 times more likely to be convicted of drug crimes than innocent white
people. Some of those disparities are caused by the type of misconduct we study here and some
are not.
Misconduct in obtaining and presenting evidence contributes substantially to the racial disparity
in murder exonerations, as we will see. On the other hand, the huge disparity in drug
exonerations primarily reflects a type of misconduct we don’t cover in this Report—racial
discrimination in choosing which people to stop or search for drugs, what is commonly called
“racial profiling.”
The Report describes many varieties of misconduct in investigations and prosecutions. Some are
always deliberate, some are rarely or never deliberate, and some may or may not be deliberate.
The Report organizes the myriad of types of misconduct into five general categories, roughly in
the chronological order of a criminal case, from initial investigation to conviction: Witness
Tampering; Misconduct in Interrogations of Suspects; Fabricating Evidence; Concealing
Exculpatory Evidence; Misconduct at Trial.
Most of the misconduct we discuss was committed by police officers and by prosecutors. We also
report misconduct by forensic analysts in a minority of cases, mostly rapes and sexual assaults,
and by child welfare workers in about a quarter of child sex abuse cases.
Some major patterns we observed:
• Official misconduct contributed to the false convictions of 54% of defendants who were
later exonerated. In general, the rate of misconduct is higher in more severe crimes.
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page iv • National Registry of Exonerations • September 1, 2020
• Concealing exculpatory evidence—the most common type of misconduct—occurred in
44% of exonerations.
• Black exonerees were slightly more likely than whites to have been victims of misconduct
(57% to 52%), but this gap is much larger among exonerations for murder (78% to
64%)—especially those with death sentences (87% to 68%)—and for drug crimes (47% to
22%).
• Police officers committed misconduct in 35% of cases. They were responsible for most of
the witness tampering, misconduct in interrogation, and fabricating evidence—and a
great deal of concealing exculpatory evidence and perjury at trial.
• Prosecutors committed misconduct in 30% of the cases. Prosecutors were responsible for
most of the concealing of exculpatory evidence and misconduct at trial, and a substantial
amount of witness tampering.
• In state court cases, prosecutors and police committed misconduct at about the same
rates, but in federal exonerations, prosecutors committed misconduct more than twice as
often as police. In federal exonerations for white-collar crimes, prosecutors committed
misconduct seven times as often as police.
We also examined disciplinary actions against officials who committed misconduct. These were
uncommon for all types of officials, and especially so for prosecutors.
We tried to determine whether official misconduct that contributes to false convictions has
become more or less frequent over the past 15 to 20 years. For most types of misconduct, we
won’t know for years to come, but we already see strong evidence that a few kinds of misconduct
have become less common: violence and other misconduct in interrogations; abusive
questioning of children in child sex abuse cases; and fraud in presenting forensic evidence. On
the other hand, the number of federal white-collar exonerations with misconduct by prosecutors
has been increasing.
In the last section we consider what led officials to commit misconduct. We conclude that the
main causes are pervasive practices that permit or reward bad behavior, lack of resources to
conduct high quality investigations and prosecutions, and ineffective leadership by those in
command. We discuss a range of possible remedies, from specific rules to changes in culture, in
cities, counties, states and the nation as a whole.
We present many other findings in the Report itself. The core of our data on official misconduct
are available online, sortable and filterable, for others to explore; go to the “OM Tags” column
here.
Samuel R. Gross
Maurice J. Possley
Kaitlin Jackson Roll
Klara Huber Stephens
September 1, 2020
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page v • National Registry of Exonerations • September 1, 2020
Use Note:
1. Common terms
It may be useful to explain some terms that we use in this Report:
Exoneration means an exoneration listed in the Registry. Every exoneration,
identified by the name of the exoneree, has a page in the Registry, and is listed on our
Summary View and Detailed View pages.
Known exonerations: We know that our list of exonerations is incomplete: we
regularly discover cases we missed. Sometimes we specify that these are “known
exonerations,” more often we don’t, but it’s true regardless.
Misconduct in an exoneration: Strictly speaking, the practice we write about is
official misconduct that contributed to a criminal conviction that was ultimately
reversed by exoneration. That’s a mouthful. For convenience, we often refer to it as
“misconduct in the exoneration” even though the misconduct was part of the process of
obtaining a conviction.
Police: Police agencies in the United States range from one-person police departments
to the FBI. The titles of sworn peace officers include Patrolman, Officer, Deputy Sheriff,
Trooper, Agent—and many more. We refer to all of them as “police.”
2. Links and Navigation
(i) The report contains numerous links to pages on the website of the National Registry
of Exonerations. Most are links to the stories of individual exonerees; some are links to
collections of cases. In both situations, almost all links go to the current versions of the
pages, not those in effect in late February 2019, when we completed the set of 2,400
exonerations that are the subject of this report. For example:
• This link goes to Ricky Jackson’s page, which was last updated in May 2020. That
page contains information we did not know when we completed the compilation
of the dataset fifteen months earlier—and (like other summaries and data on the
Registry) it may be further modified in the future.
• This link goes to a list of all exonerations with misconduct in Cook County at the
time you click on it—230 as of this writing, more in months and years to come—
not the 204 exonerations with official misconduct in Cook County among the
2,400 exonerations included in this Report.
(For technical reasons, a few links go to copies of Registry pages rather than live pages.)
(ii) The Executive Summary and the Table of Contents contain links that may help
navigate this document. The Summary contains a list of page numbers in the form of
links—like this, 9—that take you to the indicated page in the text. In the Table of
Contents you can click on any part of an entry to go to the page on which that section
begins.
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page vi • National Registry of Exonerations • September 1, 2020
(iii) Each page of the text (except the first pages of major sections) includes two
highlighted buttons:
Go to Executive Summary and Go to Table of Contents.
If you click on them, they will take you to the beginning of the Executive Summary and of
the Table of Contents, respectively.
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page vii • National Registry of Exonerations • September 1, 2020
Acknowledgements
We didn’t do this on our own. Not nearly. It took a couple of villages and a lot of friends.
This report was produced by the National Registry of Exonerations. The editors of the Registry
were essential: Barbara O’Brien, Editor in Chief; Simon Cole, Associate Editor and Director; and
Catherine Grosso, Managing Editor. They read drafts, classified cases and thought through the
project with us. The Registry staff—Ken Otterbourg, Jessica Weinstock Paredes, Meghan
Cousino, and Eva Nagao who left us this June—were equally essential. They identify the cases on
which our work is based; research, code and write them up; and maintain the website through
which the work of the Registry is available to the world. We also received invaluable support and
advice from our Advisory Board, especially Denise Foderaro, Barry Scheck, and Rob Warden,
co-founder of the Registry.
The core work of our work—researching, coding, checking and recoding information on official
misconduct in the 2,400 cases in our database—was mostly done by a dedicated group of
research assistants—some of whom also did legal research, wrote memoranda, commented on
and corrected partial drafts, and provided advice at many stages. Most were students at the
University of Michigan Law School—Christine Adams, Zachary Adorno, Claudia Arno, Jennifer
Chun, Michael Darling, Lauren Flamang, Max Greenwald, Griffin Hardy, Caroline Howe,
Connor Lang, Ginny Lee, James Millikan, Amanda Rauh-Bieri, Amanda Stephens, Jenny Stone,
Julia Xin and Eric Yff—or at the Michigan State University College of Law: Nadine Kassem and
Alison Swain. In addition, we received excellent contributions from two young lawyers, Marc
Allen and Eli Wykell, and careful statistical analyses from Josue Guevara, starting when he was
a student at Michigan, German Marquez Alcala, who works for the University of Michigan Law
Library, and Valerie King, as graduate student at the Univesity of California, Irvine, and after
she completed her degree.
The staff at the University of Michigan Law School was as skillful and helpful as always. In
particular, Cheri Fidh corrected more errors in content and format than we can count, while
Alex Lee and Richard Savitski are responsible, respectively, for the appearance and the contents
of the data we are making available online with this report. At a distance, Julie Smith designed
the appearance of the report, and Margot Friedman worked tirelessly to present it to the world.
The staff of the Innocence Project was unfailingly helpful, including especially Barry Scheck and
Rebecca Brown, who answered questions, provided information, read drafts, and suggested
additions. Elizabeth Webster, formerly of the Innocence Project, spent a summer helping us
devise our initial coding system. And our dear friends in Ann Arbor, Phoebe Ellsworth and
Alexandra Gross, read partial and full drafts of this report repeatedly over several years, made
countless corrections and suggestions, and sustained our spirits.
The Registry, and this project in particular, depend on generous financial support from many
individuals and organizations. We are particularly grateful to James and Martha Newkirk, and
to Denise Foderaro and Frank Quattrone, who encouraged and supported our work since its
inception, in many ways.
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page viii • National Registry of Exonerations • September 1, 2020
The Registry is a joint project of three universities. We are fortunate to have had the support of
the University of Michigan Law School, our original home for several years; the Michigan State
University College of Law, which took us on four years ago; and the Newkirk Center for Science
& Society at the School of Social Ecology of the University of California, Irvine, which has been
our main home since 2016.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page ix • National Registry of Exonerations • September 1, 2020
Executive Summary
•
Page
I. Introduction
1
II. Background
3
Misconduct by law enforcement has received a great deal of attention as a result
of the Black Lives Matter movement, which has focused on racial discrimination
and violence by police officers. We study a different (but overlapping) type of
behavior: misconduct that distorts evidence in criminal cases and leads to
convictions of innocent people.
3
There is a dearth of prior systematic research on police misconduct that
contributes to false convictions.
3
Prosecutorial misconduct has attracted a good deal of attention in the past decade,
primarily concealing exculpatory evidence. Several studies have found thousands
of criminal cases in which courts or other agencies determined that prosecutors
committed misconduct, but very few were disciplined for it.
3
Our database, the National Registry of Exonerations, is an ever-changing
public archive. We define “exoneration” by the conduct of public officials, and use only
non-confidential data. We list all exonerations we can find, add new cases regularly, and
modify our postings on old cases as we get more information or refine or inquiries.
7
This unique database enables us to examine all exonerations—with data from multiple
sources—to identify many cases of misconduct that cannot be found in official
decisions, and to begin to describe the causes and effects of that misconduct.
8
We cannot, however, estimate rates of misconduct in all criminal cases; and
even among exonerations we miss a great deal of official misconduct that remains
hidden.
8
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page x • National Registry of Exonerations • September 1, 2020
•
Page
We do not systematically examine misconduct by criminal defense attorneys
or by judges. That’s unfortunate, especially for ineffective legal assistance by defense
lawyers, which is probably a major contributor to convictions of innocent defendants.
The main reason is that we do not have data that would enable us to speak to those
issues.
9
III. The Frequency of Official Misconduct
11
In 54% of exonerations, official misconduct contributed to the false convictions;
usually more than one type of misconduct. Overall, male exonerees and Black
exonerees were modestly more likely to experience misconduct (with some larger
differences by race for a few particular crimes).
11
30% of exonerations include misconduct by prosecutors, 35% misconduct by
police, 3% by forensic analysts, and 2% by child welfare workers.
12
The overall rate of misconduct varies by crime, from 72% in murder cases to 32% for
most non-violent crimes. For most crimes, the rates of misconduct for prosecutors
and police are comparable. However:
12
• For drug crimes, the rate of police misconduct is nearly four times the rate of
misconduct for prosecutors.
13
• In white-collar exonerations, prosecutorial misconduct is more than five
times as frequent as misconduct by police. This gap is entirely due to the
extremely high rate of misconduct by federal white-collar prosecutors.
13
We only count misconduct that contributed to the exonerees’ false
convictions by generating false evidence of guilt or concealing true evidence of
innocence. We don’t count misconduct that can’t produce false evidence—for example,
police brutality that was not part of an interrogation—or failed attempts to produce false
evidence, such as torturing a suspect who does not confess.
13
Violent felonies account for nearly 80% of exonerations. Misconduct is generally
more common the more extreme the violence, ranging from 38% and 39% for robbery
and sexual assault cases to 72% for exonerations from death sentences. These
numbers reflect both higher rates of official misconduct in the most serious
crimes, and more diligent post-conviction reinvestigations.
15
Drug crimes make up more than 60% of exonerations for non-violent crimes. Two-
thirds of them occurred in two very different local clusters:
20
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xi • National Registry of Exonerations • September 1, 2020
•
Page
• In Chicago, 66 convicted drug offenders were exonerated after it was shown
that officers under the command of a corrupt police sergeant or his subordinates
planted evidence on them. All of those cases involved police misconduct.
21
• In Harris County, Texas (Houston), 149 defendants who pled guilty to drug
crimes were exonerated after lab tests found no illegal drugs in the materials
seized from them. Only 3% of those cases involved misconduct.
24
White-collar crimes—the second largest group of exonerations for non-violent
offenses—are primarily federal cases with a very high rate of prosecutorial
misconduct.
26
About 80% of criminal convictions in the United States are misdemeanors, but only
about 4% of exonerations, and two-thirds of those are Harris County drug crime
guilty plea cases. The remaining sliver of misdemeanor exonerations—about 1% of the
total—have a high rate of official misconduct, 58%.
26
Overall, exonerations of Black defendants have a slightly higher rate of misconduct
than those of white defendants, 57% to 52%. But the differences are greater for
murder cases (78% to 64%)—especially those with death sentences (87% to 68%)—
and drug crime exonerations (47% to 22%).
28
Almost all the official misconduct we have identified falls into five general categories
that we discuss in detail in the sections that follow:
29
1. Witness tampering occurred in about 17% of exonerations.
30
2. Misconduct in interrogations occurred in 57% of all exonerations with
false confessions, or about 7% of all cases.
31
3. Fabricating evidence happened in about 10% of cases, in three forms:
Forensic fraud—in 3% of exonerations, police officers or forensic analysts
lied about forensic evidence. Fake crimes—in 4% of exonerations, police
planted drugs or guns on innocent suspects, or lied and said the suspects had
assaulted them. Fictitious confessions—in about 2% of exonerations,
officers fabricated confessions from defendants who did not confess.
31
4. Concealing exculpatory evidence is the most common type of official
misconduct we found. It occurred in 44% of all exonerations.
32
5. Misconduct at trial occurred in about 23% of exonerations, about evenly
divided between perjury by law enforcement officers, 13%, and trial
misconduct by prosecutors, 14% (with some overlap).
33
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xii • National Registry of Exonerations • September 1, 2020
•
Page
Misconduct in interrogations occurred overwhelmingly in murder
exonerations; concealing exculpatory evidence and misconduct at trial were
most common in murder cases, followed by white-collar crimes; witness
tampering was slightly more common among exonerations for child sex abuse
exonerations than for murder; and fabricating evidence was several times
more common among exonerations for drug crimes than for any other crime.
30
• IV. Witness Tampering
34
Witness tampering occurs when a law enforcement officer tricks, persuades or
forces a witness to testify falsely against the defendant. The officer need not know
that the witnesses is testifying falsely as long as the officer does not care whether the
witness is telling the truth.
34
Witness tampering occurred in 17% of exonerations. In 5%, witnesses were forced to
give false testimony by threats, in 13% they were manipulated into doing so without
threats. (1% of cases included both types of tampering.)
35
Witness tampering occurred most often in child sex abuse cases, 28%, and murder
cases, 23%. Police participated in witness tampering in 80% of cases where it
occurred, prosecutors in 31% and child welfare workers in 14%.
35
About four-fifths of witness tampering falls into one of three categories:
1. Procuring false testimony—inducing a witness to testify to facts the
officer or prosecutor knows the witness did not perceive.
2. Tainted identifications—inducing a witness to identify a suspect at an
identification procedure, whether or not the witness recognizes the suspect.
3. Improper questioning of a child victim—repeated, insistent and
suggestive questioning of a child by officials who will not allow the child to
deny that s/he was a victim of sex abuse.
36
Procuring false testimony, with or without threats, means obtaining testimony that
both law enforcement and the witness know is false. It occurred in 6% of
exonerations, two thirds of them murder cases. Police were involved about twice as
often as prosecutors.
37
A suggestive identification procedure—for example, giving a witness a single picture
of a suspect to identify—may easily cause misidentifications, but that alone is not
misconduct. A tainted identification occurs when an officer directly or indirectly
“tells” a witness who to identify as the criminal.
39
Tainted identifications occurred in about 6% of exonerations. Three quarters were
murder and sexual assault cases; almost all were obtained by police.
39
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xiii • National Registry of Exonerations • September 1, 2020
•
Page
• In 80% of murder cases with tainted identifications, at least one witness
deliberately misidentified the exoneree; many were forced to do so.
40
• In all but one sexual assault exonerations, victims or other witnesses were
persuaded or tricked into misidentifying the exonerees by mistake; in one
case, the identification was produced by threats.
41
Improper questioning of a child victim occurred in about a quarter of child sex
abuse exonerations, primarily cases from the epidemic of child sex abuse hysteria
prosecutions in the early 1980s to the late 1990s.
41
• Police participated in improper questioning of child victims 85% of the time,
and child welfare workers did so in 71% of the cases. Most of the children
were questioned by more than one type of official.
36
• Some children who eventually testified against exonerees came to believe
their accusations; others have said that they knew that they were lying.
43
• V. Misconduct in Interrogations
45
In 12% of known exonerations—mostly murder cases—convictions were based on
false confessions by the exonerees. 57% of false confessions were obtained by
misconduct in interrogations.
45
Misconduct in interrogations is defined (if not clearly) by the Supreme Court.
Beginning in the 1940s, the Court developed an increasingly strong prohibition against
violence in interrogations. Otherwise, an interrogation violates due process of law if
under the “totality of the circumstances” it is deemed so coercive that the resulting
confession is “involuntary.”
48
False confessions are far more common in Chicago than elsewhere. In the rest
of the country, 10% of all exonerations and 18% of murder exonerations included false
confessions; in Chicago the comparable figures are 33% and 54%.
47
The same is true of misconduct in interrogations: 77% of false confessions in Chicago
were obtained by misconduct, compared to 49% elsewhere.
49
Actual or threatened violence was used in 64% of interrogations with
misconduct—36% of all exonerations with false confessions—as often as all
other forms of misconduct in interrogations combined.
50
The concentration of violence in interrogations in Chicago is particularly stark.
Violence was used to obtain 69% of false confessions in exonerations in Chicago,
but only 24% elsewhere. Half of all exonerations in the country with false
confessions that were obtained by violence are from Chicago.
50
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xiv • National Registry of Exonerations • September 1, 2020
•
Page
Much of the violence in interrogations in Chicago was due to a systematic program of
torture of Black suspects in the 1970s and 80s by Chicago Police Commander Jon
Burge and his subordinates (see Section XII).
50
Some lies, promises and threats in interrogations are permitted, and some are
prohibited. The distinction can be elusive. Other than violence, most misconduct in
interrogations consisted of prohibited lies, promises or threats.
48
53
Interrogators are allowed to lie about the facts of the investigation (“we found your
fingerprints”) but not about the law (“you’ll get sentenced to death”). They may make
vague promises (“if you confess, we can help you”), but not specific ones they can’t
keep (“if you confess, the DA won’t ask for the death penalty”). In 20% of exonerations
with false confessions, the police lied about the law or promised outcomes they
couldn’t deliver.
48
53
Police may threaten to arrest a suspect who does not confess—an act within their
power—and prosecutors may threaten to prosecute one who does not cooperate.
But threats against third parties—e.g., to arrest a spouse or child of the suspect, or to
remove minor children from her home—are prohibited. Third-party threats were
used in 8% of exonerations with false confessions.
46
48
54
Several permitted interrogation practices also contribute to false confessions:
55
• Permissible promises and threats.
• Lying about the investigation.
• Telling the suspect details of the crime (which makes it hard to separate
true confessions from false ones generated by the police).
• Interrogating a juvenile without a parent or guardian present.
One or more of these practices contributed to 70% of false confessions: 79% of those
obtained with misconduct and 57% of those obtained without it.
60
Confessions by actual or possible codefendants of the exonerees falsely
implicated exonerees in about 13% of cases.
60
Many codefendants voluntarily confessed and implicated exonerees, usually to shift
some blame away from themselves. About a third of codefendant confessions that
contributed to false convictions were obtained by the same types of misconduct that
produced most false confessions by the exonerees themselves.
61
In a third of exonerations with codefendant confessions, the exoneree also
confessed; in the rest, the exoneree did not. The net effect is that all false
confessions—by codefendants as well as by exonerees themselves—contributed to
the convictions in 21% of exonerations.
60
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xv • National Registry of Exonerations • September 1, 2020
•
Page
As with false confessions by exonerees, false codefendant confessions that helped
convict exonerees were concentrated in murder cases, and in Chicago.
61
• VI. Fabricated Official Evidence
65
In 10% of exonerations, officers falsely reported that they examined forensic
evidence that proved (or failed to disprove) the defendants’ guilt, saw the defendants
commit crimes that did not occur, or witnessed confessions by defendants who
did not confess.
65
Forensic fraud—the deliberate falsification of forensic evidence to help convict a
defendant—occurred in 3% of exonerations.
65
• Forensic fraud is a form of intentional misconduct. We do not count a
larger set of cases with forensic evidence that was (as far as we know)
unintentionally mistaken, misleading, or invalid.
65
• We only count forensic fraud by law enforcement officers, usually forensic
examiners at police crime labs or other state-run labs.
65
• There are many types of forensic fraud, but these are the most common:
1. In more than a third of the cases, analysts reported that the defendant’s
hair, saliva, blood, semen, tooth marks, etc., matched or were
consistent with those found at the crime scene, when in fact testing
had shown the opposite.
65
2. In about a quarter of the cases, forensic witnesses reported that the
defendants might have been the source of crime-scene blood, semen
or fingerprints, after forensic tests that showed that was impossible.
65
• A third of forensic fraud cases involve repeat offenders, possibly because
they are more likely than other wrongdoers to eventually get caught, after which
many of their prior cases are reexamined.
67
Fake crimes were fabricated by police in about 5% of exonerations:
68
• In about 4%, police planted evidence at the scene of the crime and claimed to
have found it there. In all but a few cases, they planted illegal drugs—especially
in a cluster in Chicago (see above, Section III.3.c.i).
68
• In about 1%, officers falsely claimed that the defendants assaulted them,
usually to cover up their own violence against the same defendants.
69
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xvi • National Registry of Exonerations • September 1, 2020
•
Page
Fabricated confessions: In about 2% of exonerations, police made up confessions
from exonerees who did not confess.
70
• In several cases, police had exonerees sign documents they did not or could
not read, which later turned out to be confessions. In most cases, they lied and
said the exonerees made unrecorded oral confessions.
70
• As usual for false confessions, fabricated confessions were more likely in
Chicago than elsewhere, but by a modest amount, 16% to 11%.
72
• VII. Concealing Exculpatory Evidence
74
Concealing exculpatory evidence contributed to the convictions of 44% of
exonerees, more than any other type of official misconduct we know of.
75
The legal duty to disclose exculpatory evidence has multiple bases:
75
• In Brady v. Maryland, in 1963, the Supreme Court announced the
‘Brady rule’: “[S]uppression by the prosecution of evidence favorable to the
accused… violate[s] due process where the evidence is material either to guilt or
to punishment.”
75
• Brady only applies if the concealed evidence is “material”—which, in this
context, means that the outcome of the trial would likely have been different if
that evidence had been known. This requirement has been widely criticized as
incoherent, inconsistent and unadministrable.
75
• In addition, rules of professional responsibility and pretrial discovery
that govern criminal cases also require the prosecution to disclose all
exculpatory evidence, regardless of “materiality.”
78
• We apply these procedural and ethical rules—and classify hiding evidence
as misconduct regardless of “materiality”—because they prescribe correct
conduct rather than define a violation of the constitution. (Plus, we too could not
classify “materiality” consistently if we tried.)
80
The rate of concealing exculpatory evidence varies by crime, from 61% for murder
to 27% in child sex abuse cases. It is so common and widespread that it happened
in 82% of all exonerations with any official misconduct.
81
Prosecutors concealed exculpatory evidence in 73% of cases in which it occurred.
That’s not surprising, since prosecutors have the duty to disclose that evidence to
the defense. We only count other officials as responsible if (as far as we know)
prosecutors were ignorant of the evidence.
82
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xvii • National Registry of Exonerations • September 1, 2020
•
Page
Police concealed exculpatory evidence in 33% of cases where it occurred (including
cases with concealing by more than one type of official), and forensic analysts did so
in 6%. In some portion of those exonerations, prosecutors did know about the concealed
evidence, but we have no record of that knowledge.
82
As far as we know, only 13% included concealed physical objects—clothing,
weapons, etc. This gap may in part reflect how effectively objects can be destroyed or
hidden, but information may linger in electronic or physical files, or the memories
of people.
83
In 63% of cases with concealed exculpatory evidence, substantive evidence of the
exonerees’ innocence was hidden—evidence that in itself helps prove the defendant’s
innocence, such as an eyewitness who named another person as the criminal.
85
In 80% of such cases, impeachment evidence that undermined testimony by
prosecution witnesses was concealed—for example, evidence that a witness who
identified the exoneree as a murderer told his brother he never saw the killing.
85
In half the exonerations with concealed exculpatory evidence, both substantive and
impeachment evidence were hidden. Often, a single item of evidence serves both
functions. “Substantive” evidence may sound more important, but concealing
impeachment evidence that eviscerates the credibility of a critical prosecution witness
can be devastating to an innocent defendant.
85
Predictably, law enforcement officials usually conceal their own misconduct. That’s
misconduct in itself, derivative concealment. For example, it’s misconduct for an
officer to plant drugs on a suspect, and it’s a separate act of misconduct to conceal
the officer’s knowledge that the suspect is innocent.
85
Evidence of other official misconduct was concealed in 26% of all exonerations,
over half of exonerations with any concealed exculpatory evidence.
88
A large variety of types of exculpatory evidence were concealed, but most fall into
several categories.
89
Impeachment evidence:
• Incentives to testify against the exonerees were concealed in 21% of
exonerations, most often deals on criminal charges against the witnesses.
89
• Inconsistent statements by prosecution witnesses that contradicted their
testimony were concealed in 14% of exonerations.
90
• Criminal records and histories of dishonesty of witnesses for the state
were concealed in 4% of exonerations.
90
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xviii • National Registry of Exonerations • September 1, 2020
•
Page
Substantive Evidence:
91
• Exculpatory forensic tests were concealed in 6% of exonerations, including
many that conclusively established the exonerees’ innocence.
91
• Alternative suspects were concealed in 12% of all exonerations—20% of
murder exonerations and 6% of other cases.
92
• Exclusions by an eyewitness—evidence that an eyewitness said the exoneree
is not the criminal—was concealed in 2% of exonerations.
92
• Alibi evidence was concealed in 1% of exonerations.
94
• Evidence that no crime was committed was concealed in 6% of exonerations,
mostly cases where police concealed the fact that they themselves framed the
defendants.
94
• VIII. Misconduct at Trial
96
At least 95% of criminal convictions in the United States are obtained by guilty
pleas rather than trial verdicts, but 80% of exonerations in the Registry followed
conviction at trial. About 28% of those trials (23% of all exonerations) included
official misconduct in court.
96
Police perjury
96
• Perjury by all law enforcement officials occurred in 14% of the trials at
which exonerees were convicted, or 13% of all exonerations (including those
after guilty pleas). In about a quarter of those cases, officials lied about forensic
testing, or about things the officials themselves claimed to have witnessed the
exonerees do or say. (See above, Section VI.)
96
• Perjury by police officers occurred in 11% of trials of exonerees. In 9% of
those trials (7% of all exonerations), officers lied about information
obtained from others.
97
• Most often, police lied about the conduct of the investigations: what a
witness had said, whether or how a lineup was conducted, etc. The most
common subject of police perjury was the conduct of interrogations at which
innocent defendants confessed.
97
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xix • National Registry of Exonerations • September 1, 2020
•
Page
• We miss a great deal of police perjury. We rarely have access to
transcripts or other detailed information about trial testimony, so we only
learn about perjury at trial if it becomes a conspicuous issue.
98
Trial Misconduct by Prosecutors
98
• Permitting Perjury
98
❖ In 1959, the Supreme Court held that a prosecutor has a
constitutional obligation to correct perjury by a state witness even
if she did not herself offer the false testimony.
99
❖ Prosecutors permitted perjury to go uncorrected in 8% of
exonerations. In most cases, the perjury was by civilian witnesses. The
most common lies were about favorable treatment the witnesses
receive in pending criminal cases of their own.
99
• Lying in Court
100
❖ It is misconduct, and punishable as contempt of court, for a lawyer to
lie in court, whether or not the lawyer in under oath.
101
❖ We know that prosecutors lied in court in 4% of exonerations. The
real rate may be higher since we only count cases with clear evidence
that prosecutors made statements they knew were false.
102
❖ About half of lies by prosecutors were made in closing argument. A
common pattern is to repeat and affirm perjury by a witness that the
prosecutor knew about but failed to correct—for example, a lie by a
witness who claimed to have no deal with the prosecutor.
102
• Improper Statements in Closing Argument or Cross-examination
103
❖ Prosecutors also commit misconduct at trial without lying, usually
in closing arguments or in questions on cross-examination.
103
❖ Prosecutors made improper closing arguments (without lying) in
3% of exonerations, ranging from statements that the prosecutor “knows”
the exoneree is guilty to outright appeals to bigotry.
104
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xx • National Registry of Exonerations • September 1, 2020
•
Page
❖ Prosecutors asked impermissible questions on cross-examination in
1% of exonerations.
106
❖ Both forms of misconduct are undoubtedly much more common
than we know. They are only visible if the defense objects at the time,
and lawyers often fail to do so, intentionally or by neglect.
104
• IX. Federal Cases
108
Federal crimes are a small and unrepresentative minority of all criminal cases in the
United States. They generate about 6% of convictions, heavily skewed to immigration,
drug and white-collar crimes.
108
Federal exonerations are a comparably small minority of all exonerations, and
equally skewed: 41% are white-collar crimes, and another half are about evenly split
between drug and violent crimes.
108
The overall rate of official misconduct is somewhat higher in federal exonerations
than in state cases, 61% compared to 54%.
109
Most misconduct in federal exonerations was committed by prosecutors, 52%
compared to 29% in state cases.
109
Federal prosecutors committed misconduct in exonerations more than twice as
often as police (52% to 20%), while state prosecutors committed misconduct less
often than police (29% to 36%).
109
Federal white-collar exonerations have striking similarities to murder
exonerations under state law. They are the most common types of exonerations in
their respective courts; many are big-ticket cases—expensive, long-running,
conspicuous; they have the highest rates of misconduct for exonerations in those
courts, 65% for federal white-collar crimes, 72% for state-court murders.
110
In federal white-collar exonerations, prosecutors committed misconduct more
than 7 times as often as police, 65% to 9%; every federal white-collar exoneration
with any official misconduct included misconduct by a prosecutor.
112
Federal white-collar cases have both the highest rate of misconduct by
prosecutors and the lowest rate of misconduct by police of exonerations in any
crime category.
112
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxi • National Registry of Exonerations • September 1, 2020
•
Page
Federal white-collar prosecutors seem to play a more dominant role in cases
that lead to exonerations than state prosecutors. They have more resources, and are
more likely to take the lead in the investigations. That role may reduce police
misconduct—but not misconduct by the prosecutors themselves.
113
• X. Discipline
115
In 17% of exonerations with official misconduct, we know that some form of
discipline was imposed on officials who participated in that misconduct.
115
Many officials who were disciplined committed misconduct in several or many
exonerations, but formal discipline was limited to one, or was imposed in a separate
case. Chicago Police Commander Jon Burge, for example, was sentenced to prison for
lying about the torture program he ran. We count that as discipline in all 19 cases in
which he or those he commanded abused the exonerees. In 70% of exonerations with
discipline, it was imposed for general patterns of behavior or in cases other than the
specific ones at hand.
115
Discipline may be imposed by three sets of authorities: employment discipline by
the agencies that employ the misbehaving officials; professional discipline by
regulatory bodies that certify or license their professions (this category includes a few
instances of courtroom discipline of prosecutors by judges); and criminal discipline,
convictions by courts for misconduct that violates criminal laws.
117
Disciplining Prosecutors
119
• Prosecutors are hardly ever disciplined for misconduct that contributes to
false convictions. We know of some discipline for prosecutors in 4% of
exonerations with prosecutorial misconduct. In most of those cases, the
discipline was comparatively mild.
120
• Eleven prosecutors were disciplined by the offices that employed them, but
just two were fired (and four resigned or retired); 14 were disciplined by bar
authorities or courts, but only three were disbarred. Only two prosecutors
have been convicted of crimes for misconduct in exonerations, both in
notorious cases, and both received nominal sentences.
120
Disciplining Police
121
• Police officers were disciplined in 19% of exonerations with police
misconduct, about five times the rate for prosecutors.
119
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxii • National Registry of Exonerations • September 1, 2020
•
Page
• In almost 80% of these cases, officers were convicted of crimes; in 20% they
were disciplined by the police forces for which they worked. We know of no
professional discipline of police officers.
120
• In 127 exonerations, police officers who committed misconduct were
convicted of crimes for misconduct of the sort they committed in those cases—
but not 127 separate officers. As we explained, the conviction of a single serial
offender may count as discipline in many cases.
121
• Even so, at least 30 officers were convicted of crimes (compared to two
prosecutors) and some received long prison sentences.
122
• Disciplinary records of police officers are often concealed by their
employers, unions, and professional agencies. As a result, we have no doubt
missed cases of employment and professional discipline of police.
122
Disciplining Forensic Analysts
124
• Forensic analysts were disciplined in 47% of exonerations in which their
misconduct was discovered, a much higher rate than prosecutors or police.
Four-fifths of them were disciplined by their employers, a fifth by
professional agencies, and a few were convicted of crimes.
124
• As with police, several analysts who were disciplined were repeat offenders,
each of whom participated in multiple cases. We know of 35 exonerations with
discipline for forensic misconduct, but only 13 separate analysts were punished,
six of whom accounted for 80% of the cases.
124
• XI. Changes in Official Misconduct over Time
127
It is difficult to detect decreases in misconduct in exonerations because of the long
time lag from conviction (the last date for occurrence of misconduct) to
exoneration (the earliest date when the misconduct can be added to our data).
127
For example, there were six times more exonerations with misconduct in murder
convictions in the 16 years from 1987 through 2002 than in the 16 years from 2003
through 2018—but because the average time to exoneration for such cases is 17
years, more cases from the later period will continue to emerge for years.
128
In three contexts, official misconduct decreased so sharply that we believe the
decline is real despite the difficulty of identifying decreases:
128
• Improper questioning of child victims all but stopped after the end of the
child sex abuse hysteria epidemic that ran its course in the late 1990s.
129
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxiii • National Registry of Exonerations • September 1, 2020
•
Page
• Misconduct in interrogations, especially violence, has dropped to a small
fraction of what we saw for convictions before 2003.
129
• Forensic fraud has declined sharply in all exonerations from convictions
since 2003, and among those with forensic evidence problems.
130
On the other hand, the rate of federal white-collar crime exonerations with official
misconduct has doubled among exonerations for convictions since 2003.
131
• Increases in misconduct—unlike decreases—are not hard to spot. In fact,
since more exonerations in recent cases with misconduct will continue to occur,
the size of an observed increase will go up over time.
131
• XII. Discussion and Conclusions
133
Why do Law Enforcement Officials Commit Misconduct?
133
We address this question by examining the conduct of the officials who committed or
permitted misconduct that led to many false convictions.
We conclude that the main causes are systemic: pervasive practices that permit or
reward bad behavior; lack of resources to train, supervise and conduct high
quality investigations and prosecutions; and ineffective leadership by police
commanders, crime lab directors and chief prosecutors.
133
• Ken Anderson, was the district attorney of Williamson County, Texas, who
prosecuted Michael Morton for the murder of his wife, and obtained a life
sentence in 1987. Anderson concealed potent exculpatory evidence that
could have cleared Morton and led to the real killer—who killed another
woman in 1988. Morton was exonerated by DNA in 2012.
134
❖ Why did Anderson conceal this evidence? Our best guess is that he
believed Morton was guilty, paid little attention to evidence to the
contrary—and concealed that evidence because that was his regular
practice, it made winning easier, and no one had stopped him before.
135
❖ Anderson set the tone for his office. We know that one subordinate
followed Anderson’s lead and concealed evidence to convict an
innocent defendant. There were probably others.
135
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxiv • National Registry of Exonerations • September 1, 2020
•
Page
• Richard M. Daley was chief prosecutor in Chicago in 1982 when he received
a detailed report of torture of a murder suspect by then Chicago Police
Lieutenant Jon Burge, who ran a systematic program of torturing suspects,
mostly Black men, in the1970s and ‘80s. Daley ignored it, and his office
continued to use confessions that Burge obtained by torture.
136
❖ Why did Jon Burge act as he did? Sadistic racists exist, and some
become police officers. The real question is why he was not stopped,
given that many people—including Daley—knew what he was up to.
138
❖ Burge’s superior in the police department could also have stopped
him, but Daley had more power. He could have stopped using
confessions obtained by torture, and he could have prosecuted
Burge and his men for numerous violent felonies—but he didn’t.
139
❖ Why was Burge given free rein? Most likely, Daley and others
thought the defendants were guilty, wanted murder convictions,
didn’t worry about the means—and didn’t mind the torture of Black
men they believed were murderers.
140
• Joyce Gilchrist was fired as a supervisor of the Oklahoma City Police
Laboratory in 2001, after 16 years at that lab. By then, Gilchrist was known to
have committed forensic fraud in several cases. That count has grown to
dozens of forensic fraud cases, including six exonerations.
140
❖ Why did Gilchrist pursue a career in forensic fraud? It made her a
star. She received a citation from the police department, a
commendation from the district attorney, an early promotion, and
was named Civilian Police Employee of the Year.
141
❖ Why were police and prosecutors so enthusiastic about Gilchrist?
They were warned about her, and some thought her results were too
good to be true, but she got convictions other analysts couldn’t—so
they used her.
142
• Officers Iannotto, Palmer, Pecorale, Martin, Visconti and Bishop, and
Detective Massanova all participated in the investigation of a fatal
shooting in New York in November 1990. They soon identified an innocent
suspect and brought him to the scene—in handcuffs and wearing a jacket turned
inside out to resemble the shooter’s jacket—where several witnesses urged each
other to identify him. The next day they put the disheveled suspect in a lineup
with well-groomed police cadets, and he was misidentified again.
143
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxv • National Registry of Exonerations • September 1, 2020
•
Page
❖ Why did these officers conduct what an appellate court later
describe as “a series of identifications [that] were both improper and
prejudicial”? It seems to have been just routine: they thought they
had the shooter, and got the evidence to convict him in the easiest,
quickest manner.
144
Can We Reduce Official Misconduct in Criminal Cases?
144
We discuss possible changes in three contexts: possible Reforms that affect rules,
resources, and accountability; the Local Leadership and Culture of crime labs,
police forces and prosecutors’ offices; and National Patterns of law enforcement,
as shaped by the federal Department of Justice and by national culture.
144
Reforms
145
• Rules
145
❖ Procedural rules. In response to Michael Morton’s exoneration, the
state legislature provided “open file” discovery in criminal cases in
Texas—a procedural rule regulating conduct after evidence has been
gathered. Such rules may reduce some types of misconduct, if they
are enforced.
145
❖ Evidence-gathering rules specify, for example, how a lineup should
be run, or that interrogations must be recorded. They directly
improve the quality of investigations; they may also prevent
misconduct in investigations more effectively than procedural rules—if
they are obeyed.
145
Resources. In 1992, New York had 2,000 plus murders; in 2018, a wealthier
city had 287. Lack of resources—because of huge caseloads or other reasons—
tempts officials to close cases by cheating, lying and concealing, and makes
suspects vulnerable to misconduct, because police don’t look for evidence
that clears them, and overworked defense lawyers can’t fill the gap.
146
• Accountability. Discipline for misconduct in past exonerations is too slow
and uncommon to prevent much misconduct in other cases. Comparatively
mild contemporaneous sanctions for low level infractions, by the agencies
that employ the officials, are likely to be more effective—but that is a form of
ongoing supervision that requires adequate resources.
149
Local Leadership and Local Culture
153
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxvi • National Registry of Exonerations • September 1, 2020
•
Page
The American system of criminal justice is run primarily by thousands of
independent local prosecutors, police chiefs and other administrators. Most
reforms depend on their leadership and ability to change local work cultures.
155
• Crime Labs in the United States are usually run by local police departments.
Experts agree that only independent crime labs can eliminate conflicts of
interest and provide reliable technical supervision. In 2014, after a run of
disastrous errors and misconduct, Houston replaced its police lab with a
highly regarded independent lab—after a costly 12-year transition.
156
• Police. There’s wide agreement that police reform in America is hard to
achieve and harder to sustain, in part because of the influence of police
culture. But the reforms that receive most attention concern police authority
over civilians, especially use of force and race relations. Those that concern us—
procedures for conducting criminal investigations—may be easier to
attain.
160
❖ Recorded interrogations are the most effective means for preventing
false confessions and misconduct in interrogations. In 2002, recording
was required in two states; by 2019, it was 24 states and the federal
government (which had prohibited it)—a sea change that was led by
numerous local police forces that adopted the reform before their
states.
162
❖ Improved eyewitness identification procedures prevent both false
convictions and misconduct in identification procedures. By 2020, 31
states had reformed their identification practices in some manner.
As with recorded interrogations, these statewide reforms were adopted
after hundreds, if not thousands, of local police departments did so on
their own.
163
• Prosecutors. Chief local prosecutors are the most powerful actors in
our system of criminal justice. Like police chiefs, they are constrained by local
culture and politics, but they have greater control over their own agencies.
They also have unreviewable power to decide who to charge and for what
crimes, and effectively control the punishments most convicted criminals
receive.
166
• A chief prosecutor can prevent misconduct that causes false convictions in
many ways: order deputies to follow correct procedures; discipline those
who commit misconduct; drop (or not file) cases tainted by misconduct;
prosecute officials who abuse witnesses or suspects or obstruct justice;
reinvestigate past convictions of defendants who might be innocent.
168
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxvii • National Registry of Exonerations • September 1, 2020
•
Page
• In the past dozen years, more than 60 local prosecutors, many in major
cities, have created Conviction Integrity Units (CIUs) that helped
exonerate hundreds of innocent defendants. Exonerations mostly occur
long after convictions, but may deter some misconduct all the same.
168
• Since 2014, several progressive prosecutors have been elected in major
counties on platforms that include preventing false convictions. All inherited
or created CIUs, and several have enacted policies to prevent future false
convictions: open file discovery, and lists of police officers they will not call
as witnesses because of past misconduct.
169
• Progressive prosecutors might also limit misconduct by police and
forensic analysts by scrutinizing the evidence they present, refusing to
file charges when they have doubts, and if warranted, prosecuting officials for
criminal conduct.
169
• Progressive prosecutors have attracted substantial institutional and
political opposition from police and police unions, from judges, and from other
prosecutors. Their impact will depend on politics: Will those in office be
reelected? Will others join their ranks? Time will tell.
170
National Patterns
172
Even successful reforms by crime lab directors, police chiefs and local
prosecutors will, at best, turn America into a patchwork of counties with widely
divergent practices, some effective at combating misconduct, some not—unless
change takes place at the national level.
172
• The United States Department of Justice (DOJ) can foster reform, lead
by example (as it did on recording interrogations in 2014), and it can take direct
action to curb local police misconduct. Recently, it has retreated on all
fronts.
173
❖ Reforming forensic science. For several years after 2009 DOJ
played a leading role in a coordinated effort to reform the use of
forensic science in the United States. That was terminated in 2017.
173
❖ Policing local police. Federal prosecutions of police officers who
planted drugs on defendants have led to the dismissal of hundreds of
cases, including 66 exonerations in Chicago. Before 2017, DOJ also
obtained 40 decrees requiring police departments to systematically
change their practices—but not since.
175
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxviii • National Registry of Exonerations • September 1, 2020
•
Page
❖ Leading prosecutors by example. Prosecutorial misconduct in
federal white-collar cases is a startling example of bad practice. It’s
more common than in any other category of exonerations, and the
number of exonerations with prosecutorial misconduct has increased
in the past 18 years.
176
• National Culture. We discussed the culture of individual offices and
departments, or particular counties. But culture also exists—and can
change—at a national level.
177
❖ Questioning children. As we discussed, a nationwide practice of
abusive questioning of supposed victims of child sex abuse has
been abandoned since the mid-1990s.
177
❖ Forensic fraud. A similar change may be taking place with forensic
fraud, at least in investigations of violent crimes. The number of known
cases has decreased by more than 90% since 2003.
178
❖ Violence in interrogations. In 1931, a national commission
initiated a program to eliminate violence in interrogations, which
was routine across the country. By the late 1960s, beatings and torture
were rare; since 2003, they’ve nearly disappeared. This was a major
cultural change in criminal investigation in America. We could see
another.
179
Coda: A quick summary of our thoughts on reforms.
181
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxix • National Registry of Exonerations • September 1, 2020
Table of Contents
I. Introduction ............................................................................. 1
II. Background ............................................................................. 3
1. CONTEXT ....................................................................................................................... 3
2. THIS STUDY ................................................................................................................... 6
a. Our data ........................................................................................................................ 6
b. Advantages and limitations .......................................................................................... 9
III. The Frequency of Official Misconduct .................................. 11
1. IN GENERAL .................................................................................................................11
2. THE COMMISSION AND THE DISCOVERY OF MISCONDUCT ...................................13
3. MISCONDUCT BY CRIME .............................................................................................15
a. Violent Felonies ........................................................................................................... 15
b. Non-violent crimes in general .................................................................................... 20
c. Drug crimes .................................................................................................................. 21
i. Group exonerations: Cook County, Illinois ........................................................... 21
ii. Guilty pleas: Harris County, Texas ..................................................................... 24
d. White-collar Crimes .................................................................................................... 26
e. Misdemeanors............................................................................................................. 26
4. RACIAL PATTERNS ......................................................................................................27
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxx • National Registry of Exonerations • September 1, 2020
5. CATEGORIES OF MISCONDUCT .................................................................................29
a. Witness Tampering ..................................................................................................... 30
b. Misconduct in Interrogations ...................................................................................... 31
c. Fabricated Official Evidence......................................................................................... 31
d. Concealing Exculpatory Evidence .............................................................................. 32
e. Misconduct at Trial ..................................................................................................... 33
IV. Witness Tampering .............................................................. 34
1. GENERAL PATTERNS IN WITNESS TAMPERING ......................................................35
2. PROCURING FALSE TESTIMONY ...............................................................................37
3. TAINTED IDENTIFICATIONS ........................................................................................39
4. IMPROPER QUESTIONING OF A CHILD VICTIM.........................................................41
V. Misconduct in Interrogations ............................................... 45
1. BACKGROUND .............................................................................................................45
a. Misconduct and permissible interrogation techniques ................................................ 46
b. The frequency of false confessions, in Chicago and elsewhere ..................................... 47
2. WHAT COUNTS AS COERCIVE MISCONDUCT IN AN
INTERROGATION? .......................................................................................................48
3. COERCIVE MISCONDUCT IN INTERROGATIONS THAT PRODUCE
FALSE CONFESSIONS .................................................................................................49
a. Violence ......................................................................................................................... 50
i. Torture in Chicago .................................................................................................. 50
ii. Violence in interrogations of suspects with mental disabilities ............................. 53
b. Sham Plea Bargaining and other lies about the law ..................................................... 53
c. Threats to Third Parties ................................................................................................ 54
4. PERMITTED INTERROGATION PRACTICES THAT LEAD TO FALSE
CONFESSIONS .............................................................................................................55
a. Lying about the investigation ..................................................................................... 55
b. Permissible promises and threats .............................................................................. 56
c. Feeding the suspect details of the crime ...................................................................... 57
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxxi • National Registry of Exonerations • September 1, 2020
d. Interrogating a juvenile with no parent present ........................................................ 59
e. General patterns ......................................................................................................... 59
5. MISCONDUCT IN THE INTERROGATION OF CODEFENDANTS ................................60
VI. Fabricated Official Evidence ................................................ 65
1. FORENSIC FRAUD........................................................................................................65
2. FAKE CRIMES ...............................................................................................................68
a. Planted evidence ........................................................................................................... 68
b. Phony assaults............................................................................................................... 69
3. FABRICATED CONFESSIONS......................................................................................70
VII. Concealing Exculpatory Evidence ......................................... 74
1. THE DUTY TO DISCLOSE EXCULPATORY EVIDENCE ..............................................76
a. Brady v. Maryland and the “Materiality” Requirement.............................................. 76
b. Other Legal Bases for the Duty to Disclose................................................................. 78
i. Professional Responsibility ....................................................................................78
ii. Pretrial Discovery ................................................................................................... 79
2. CONCEALING EXCULPATORY EVIDENCE, BY CRIME..............................................81
3. WHO DOES THE CONCEALING? .................................................................................81
4. WHAT WAS CONCEALED? ..........................................................................................83
a. Objects vs. Information ............................................................................................... 83
b. Substantive Evidence vs. Impeachment ..................................................................... 85
c. Concealing Other Misconduct ..................................................................................... 86
5. CATEGORIES OF CONCEALED INFORMATION .........................................................89
a. Impeachment of prosecution witnesses ..................................................................... 89
i. Incentives to testify ................................................................................................ 89
ii. Inconsistent statements ........................................................................................ 90
iii. Criminal records and histories of dishonesty ..................................................... 90
b. Substantive evidence of innocence .............................................................................. 91
i. Forensic tests ........................................................................................................... 91
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxxii • National Registry of Exonerations • September 1, 2020
ii. Alternative suspects ............................................................................................... 92
iii. “I don’t see him” and “Not the guy” ....................................................................... 93
iv. Alibi evidence ......................................................................................................... 94
v. No crime ................................................................................................................ 94
VIII. Misconduct at Trial ............................................................... 96
1. POLICE PERJURY ........................................................................................................96
2. TRIAL MISCONDUCT BY PROSECUTORS ..................................................................98
a. Permitting Perjury ........................................................................................................ 98
b. Lying in Court ............................................................................................................. 100
c. Improper Statements in Closing Argument or Cross-examination .............................103
IX. Federal Cases ...................................................................... 108
1. WHITE-COLLAR CRIMES ........................................................................................... 110
2. MISCONDUCT BY PROSECUTORS ........................................................................... 112
X. Discipline ............................................................................. 115
1. IN GENERAL ............................................................................................................... 115
2. DISCIPLINE BY CATEGORY OF GOVERNMENT OFFICIAL ..................................... 119
a. Prosecutors ................................................................................................................... 119
b. Police officers ............................................................................................................... 121
i. Criminal conviction in related cases ...................................................................... 121
ii. Unreported employment and professional discipline .......................................... 123
c. Forensic Analysts ......................................................................................................... 124
3. SUMMARY ................................................................................................................... 125
XI. Changes in Official Misconduct over Time ......................... 127
1. IN GENERAL ............................................................................................................... 127
2. IMPROPER QUESTIONING OF A CHILD VICTIM....................................................... 129
3. VIOLENCE AND OTHER MISCONDUCT IN INTERROGATIONS ............................... 129
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page xxxiii • National Registry of Exonerations • September 1, 2020
4. FORENSIC FRAUD...................................................................................................... 130
5. FEDERAL WHITE-COLLAR CRIME CASES ............................................................... 131
XII. Discussion and Conclusions ................................................ 133
1. WHY DO LAW ENFORCEMENT OFFICIALS COMMIT MISCONDUCT? ................... 134
a. Ken Anderson – Routine Practice ................................................................................ 134
b. Richard M. Daley – Encouraging Brutality ................................................................. 136
c. Joyce Gilchrist – Rewarding Fraud ............................................................................. 140
d. Officers Iannotto, Palmer, Pecorale, Martin, Visconti and Bishop, and Detective
Massanova – Closing Cases ............................................................................................. 143
2. CAN WE REDUCE OFFICIAL MISCONDUCT IN CRIMINAL CASES? ....................... 144
a. Categories of Reforms .................................................................................................. 145
i. Rules........................................................................................................................ 145
ii. Resources ............................................................................................................... 149
iii. Accountability ...................................................................................................... 153
b. Local leadership and local culture ............................................................................... 155
i. Crime Labs .............................................................................................................. 156
ii. Police ...................................................................................................................... 160
iii. Prosecutors ........................................................................................................... 166
c. National Patterns ......................................................................................................... 172
i. The United States Department of Justice .............................................................. 173
ii. National culture .................................................................................................... 177
3. CODA ........................................................................................................................... 181
Methodological Appendix ............................................................ 182
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 1 • National Registry of Exonerations • September 1, 2020
I. Introduction
The National Registry of Exonerations lists all known defendants who were convicted of crimes
in the United States and then exonerated by new evidence of innocence since 1989—a total of
2,400 cases as of February 27, 2019. In 1,296 of those cases, 54%, misconduct by government
officials contributed to the defendants’ wrongful convictions. More than a third of all
exonerations included misconduct by police officers, nearly as many involved misconduct by
prosecutors, about one in seven included misconduct by other government officials, and quite a
few had misconduct by more than one type of government official.
To state the obvious, this is not a measure of the frequency of official misconduct in all criminal
cases. We cannot estimate the general rate of misconduct in criminal investigations or
convictions from data on innocent defendants who were convicted and then exonerated. What
we can say is that official misconduct is a major cause of convictions of innocent defendants.
The exonerations in which the misconduct occurred run the gamut of crime. At one end of the
spectrum, 93 innocent defendants were sentenced to death at least in part because of official
misconduct. For example:
On May 19, 1975, Frank Harold was gunned down and killed in a robbery in
Cleveland, and a bystander was seriously wounded. Within days, 12-year-old
Eddie Vernon told the police that he had witnessed the murder and identified 18-
year-old Ricky Jackson as the gunman and brothers Ronnie and Wiley
Bridgeman as accomplices who fled the scene in a car with Jackson.
Jackson and the Bridgeman brothers were tried separately and convicted in
August and September, 1975. All three were sentenced to death. No physical
evidence connected them to the crimes, and none of the three had a criminal
record. In each case, the conviction depended entirely on Eddie Vernon’s
testimony.
At a hearing in November 2014, Vernon testified that he had not seen the robbery
or murder but had heard a rumor that Jackson and the Bridgemans were
involved, and he came forward and lied because he wanted to help the police. He
said that he had tried to recant before trial, but the police told him that because
he was too young to go to jail they would arrest his parents for perjury if he
Government Misconduct and
Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 2 • National Registry of Exonerations • September 1, 2020
backed out—so he went ahead and testified at the trials. The police never revealed
that conversation to the court or the defendants.
By the time of the hearing, lawyers for the defendants had also learned that a
suspect, whose car was seen at the scene—but who was ignored after Vernon
came forward—pled guilty in 1976 to more than a dozen counts of aggravated
robbery.
All three defendants were exonerated in 2014. Their death sentences had been
commuted to life in prison. Ronnie Bridgeman had been paroled after 29 years in
prison; Jackson and Wiley Bridgeman had each served more than 39 years for a
crime they did not commit.
At the other end of the scale, some exonerations with official misconduct were for
misdemeanors:
In July 2014, Wassillie Gregory was charged with “harassment” of a police officer,
a misdemeanor, in Bethel, Alaska. The officer wrote in his report that Gregory
was “clearly intoxicated” and that “I kindly tried to assist Gregory into my cruiser
for protective custody when he pulled away and clawed at me with his hand.”
Gregory pled guilty, without the benefit of a defense lawyer. Normally, that
would be the end of the story. But this arrest was witnessed by an anthropologist
from Arizona who was doing seasonal research in Alaska, and she filed a
complaint with the police department. Gregory was exonerated a year later
because a surveillance video showed the officer handcuffing him and then
repeatedly slamming him onto the pavement.
The police in Ricky Jackson’s case threatened the central witness in the case and forced him to
testify falsely against the defendants. They also concealed critical exculpatory evidence, the fact
that the only supposed eyewitness tried to recant his accusation and told them it was a lie.
In some ways, the misconduct by the police officer in Wassillie Gregory’s case was worse than
what Ricky Jackson and the Bridgeman brothers endured. He beat and seriously injured the
defendant with no justification—and, of course, concealed the beating—and then he lied and
framed the defendant for an assault that never occurred. This all happened, however, in a far
less serious context, a misdemeanor prosecution for which the defendant received a suspended
sentence and probation.
Many types of misconduct by prosecutors, police officers and others contributed to the
convictions of these exonerated defendants. We will describe them in detail, but we begin with
background information on the issue and the nature of our data, followed by a brief summary of
the range of misconduct we have observed.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 3 • National Registry of Exonerations • September 1, 2020
II. Background
1. CONTEXT
Misconduct by law enforcement officials in the United States has received a lot of attention in
recent years. The lion’s share has been directed at police misconduct and was catalyzed by Black
Lives Matter,
1
a national movement that gained prominence in 2014 after the police killings of
Michael Brown in Ferguson, Missouri, and Eric Garner in New York City, and has grown
immensely in size and impact since the killing of George Floyd in Minneapolis in 2020.
2
The main focus of Black Lives Matter, and of the demands and proposals for reform it has
generated, is police violence and police treatment of African Americans and other people of
color. Our research is limited to misconduct that produces false or misleading evidence that is
used to convict innocent people, or that conceals true evidence that could help clear them. For
police, that means misconduct in investigating crimes—one of the less visible of the many tasks
they perform.
Most unjustified police violence, including the killing of unarmed Black men and women, does
not involve convicting innocent people in court. But there is an overlap. Some police
investigations do include violence, racism or both, as we discuss, and the movement that was
sparked by these terrible killings has also called attention to other types of misconduct in
criminal cases by police and by prosecutors as well.
Prosecutors work entirely within the legal system; obtaining criminal convictions is their main
task. Prosecutorial misconduct is primarily misconduct that may contribute to false convictions.
As the number of exonerations has climbed over the past 30 years, concern about prosecutorial
misconduct has also increased.
Over the past decade, the most prominent voice on prosecutorial misconduct was probably that
of Alex Kozinski, former Chief Judge of the United States Court of Appeals for the Ninth Circuit.
In a widely-quoted opinion in 2013, Judge Kozinski wrote that a major form of prosecutorial
1
See Black Lives Matter, https://blacklivesmatter.com/.
2
Larry Buchanan, Quoctrung Bui and Jugal K. Patel, Black Lives Matter May Be the Largest Movement in U.S.
History, N.Y. TIMES, July 3, 2020.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 4 • National Registry of Exonerations • September 1, 2020
misconduct—concealing exculpatory evidence—had “reached epidemic proportions in recent
years,”
3
and that courts were partly responsible because they failed to take action against the
offending prosecutors.
In a law review article two years later, Kozinski expanded at length on that critique, and
observed that “there are disturbing indications that a non-trivial number of prosecutors—and
sometimes entire prosecutorial offices—engage in misconduct that seriously undermines the
fairness of criminal trials.”
4
One of the main indications Judge Kozinski pointed to was the high
number of exonerations since DNA exonerations began in 1989.
Prior research on official misconduct roughly tracks public attention. There has been a great
deal of interest and writing about misconduct by police officers in their interactions with
civilians on the street: corruption, violence, racial and ethnic prejudice.
5
That behavior, however
abhorrent, is not the sort of misconduct we address here since it does not produce false evidence
of guilt or conceal true evidence of innocence; and little of the writing on police misconduct
analyzes systematic data on police behavior.
6
There is more systematic research on misconduct by prosecutors, mostly by journalists.
In 1999, Ken Armstrong and Maurice Possley reported in the Chicago Tribune that since 1963,
at least 381 homicide convictions across the United States were reversed “because prosecutors
concealed evidence suggesting innocence or presented evidence they knew to be false.”
7
Virtually
no disciplinary actions were taken against the hundreds of prosecutors involved: one was fired
but reinstated, another was suspended for 30 days. A later article in the same series identified
42 prosecutors in Chicago who obtained criminal convictions that were later reversed because of
3
United States v. Olsen, 737 F.3d 625, 631 (9th Cir. 2013).
4
Alex Kozinski, Preface: Criminal Law 2.0, 44 Geo. L.J. Ann. Rev. Crim. Proc. iii, xxii (2015).
5
See, e.g., Rayman, Graham, The NYPD Tapes: Inside Bed-Stuy's 81st Precinct, The Village Voice, May 4, 2010,
for the first in a series of articles on police misconduct in New York City; links to the entire series can be found at
https://perma.cc/56U2-7J6V; Josefa Velasquez, Greg B. Smith & Reuven Blau, The Complaint Files NYPD Unions
Don’t Want You to See, The City, July 31, 2020 (published in partnerships with ProPublica).
6
The major exceptions are studies of racial profiling—the discriminatory selection of African Americans and other
minorities for police stops, searches, and occasionally arrests and prosecutions. See Jeffrey Fagan, Amanda Geller,
Garth Davies, and Valerie West, "Street stops and Broken Windows revisited: The demography and logic of
proactive policing in a safe and changing city," in Race, ethnicity, and policing: New and Essential Readings, edited
by Stephen K. Rice and Michael D. White. New York, NY: NYU Press (2010); Amanda Geller & Jeffrey Fagan, Pot as
Pretext: Marijuana, Race, and the New Disorder in New York City Street Policing, J. OF EMPIRICAL LEGAL STUD. Vol. 7
(4), pp. 591-633 (2010); Samuel R. Gross & Katherine Y. Barnes, Road Work: Racial Profiling and Drug Interdiction
on the Highway, 101 Mich. L. Rev. 651 (2003).
Racial profiling does produce some criminal convictions, but in the contexts that have been studied systematically,
the great majority of its victims are not arrested let alone convicted. These widespread practices of systematic racial
discrimination produce few false convictions. Those who are charged are usually guilty—of minor crimes that are
routinely ignored when committed by white people.
7
Ken Armstrong & Maurice Possley, Trial & Error: How Prosecutors Sacrifice Justice to Win. Part 1: The Verdict:
Dishonor, CHI. TRIB. (Jan. 11, 1999). Sixty-seven of the defendants in those cases were sentenced to death; by 1999,
24 of those death-sentenced defendants had been exonerated.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 5 • National Registry of Exonerations • September 1, 2020
their misconduct, and who not only escaped any meaningful adverse consequences, but went on
to become judges.
8
In 2003, the Center for Public Integrity released a study of more than 11,000 state court
criminal cases across the country since 1970 in which prosecutorial misconduct was alleged.
Courts reduced sentences, dismissed charges, or vacated convictions in more than 2,000 of
those cases, but only 44 prosecutors were the subject of state bar complaints, and of those, only
two were disbarred and 12 were suspended.
9
A study in the Pittsburgh Post-Gazette in 1998 found much the same thing for prosecutorial
misconduct in federal criminal cases. The Department of Justice investigated only 9 percent of
some 4,000 complaints of misconduct by its prosecutors over 20 years, and of those, only 4
percent were found to have merit—approximately 15 cases all told.
10
Twelve years later, USA
Today reported that little had changed. From 1997 to 2010, judges found misconduct by federal
prosecutors in 201 cases, but only six were disciplined by bar authorities.
11
The most detailed study on the subject was released by Kathleen Ridolfi and Maurice Possley of
the Northern California Innocence Project in 2010.
12
They collected all decisions in which courts
found that state prosecutors in California had committed misconduct in trials that led to
convictions, 707 cases from 1999 through 2007. Only 159 of those convictions were reversed—in
the remainder, the misconduct was deemed “harmless”—and, despite a legal obligation that
California courts report all such findings to the California State Bar, only six California
prosecutors were disciplined in any manner for misconduct in a criminal case in that nine-year
period.
Taken together, the studies of prosecutorial misconduct reached two main conclusions: (i) a
substantial number of prosecutors commit misconduct in criminal cases, and (ii) almost none
are disciplined for it. This is an important contribution to our understanding of the problem.
The journalists who conducted them did an impressive job of searching through thousands of
cases to locate the small minority in which courts found that misconduct had occurred, and then
determining whether the prosecutors involved were sanctioned.
8
Ken Armstrong & Maurice Possley, Trial & Error: How Prosecutors Sacrifice Justice to Win. Part 5: Break Rules Be
Promoted, CHI. TRIB. (Jan. 14, 1999). A third article in the Chicago Tribune, by Armstrong and Steve Mills several
months later, found that, “More than 10 percent of Illinois’ death-penalty cases have been reversed …because
prosecutors took some unfair advantage that undermined a trial’s integrity.” Ken Armstrong & Steve Mills, The Failure
of the Death Penalty in Illinois: Part 1: Death Row Justice Derailed: Bias, Errors, and Incompetence in Capital Cases
Have Turned Illinois’ Harshest Punishment into Its Least Credible, Chi. Trib. (Nov. 14, 1999).
9
CTR. FOR PUB. INTEGRITY, HARMFUL ERROR: INVESTIGATING AMERICA’S LOCAL PROSECUTORS (2003). See Michael J.
Sniffen, “Study finds thousands of cases of misconduct by prosecutors across U.S.,” Journal Times (June 26, 2003).
10
William Moushey, Win at All Costs, Pitt. Post-Gazette, Nov. 22, 1998, at A.
11
Brad Heath & Kevin McCoy, Prosecutors’ Conduct Can Tip Justice Scales, USA Today (Sept. 23, 2010).
12
Kathleen M. Ridolfi & Maurice Possley, Preventable Error: A Report On Prosecutorial Misconduct In California,
1997–2009, A Veritas Initiative Report (2010). See also Richard A. Rosen, Disciplinary Sanctions against
Prosecutors for Brady Violations: A Paper Tiger, 65 N.C. L. REV. 693 (1987).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 6 • National Registry of Exonerations • September 1, 2020
These studies, of course, have limitations. The central one is that they are all based on official
findings that misconduct occurred, usually in written opinions by judges (or in the Pittsburgh
Post-Gazette study, internal Department of Justice memoranda). Most criminal cases, with or
without misconduct, do not produce written court opinions; in fact, most convictions based on
guilty pleas leave virtually no substantive records at all. In cases that do include court opinions,
misconduct is often overlooked by lawyers and judges alike because nobody knew about it at the
time. For other convictions, valid claims of prosecutorial misconduct are raised but rejected by
courts because critical evidence to prove those claims has not yet come to light. We see that
regularly in the records of cases that eventually do produce exonerations—and then, even when
misconduct is an undisputed fact, the case may simply be dismissed with no formal finding.
In short, because prior studies rely on cases with findings of misconduct, and include no data on
cases without misconduct, they can’t speak to the rate of misconduct in all criminal cases, or in
any subset of them, or to the features that distinguish cases with misconduct from those
without, or to differences between the acts of misconduct that are identified by courts and those
that they miss.
Our own study has a much richer database. We look at all convictions that end in exoneration,
whether or not official misconduct occurred or was found by a court. Exonerations receive much
more attention than run of the mill convictions, and we have done the best we can to assemble
detailed information on all exonerations that we know about. That enables us to identify many
instances of misconduct that cannot be found in court opinions.
13
In other words, our data enable us to avoid some of the limitations imposed of prior research—
but not all.
2. THIS STUDY
a. Our data
Our data come from the files assembled by the National Registry of Exonerations on all known
exonerations in the United States since the beginning of 1989. We define an exoneration as
follows:
“A person has been exonerated if he or she was convicted of a crime and, following a
post-conviction reexamination of the evidence in the case, was either: (1) declared to be
factually innocent by a government official or agency with the authority to make that
13
For example, Ridolfi & Possley report that improper closing arguments were by far the most common type of
misconduct by prosecutors found by California courts, about seven times more frequent than concealing exculpatory
evidence. Ridolfi & Possley, supra, at p. 25. As you’ll see, our data on exonerations show that prosecutors conceal
exculpatory evidence about ten times more often than they make improper arguments to the jury. Infra sections VII.3
and VIII.2.c. That reflects the different cases we consider.
Ridolfi & Possley’s cases are overwhelmingly initial direct appeals from criminal convictions, proceedings that are
limited to reviewing the record of what happened in the trial court. Improper arguments are made in open court and
are visible in trial transcripts. Concealed exculpatory evidence, by its very nature, is rarely known at that stage of the
proceedings, but it’s common among cases that end in exoneration, typically after long reinvestigations of the cases,
and usually with no written finding on misconduct.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 7 • National Registry of Exonerations • September 1, 2020
declaration; or (2) relieved of all the consequences of the criminal conviction by a
government official or body with the authority to take that action.”
14
We rely on the decisions of government officials to classify a case as an exoneration. Our task is
limited to determining whether a convicted defendant was declared innocent or relieved of all
consequences of the conviction, and—if the latter—whether that action “was the result, at least
in part, of evidence of innocence that either (i) was not presented at the trial at which the person
was convicted; or (ii) if the person pled guilty, was not known to the defendant and the defense
attorney, and to the court, at the time the plea was entered.”
15
The Registry relies entirely on non-confidential data, primarily court records and exhibits, other
official documents, news reports, and interviews with attorneys and others who worked on the
cases in their professional capacity.
The Registry is an ever-changing public archive. We add cases steadily, about 200 a year, both
exonerations that occurred in that year and ones from years or decades before. This report is
based on a snapshot of the data in the archive (or rather, several snapshots). It includes the first
2,400 exonerations posted in the Registry, those posted as of February 27, 2019. The total has
grown to 2,663, as of this writing, and counting.
Those 2,400 cases, however, have not remained entirely static. We invite corrections and new
information on cases we have listed, and we update and amend them regularly. We have
removed several cases over the past eight years when new information has led us to conclude
that they did not in fact meet our criteria, and we have changed many more. We have already
learned, for example, that two exonerations among the first 2,400 in the Registry included
official misconduct that we did not know about when we stopped coding the data for this
Report.
16
We are not aware of any situation in which these changes affect any general pattern we
discuss.
We link to the Registry many times in this report. For example, in the Introduction we linked to
our write-up of the exoneration of Ricky Jackson. The write-up that link takes you to is the
current one; it includes information on a civil lawsuit by Jackson that was settled in May 2020,
more than a year after we closed the book on additional exonerations for this Report. Similarly,
this link will take you to the list of all exonerations in the Registry with official misconduct as of
14
See National Registry of Exonerations, Glossary, which also includes a more detailed definition.
15
Id.
16
Codefendants Jean Dorval, Duquene Pierre, and James Louis were convicted of murder in New Jersey in 1996.
Pierre and Dorval were exonerated and added to the Registry in 2016 and 2018, respectively. We knew of no
misconduct in their cases, and coded them accordingly. In June 2020, months after we completed all additions and
corrections to the 2,400 exonerations that are the subject of this report, James Louis was also exonerated and we
learned for the first time that the convictions of all three men were tainted by serious official misconduct, concealing
exculpatory evidence and witness tampering. Louis’s case is not included in this Report because he was exonerated
after February 2019. Dorval and Pierre’s cases are in the database for this Report, but they are coded as not
including official misconduct because as of March 2020, we didn’t know of any.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 8 • National Registry of Exonerations • September 1, 2020
when you click—1,443 on September 1, 2020—not the set of 1,296 exonerations with misconduct
as of February 27, 2019, that this Report is based on.
“Official Misconduct” is a term of art. In general, it means that a prosecutor, police officer, or
less frequently, forensic analyst or child welfare worker violated an official duty in the
investigation or prosecution of a criminal case, and that violation contributed to the conviction
of a defendant who was later exonerated. With a handful of exceptions, all acts of misconduct we
considered fell into five broad categories: Witness Tampering, Misconduct in Interrogations,
Fabricating Evidence, Concealing Exculpatory Evidence, and Misconduct at Trial.
A few types of official misconduct are public and visible for all to see—improper closing
argument, for example. Most, however, are concealed—witness tampering, concealing
exculpatory evidence, forensic fraud, and so forth. We determine whether misconduct occurred
based on the information available to us, which inevitably means that we miss misconduct that
was successfully hidden.
17
Some forms of misconduct are deliberate by definition—perjury by a law enforcement official,
or forensic fraud—and some are always deliberate in operation, such as violence in
interrogations. Some are usually deliberate, but not inevitably—for example, misconduct in
interrogations that does not include violence—and some, in particular concealing exculpatory
evidence, may or may not be deliberate. One form of witness tampering—abusive questioning of
a child victim—was probably never a deliberate act of misconduct because the officials involved
believed it was appropriate and necessary.
The findings we report come from two overlapping datasets. We began this project in 2014 by
coding detailed data on official misconduct for the first 1,361 cases in the Registry, those that
had been posted by May 13, 2014. In April 2016, we added a series of new codes on official
misconduct to all past cases we had posted, and to our ongoing protocol for gathering and
recording data on cases that we add. Those data, which are available for all exonerations in the
Registry, present and future, are the main source we rely on in this Report.
There are, however, a few issues on which we have more detailed data for the 1,361 cases that
were posted by mid-May 2014. In particular, for some types of misconduct, that earlier study is
the only source of data on the categories of the officials—prosecutors, police, etc.—who
committed that misconduct. We discuss our use of these two datasets in more detail in the
Methodological Appendix.
For now, we note that when we report a proportion that is based on the earlier, 1,361 case
dataset, we italicize the fraction from which that proportion is derived, but when the proportion
is based on all 2,400 exonerations, we do not. Thus, we would describe the proportion of all
exonerations with any official misconduct in the 1,361 dataset as “56% (768/1,361),” while in the
2,400 dataset, the same proportion would be “54% (1,296/2,400).”
17
See infra Sections III.2 and III.3.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 9 • National Registry of Exonerations • September 1, 2020
b. Advantages and limitations
Our database is unique. We have detailed information on all known convictions of innocent
defendants who were later exonerated. These are the criminal cases in which official misconduct
produces its worst results: the conviction of innocent defendants. Because we have data on all
exonerations, not just those with misconduct, we know that official misconduct played a role in
more than half of the convictions of the exonerated defendants. Because we have detailed
information about these convictions, we can begin to describe how this sort of misconduct
operates—who does what, when, how, and in which cases.
There’s a qualification. As we pointed out, we miss misconduct that was successfully hidden, so
the rates we report are liable to be under-estimates.
We cannot, however, use these data to estimate the rate of misconduct in all criminal
convictions. We’re confident that official misconduct that might contribute to a false conviction
is much more common in cases in which false convictions did in fact occur than in those in
which they did not, but we have no measure of that second rate. That would require detailed
data on a representative sample of all criminal convictions in the United States. No such data
exist, and they would be nearly impossible to assemble. As we pointed out, the vast majority of
convictions are based on guilty pleas, which usually produce little or no record of the process
that led to the plea; and, unlike exonerations, almost none have the benefit of a reexamination of
the evidence.
Nor can we generalize from these data on exonerations to all convictions of innocent people. As
we discuss below,
18
there are strong reasons to believe that the vast majority of people who are
convicted of crimes they did not commit are never exonerated, especially those who are falsely
convicted of property or drug crimes and sentenced to probation or short terms in jail. The
frequency and the types of misconduct committed in those erroneous convictions may be
different from what we see in cases of defendants who were exonerated, usually after spending
years in prison for convictions for violent crimes.
Finally, a few words on two issues we do not address:
• We do not systematically discuss misconduct by criminal defense attorneys in
representing their clients. Strictly speaking, misconduct by defense attorneys is not
“official misconduct” because even public defenders, who are employed by the
government, represent the defendants not the state. Nonetheless, we would have
included defense attorneys in this study if we could, since they are all officers of the
court, most are hired by the state, and—judging from anecdotal evidence—their
misconduct and incompetence may do as much to produce false convictions as
misconduct by prosecutors and police officers combined.
But we can’t. The failures of defense counsel are overwhelmingly sins of omission,
especially the failure to investigate their clients’ cases. The absence of action is hard to
18
Section III.2 and Section III.3.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 10 • National Registry of Exonerations • September 1, 2020
spot. A failure to even try to contact persuasive alibi witnesses will rarely be apparent at
trial, and almost never when a guilty plea is taken. Unless such failures are actually
litigated—which is uncommon—they are likely to remain unknown.
19
• We do not discuss judicial misconduct for the same reason: absence of data. We know of
a few exonerations after trial with severe misconduct by judges. At Lamonte McIntyre’s
murder trial in Kansas in 1994, for example, neither the judge nor the prosecutor
revealed that they had been involved in a romantic relationship with each other. But
such cases are rare. It may be that the underlying misconduct is also rare, but we
wouldn’t know. It is widely believed that judges are, to say the least, unreceptive to
claims that they, or other judges, committed misconduct that prejudiced a litigant. As a
result, except in the clearest cases, lawyers steer away from such claims and focus on less
pejorative forms of judicial “error.” If it’s not litigated or investigated, we are not likely to
know about it.
19
See Samuel R. Gross & Michael Shaffer, Exonerations in the United States,1989 – 2012, at 41-43, National
Registry of Exonerations (2012), for a more complete description of this problem. We do list some exonerations in
which we know that the legal defense was severely ineffective, but only as examples.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 11 • National Registry of Exonerations • September 1, 2020
III. The Frequency of Official
Misconduct
1. IN GENERAL
Official misconduct contributed to the conviction of innocent defendants in 54% of known
exonerations—in most cases, more than one type of misconduct. There are modest differences in
the rates of official misconduct by the sex and race of the exonerees. The overall rate is 55% for
male exonerees and 46% for females, and, among major racial and ethnic groups, ranges from
57% for Black exonerees to 51% for whites. See Table 1. There are, however, considerably larger
differences by race among exonerations for some crimes, and for specific types of misconduct, as
we’ll see in Section III.4.
Table 1: Proportions of Exonerations with Official Misconduct by Characteristics of
Exonerees
Exonerees
Male
55%
Female
46%
White
52%
Black
57%
Hispanic
53%
Other Race or Ethnicity
40%
All Exonerees
54%
About a third included misconduct by prosecutors and slightly more involved misconduct by
police—the two types of officials who play central roles in all criminal convictions. Forensic
analysts and child welfare workers participate in many fewer criminal cases, and committed
misconduct in only 3% and 2% of exonerations, respectively. See Table 2.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 12 • National Registry of Exonerations • September 1, 2020
Table 2: The Frequency of Misconduct in Exonerations by Category of Official Actor*
Officials
Prosecutors
30%
Police
35%
Forensic Analysts
3%
Child Welfare Workers
2%
All Official Actors
54%
* Some exonerations include misconduct by officials in more than one category.
Exonerations with misconduct by child welfare workers (50) were all child sex abuse cases
(including two murder prosecutions that also involved child sex abuse). Those with misconduct
by forensic analysts were overwhelmingly murders (33/75) and sexual assaults (29/75), with a
scattering of other crimes. Misconduct by prosecutors and by police are tabulated separately in
Table 3. They occur regularly in exonerations from convictions in every category of crime.
Table 3: Proportion of Exonerations with Misconduct by Prosecutors and Police, by
Crime
Misconduct by
Prosecutors
Misconduct
by Police
ALL OFFICIAL
MISCONDUCT*
Murder (908)
44%
48%
72%
Child Sex Abuse (270)
25%
28%
44%
Sexual Assault (320)
20%
22%
39%
Robbery (122)
27%
17%
38%
Other Violent Crimes (270)
30%
36%
55%
Drug Crimes (317)
9%
34%
39%
White-collar Crimes (63)
57%
11%
62%
Other Non-Violent Crimes (130)
12%
25%
32%
ALL CRIMES (2,400)
30%
35%
54%
* The numbers in this column include cases with misconduct by forensic analysts and child welfare workers; some
exonerations include misconduct by officials in more than one category.
In general, the rates of misconduct by prosecutors and police occurred at comparable rates
among exonerations for particular crimes, with two exceptions:
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 13 • National Registry of Exonerations • September 1, 2020
Among drug crime exonerations, the rate of misconduct by police (34%) is almost four
times the rate of misconduct by prosecutors (9%). This is partly due to clusters of cases
in which police planted drugs on innocent suspects.
20
Among white-collar exonerations, the rate of misconduct by prosecutors (57%) is more
than five times the rate of misconduct by police (11%). White-collar exonerations have
both the highest rate of misconduct by prosecutors and the lowest rate of misconduct by
police of any crime category. That pattern is driven entirely by federal white-collar
exonerations: they constitute almost three-quarters of all white-collar cases (46/63), and
65% of them involved misconduct by prosecutors (30/46).
21
As a result, the rate of
prosecutorial misconduct is higher for white-collar crimes than for any other
exonerations, and the overall rate of misconduct is second only to murder.
These tabulations only include known acts of misconduct that contributed to the convictions of
the exonerated defendants. Many serious acts of misconduct don’t qualify. In some cases, there
was misconduct that was unrelated to the determination of the defendant’s guilt or innocence.
William Carter, for example, was beaten by police officers who planted drugs on him and
arrested him twice in Chicago in 2006—but beating up a suspect who is not questioned produces
no false evidence of guilt and conceals no true evidence of innocence. Neither does lying about
the source of information in an application for a search warrant, as an officer did in Joseph
Green’s case, since such lies play no part in the trial at which a defendant’s guilt is determined.
In other cases, state officials tried to obtain false evidence of guilt, or to hide true evidence of
innocence, but failed. Nathan Dwight, for example, was told by a detective in Rockwell, Georgia,
that he’d better come clean and confess because the all-white jury he would face would see him
as a “straight-up nigger”—but Dwight did not confess. And in Glen Nickerson’s case, police
officers threatened to charge alibi witnesses as accomplices and to have their children taken
away by the state if they testified on his behalf, but the witnesses testified anyway.
2. THE COMMISSION AND THE DISCOVERY OF MISCONDUCT
Since most people who commit misconduct hide it, if they can, we can only study those acts that
come to light. The frequency of the known official misconduct that we discuss here is a function
both of the underlying rate at which misconduct is committed and the proportion of cases in
which it is uncovered. Both factors vary from one category of exoneration to another.
Commission. Consider the case of Homer Taylor III:
In 1988 Taylor pled guilty to third-degree statutory rape in Gray Harbor,
Washington, for having consensual sex with an underage female in 1982 when he
was 24. Twenty-two years later, in 2010, Taylor was convicted for failing to
register as a sex offender and sentenced to 43 months in prison. The following
20
See Section III.3.c.i
21
See infra Section III.3.d and Section IX. By comparison, only 35% of the small number of state law white-collar
exonerations include misconduct by prosecutors (6/17).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 14 • National Registry of Exonerations • September 1, 2020
year he was exonerated because within a year of his guilty plea, Washington State
repealed the statutory rape law under which he was convicted, and the
Washington sex offender registration statute did not apply to convictions under
that repealed law. Apparently, nobody had noticed this issue before Taylor’s
appeal.
22
The Registry includes 24 exonerations of defendants like Taylor who were wrongfully convicted
of failing to register as a sex offender. None of them include any known official misconduct,
probably because none occurred. In all of these cases, the underlying problem is clearly that the
defendants, police officers, lawyers and judges involved—all of them—mistakenly believed that
the defendant was required to register when in fact he was not. It’s theoretically possible that
some government official deliberately arrested, charged or prosecuted a defendant for failing to
register while knowing that he was not obliged to do so, but we see no indication that ever
occurred.
On the other hand, Duarnis Perez, who was exonerated in 2006 after serving four years in
prison for illegally entering the United States, was a victim of official misconduct. A federal
judge found that the government had violated Perez’s constitutional rights by failing to inform
him at trial that he could not be guilty of that crime because, unbeknownst to him and his
family, he had become an American citizen 12 years earlier, at age 15, when his mother was
naturalized as a citizen.
Perez is one of four exonerees in the Registry who were convicted of immigration law offenses.
All four cases include official misconduct. Four exonerations are too few to reach any general
conclusions, but it looks like law enforcement officials are more likely to commit misconduct to
convict defendants accused of immigration offenses than to convict those charged with failing to
register as sex offenders.
Discovery. As we’ll see, the Registry includes 149 cases from Harris County, Texas (Houston), in
which defendants who pled guilty to drug possession were exonerated by lab tests that
unexpectedly showed that the materials seized from them contained no illegal drugs.
23
None of
those cases included any police misconduct, but that doesn’t mean none occurred. Officers may
have lied about where they found the suspected drugs, or about what the suspects said—both
items of evidence that might show that the arrested defendants knowingly possessed the
materials—but once the serendipitous lab results appeared, no one bothered to look any further
because exoneration was inevitable. In many murder exonerations, by contrast, finding official
misconduct is essential just to get a hearing on a claim of innocence.
22
State v. Taylor, 259 P.3d 289, 162 Wash. App. 791 (2011). The decision in Taylor led to three other Washington
State sex offender registration exonerations: the exoneration of Ollie Church, and two separate exonerations of
Michael Wheeler.
23
See infra Section III.3.c.ii.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 15 • National Registry of Exonerations • September 1, 2020
3. MISCONDUCT BY CRIME
a. Violent Felonies
Almost 78% of known exonerations in the United States are for violent felonies (1,872/2,400).
Table 4 displays the rate of known misconduct in exonerations for violent crime convictions,
with murder cases subdivided by penalty. In general, the more severe the crime, the higher the
rate of misconduct, ranging from 79% in murder cases with death sentences to 38% in robbery
cases.
Table 4: Rate of Official Misconduct in Exonerations for Violent Felonies, by Crime*
HOMICIDE
Murder (908):
72%
Death Penalty (121)
79%
Life Imprisonment (377)
73%
Term Less Than Life (410)
69%
Manslaughter (45)
60%
Felonious Assault (82)
57%
Attempted Murder (50)
48%
Arson (23)
48%
Child Sex Abuse (269)
44%
Kidnapping (15)
40%
Sexual Assault (320)
39%
Robbery (122)
38%
ALL CASES (2,400)
54%
* Table 4 does not tabulate separately 53 exonerations for violent crimes that account for fewer than 10 cases each
in our data, such as Child Abuse (9 cases) and Accessory to Murder (8 cases).
Why do exonerations for severe crimes include more known misconduct than those for less
severe offenses? The answer seems to involve both of the components of known misconduct:
official misconduct is more likely to occur in exonerations for more aggravated crimes,
especially murder, and misconduct that occurs in such cases is more likely to be detected.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 16 • National Registry of Exonerations • September 1, 2020
The starting point is a well-established fact about exonerations: the more serious the crime, the
higher the rate of exoneration.
24
A recent article estimates the relative rates of exoneration in
the Unites States from 1989 through 2016 for robbery, rape, non-capital murder and death
sentences, by setting the lowest rate—that for robbery—at 1, and calculating the ratios for other
crimes. The results are startling. The estimated rate of exoneration for non-capital murders is
more than four times the rate for rape and 37 times the rate for robbery, and the rate for death
sentences is about six times that for other murder convictions. See Table 5.
25
Table 5: Estimated Relative Exoneration Rates
The difference between the rates of exoneration for rape and robbery can be explained in large
part by the use of DNA testing to prove the innocence of convicted defendants: biological trace
evidence that can be tested for DNA is common in rape cases, but rare among robberies.
26
There is, however, no technological advantage in proof of innocence that explains the much
higher rate of exoneration for murders—especially for capital murders—than for lesser crimes.
On the contrary, most murder investigations are handicapped by the absence of live victims to
24
This pattern holds for felony convictions in the United States. Our data are limited to the United States, and include
too few misdemeanor exonerations to draw any general conclusions.
25
Samuel R. Gross, What We Think, What We Know and What We Think We Know About False Convictions, 14
OHIO J. CRIM. L. 753, 766 (2017).
26
The need for post-conviction DNA testing in rape cases has been diminishing, as DNA testing is increasingly used
before trial to prevent false convictions in the first place. As a result, the number of sexual assault exonerations with
DNA evidence has drifted downward in the last 18 years. See Table A:
Table A: Sexual Assault Exonerations with DNA Evidence by Year
2002
2003
2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
2014
2015
2016
2017
2018
2019
18
12
6
10
13
10
7
11
9
5
7
7
5
5
4
8
5
4
210
37
9
1
Death
Sentences
Non-Capital
Murder
Rape Robbery
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 17 • National Registry of Exonerations • September 1, 2020
describe the crimes and identify the criminals. Instead, the plausible explanation is that we
devote more time and resources to murder cases, at every stage:
(i) Wrongful convictions are more likely to occur in murder cases because police and
prosecutors work harder to secure murder convictions in cases with weak evidence than
they do for lesser crimes.
(ii) False murder convictions are more likely to be identified, and the innocent
defendants exonerated, because defense attorneys, journalists, courts and law
enforcement officers devote much more time and many more resources to doing so,
especially if the defendant might be put to death.
Both of those processes also increase the rate of known misconduct among murder
exonerations, along with overall rates of known exonerations.
Commission of misconduct. Murders are much more likely to produce criminal convictions than
other violent crimes. Most other felonies are never reported to the police. In 2018, for example,
only 43% of violent crimes other than murder were reported to the police,
27
but almost all
homicides are known to authorities because the bodies of the deceased must be interred or
cremated, and the cause of death will be determined to be homicide. Among crimes that are
known to the police, the clearance rate—the proportion of reported crimes for which the
criminal is identified and if possible arrested—was about double for murder, at 62%, than other
violent crimes such as rape, 33%, or robbery, 30%.
28
This is as it should be. Police and prosecutors work harder on murder cases because murder is
far worse than other crimes. They’re more motivated to solve murders, and they’re under more
pressure to do so. The main effect is what we want: more murderers are caught and convicted
than criminals who commit other crimes. An investigation that would be closed without arrest if
it were a mere robbery may end in a conviction if the robber killed one of his victims. But the
time and attention that are devoted to murders can also produce mistakes. Because the
authorities pursue difficult murder cases with weak evidence—cases that would be abandoned if
nobody had been killed—there is higher risk that innocent defendants will be convicted.
That strong impulse to secure convictions can also lead to misconduct. If a murder cannot
readily be solved, the authorities may be tempted to cut corners, jump to conclusions, and—if
they believe they have the killer—manufacture evidence to clinch the case, or hide evidence that
suggests innocence. The danger that they will go too far is magnified to the extent that the killing
is brutal and horrifying and attracts public attention—factors that also increase the likelihood
that the murder will be treated as a capital case. The rape-murder of a young girl is a telling
example:
In February 1983, ten-year-old Jeanine Nicarico was abducted from her home in
Naperville, Illinois, raped, and killed—a crime of stunning brutality. The murder
27
Rachel E. Morgan & Barbara A. Oudekerk, Criminal Victimization, 2018, U.S. Dep’t of Justice (Sept. 2019).
28
Fed. Bureau of Investigation, 2018 Crime in the United States, (last visited Mar. 4, 2020).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 18 • National Registry of Exonerations • September 1, 2020
was the subject of a long, frustrating, unsuccessful investigation—a humiliating
public failure. Thirteen months after the murder—and less than two weeks before
the local prosecutor stood for reelection—three men were indicted: Rolando Cruz,
Alejandro Hernandez, and Stephen Buckley. Cruz and Hernandez were convicted
and sentenced to death in 1985.
Ten years and two appeals later, at Cruz’s third trial, the case fell apart when a
police officer admitted he had given false testimony, perhaps inadvertently, when
he provided critical corroboration to the claim of other officers that Cruz had
made an unrecorded statement that included details of the killing that only the
criminal would know. The judge entered a judgment of acquittal. Apparently,
working under intense pressure, the police convinced themselves that they knew
who killed Jeanine Nicarico, and then manufactured evidence to convince
prosecutors and for use in court.
What if this had not been a murder case? What if the criminal had taken expensive jewelry from
the Nicarico home rather than a child—or even if he had knocked out a family member, or set
the home on fire? As long as no one was killed, chances are the case would have been closed
after a comparatively short investigation with no arrests, no trials, no police perjury, and no
false convictions.
We have no data on charges that were not pursued in less horrific crimes, but we see echoes of
this stark difference in the investigations of cases that were prosecuted and ultimately ended in
exoneration. For example, just over a quarter of all exonerations include one or more
recantations by prosecution witnesses (620/2,400), primarily murder and child sex abuse cases.
In 71% of murder cases with recantations, recanting witnesses reported that police or
prosecutors pressured them to lie (260/365); in child sex abuse cases, only 39% of recanting
witnesses said they were pressured to lie by law enforcement officials (50/129).
29
This suggests
that when the evidence in a child sex abuse investigation is weak, the authorities are more likely
to close the case with no charges, but when it’s a murder case, they’re more likely to pressure
witnesses to testify against the defendants in court—which sometimes leads those witnesses to
lie, and produces false convictions.
Discovery of misconduct. In many cases, misconduct that contributed to a false conviction only
came to light as a result of comprehensive, painstaking, long-running investigations. Murder
cases, especially those that result in death sentences, are the main candidates for detailed post-
conviction investigations. John Thompson’s case is a clear example:
In 1984, John Thompson was arrested for a highly-publicized robbery murder in
New Orleans. After his picture was published in local newspapers, he was
misidentified as the robber in an earlier carjacking. The prosecutor went to trial
in the carjacking case first, and got a robbery conviction in April 1985, so he could
29
On the other hand, 47% of child sex abuse recanters said that they had been pressured to lie by civilians, mostly
family members (60/129), compared to 7% of recanters in murder exonerations (25/365).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 19 • National Registry of Exonerations • September 1, 2020
use that conviction against Thompson at his murder trial—which he did, and he
got a death sentence a month later.
Fourteen years later, 30 days before Thompson was set to be put death, an
investigator conducted a last-ditch effort to prevent Thompson’s execution. He
discovered that there was a blood stain from the robber on the clothing of one of
the carjacking victims, that the stain was tested by the prosecution before trial,
and that the test showed that the blood did not come from Thompson. All of this
had been concealed from the defense by prosecutors, who went so far as to
improperly remove the piece of cloth with the blood stain from the police
property room and eventually lose or destroy it.
30
Thompson’s carjacking conviction was dismissed. Three years later he was
granted a new trial in the murder case, and in 2003 he was acquitted by a jury
that deliberated for 35 minutes. By then Thompson’s attorneys had also
discovered that the prosecution had concealed the fact that a witness who
testified that Thompson admitted to the murder had received a reward for his
testimony, and that prosecutors had allowed that witness to lie about that at trial.
What stands out about Thompson’s case is not that it took so long to find the misconduct, but
that it was ever found at all. As former Chief Judge Alex Kozinski of the United States Court of
Appeals for the Ninth Circuit observed, “If a prosecutor fails to disclose exculpatory evidence to
the defense, who is to know?”
31
Thompson was only exonerated from his robbery conviction because he was also convicted of
murder and sentenced to death. He was saved by a scorched-earth reinvestigation that
uncovered deeply-hidden facts. Investigations like that are expensive, time consuming and
extremely rare in criminal cases with poor defendants—except for some defendants who are
sentenced to death.
If Thompson had merely been convicted of robbery (and sentenced to 49 years in prison, as he
was) the misconduct in the carjacking case would never have been discovered. Of course, the
outrageous misconduct in Thompson’s carjacking case would probably never have occurred if
prosecutors had not treated his robbery trial as the first act of a capital murder trial.
Some murder exonerations do not require long, searching investigations. In November 2011, in
Brown County, Texas, Randall Philen was convicted of murdering his brother two years earlier.
Seven weeks later, a confidential informant gave the police the names of the real killers: four
men who broke into the victim’s home to rob him of drugs and money and killed him in the
process. Within a day, one of the four confessed and his fingerprints matched a print found at
the crime scene. Philen was released immediately and exonerated two months later.
30
Connick v. Thompson, 563 US 51, 83-83 n.4-6 (2011) (Ginzburg, J dissenting).
31
Alex Kozinski, Preface: Criminal Law 2.0, 44 GEO. L.J. ANN. REV. CRIM. PROC. iii, xxiii (2015).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 20 • National Registry of Exonerations • September 1, 2020
Philen’s fast-path exoneration did not uncover any misconduct—or, as far as we know, anything
else about the investigation and prosecution that led to his conviction. There was no need. The
first batch of new evidence included the identities of the real killers. One reason we know of
more misconduct in exonerations for murder than for rape is that murder exonerations include
many hard cases, like Thompson’s, and few easy ones, like Philen’s.
There was no time to re-investigate Philen’s case for misconduct before his exoneration, but if
anybody had been interested, there were avenues to explore after exoneration. At his week-long
trial two months earlier, the prosecutors presented extensive (and misleading) forensic
evidence, and testimony from family members and others about the relationship between Philen
and his dead brother, and about their drug dealings. Either category of evidence, or both, might
have been obtained by misconduct. But there was little incentive to pursue a claim for
compensation for the 49 days he spent in custody after conviction, and no claim of any sort was
filed. Philen’s case is an outlier among murder exonerations. The average time a murder
exoneree spent in custody from conviction to release was 13.9 years, more than 100 times what
Philen served.
b. Non-violent crimes in general
Non-violent crimes account for more than 80% of felony convictions in state courts in the
United States
32
(and, as best we can tell, at least a similar proportion of misdemeanor
convictions
33
), but only 19% of felony exonerations involve non-violent crimes (429/2,311). That
means that the overall exoneration rate for violent felonies is about 17 times higher than the rate
for nonviolent felonies, and the exoneration rate for murders is about 100 times greater—a huge
gap. Unless the accuracy of determining guilt or innocence for non-violent crimes puts murder
and sexual assault trials to shame, only a miniscule minority of wrongful convictions for crimes
without violence are ever reconsidered or reversed.
Overall, official misconduct contributed to the false convictions in 40% of exonerations for non-
violent crimes (205/510). That rate, however, varies greatly depending on the specific category
of the crime, and the nature of the investigations that proved the innocence of the defendants.
About 8% of exonerations for non-violent crimes involve several types of property crimes
(40/510),
34
and another 14% are for a variety of what we call, collectively, “regulatory” crimes,
such as failure to pay child support or register as a sex offender, immigration offenses, illegal
32
See e.g., Bureau of Justice Statistics, Felony Sentences in State Courts, 2006- Statistical Tables, NCJ 226846
(revised Nov. 22, 2010).
33
There are no systematic data on the nature of misdemeanor convictions in the United States. See generally
Alexandra Natapoff, Punishment Without Crime, Basic Book: New York (2018); but see, Sandra Mayson & Megan
Stevenson, Misdemeanors by the Numbers, 61 B.C. L. REV. 971 (2020), (containing a tabulation of misdemeanor
filings in 8 jurisdictions across the country; in one (Chicago), violent misdemeanors appear to account for perhaps
30% of total; in all others, less than 20%. Id. at p. 1000, figure 3).
34
These cases include burglary and other unlawful entries, destruction of property, possession of stolen property,
and theft.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 21 • National Registry of Exonerations • September 1, 2020
possession of a gun, and so forth (74/510).
35
The rates of official misconduct among these
exonerations are comparatively modest, 25% for property offenses (10/40) and 31% for
“regulatory” crimes (23/74).
Three-quarters of exonerations for non-violent crimes involve two well-defined but very
different categories of crime: drug crimes, which make up 62% of non-violent crime
exonerations (317/510), and white-collar crimes, which make up 12% (63/510). We say a bit
more about these categories in the sections that follow.
c. Drug crimes
An exoneration for a violent felony usually begins with a substantial claim that a specific
convicted defendant is innocent. That may lead to a reinvestigation of the crime, which may in
turn lead to exoneration—and sometimes uncover misconduct along the way.
Two-thirds of drug-crime exonerations, however, were not the result of distinct investigations of
individual cases. Instead, they occurred in two very different clusters of cases, 65 exonerations
in Cook County, Illinois (Chicago), and 149 in Harris County, Texas (Houston). The
exonerations in each cluster attracted attention because of common features that made it
comparatively easy to obtain exonerations. The common features, however, were very different
from one county to the other.
The overall rate of misconduct in drug-crime exonerations is 39% (123/317), but that average
masks a stark disparity between these clusters. In Cook County, Illinois, 100% of drug-crime
exonerations included official misconduct (73/73), while in Harris County, Texas, only 3% of
drug crime exonerations included misconduct (4/149). Among drug crime exonerations from
the rest of the country, the rate of official misconduct was 49% (46/94.)
36
i. Group exonerations: Cook County, Illinois
On August 4, 2006, 28-year-old George Almond was visiting a friend at a public
housing development in Chicago, Illinois, when Police Sergeant Ronald Watts
and officers under his command raided the building looking for drugs. Watts had
a reputation for planting drugs on people, so Almond remained in the apartment
until he thought the police had left. When he walked into the hallway, however,
Watts was there.
Watts ordered Almond to stop and asked him for money. When Almond said he
only had a few dollars, Watts arrested him for possession of narcotics. At the
police station, when Almond asked what he was being charged with, an officer
placed 20 baggies of heroin on the table and Watts said the drugs were Almond’s.
35
These cases also include filing a false report, solicitation, and traffic offenses.
36
We also know of seven drug crime exonerations in Harris and Cook counties that are not part of the clusters we
describe below.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 22 • National Registry of Exonerations • September 1, 2020
On October 5, 2006, Almond pled guilty in Cook County Circuit Court to
possession of a controlled substance. He was sentenced to two years in prison
and released 16 months later, in February 2008.
Normally, that would be it. It’s basically impossible to win a case that pits the word of an
accused drug dealer against that of two or three police officers—even if somebody cared enough
to fight that battle on behalf of a drug defendant who pled guilty and was released after 16
months in prison.
Four years later, however, Watts and one of his officers were arrested after they were caught on
camera stealing money from a confidential FBI informant posing as a drug courier. Federal
prosecutors said that Watts “used his badge and his position as a sergeant with the Chicago
Police Department [CPD] to shield his own criminal activity from law enforcement scrutiny. He
recruited another CPD officer into his crimes, stealing drug money and extorting protection
payments from the drug dealers who terrorized the community that he… had sworn to
protect.”
37
In 2013, Watts pled guilty to taking money from the fake drug courier; he was
sentenced to 22 months in federal prison.
Watts’s conviction ended his police career and prevented him from framing new defendants, but
it did not, on its own, affect George Almond’s conviction. That process depended on another of
Sergeant Watts’s victims, Ben Baker. Baker was arrested twice by Watts and his team (the first
time charges were dismissed)—and then a third time, with his wife, after he filed a complaint
against the police for planting drugs on him. Baker went to trial and was sentenced to 18 years,
after which he pled guilty to another four years in a deal that kept his wife out of prison.
In early 2016, after years of litigation on his behalf by the Exoneration Project at the University
of Chicago Law School, Baker was exonerated from both convictions—which led to claims by
many other defendants that Watts and officers working under him planted drugs on them too
when they refused or were unable to pay the extortion money the officers demanded.
Reinvestigations by the Exoneration Project and other civil rights and criminal defense
attorneys, and by the Conviction Integrity Unit of the Cook County State’s Attorney’s Office, led
to two additional Sergeant Watts exonerations in 2016, 17 exonerations in 2017, 31 in 2018
(including one defendant who was framed for possession of an illegal gun rather than drugs),
and 14 in early 2019—a total of 66 through February 2019.
38
George Almond himself was
exonerated in November 2018, more than ten years after he was released from prison.
37
United States vs. Ronald Watts, United States District Court Northern District of Illinois Eastern Division,
Government’s Response to Defendant Ronald Watts’ Sentencing Memorandum, October 2, 2013; PACER document
Case: 1:12-cr-00087 Document #: 84 Filed: 10/02/13 Page 1 of 12 Page ID #:431.
38
As of this writing, 11 additional Watts exonerations have been added to the Registry among the 261 exonerations
since the initial 2,400, for a total of 77.
We also know of six Cook County drug exonerations that occurred before 2016. All included official misconduct, but
none involved Sergeant Watts.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 23 • National Registry of Exonerations • September 1, 2020
What Watts and his crew did in Chicago is hardly unique. It’s one of 17 collections of cases
across the country, from 1990 through 2018, that we have described as “group exonerations.”
39
Exonerations, of course, can be grouped in any number of ways, but what we mean by a “group
exoneration” in this context is very specific: The exoneration of a group of defendants who were
deliberately framed for crimes that did not occur as part of a large-scale pattern of police
perjury and corruption.
In almost all the group exonerations of that sort, police officers framed innocent defendants by
the same method used by Sergeant Watts and his men in Chicago: they planted drugs on them,
or simply booked the drugs in evidence and said they were found in searches of the defendants,
their cars or their belongings.
Those frame-ups produced false convictions for comparatively minor crimes. Most defendants
were sentenced to probation, or to several months, or one or two years in custody. It’s
prohibitively expensive to try to establish innocence in such cases; it basically never happens—
unless some feature of the case makes the process quick and easy. For group exonerations, the
common practice of misconduct by an identified group of police officers fills that function: once
that’s been established, proving innocence in a particular case is relatively easy.
Most of the exonerees in these group exonerations are not listed in the Registry because we
know too little about their cases to describe them individually, or to be confident that the
exonerations reflect a high probability of innocence.
40
In some of those group exonerations, after the underlying pattern of corruption and perjury
came to light, specific exonerations were handled summarily and received little or no separate
attention. As a result, we often don’t know basic facts about individual exonerations: not the
dates of arrest, conviction, and exoneration; not the facts of the alleged crimes; not the mode of
conviction or the sentence; not the evidence of innocence that led to the exonerations;
sometimes not even the names of the exonerated defendants. In short, we have too little
information on many defendants in group exonerations to include them in our database of
individual exonerations.
Because of these limitations, we decided in 2012 not to include group exonerations in the
Registry, although we did discuss the ones we knew about in our reports.
41
As of 2018, we knew
of 17 such groups across the country, involving more than 2,500 exonerations, from 1990
through the present. The defendants were overwhelmingly black or Hispanic, and they were
almost always framed for drug possession or distribution.
42
39
See Nat’l Registry Of Exonerations, Mass Exonerations and Group Exonerations Since 1989 (Apr. 9, 2018).
40
See generally Russell Covey, Police Misconduct as a Cause of Wrongful Convictions, 90 Wash. U. L. Rev. 1133
(2013).
41
Samuel R. Gross, et.al., Race and Wrongful Convictions in the United States, 20 (Mar. 7, 2017); see also Nat’l
Registry of Exonerations, Mass Exonerations and Group Exonerations Since 1989 (Apr. 9, 2018).
42
See The Nat’l Registry Of Exonerations, Mass Exonerations and Group Exonerations Since 1989 (Apr. 9, 2018).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 24 • National Registry of Exonerations • September 1, 2020
In the past two years, we have focused greater attention on the group exonerations. We plan to
release a great deal of new information on them in the near future. We are expanding our focus
to include some groups of convictions that were dismissed because of different types of
systematic official misconduct.
In the process, we also decided that the Registry should include those individual cases within
group exonerations for which we have sufficient information to know that the exoneration was
the result of an individual post-conviction re-examination of the evidence in the case. We have
started to re-examine group exonerations in light of this decision; so far, the cases we have
considered and included are the 66 group Watts exonerations from Cook County.
43
Needless to say, all the exonerations in the Watts group, past and future, are for false
convictions that were obtained by criminal misconduct by police officers—by definition. So too
are thousands of exonerations, in similar groups, that have not made it onto the Registry, at
least not so far; like the Watts group, they are overwhelmingly drug crime cases.
The drug frame-up group exonerations are an important illustration of a general issue: even
outrageous and locally notorious misconduct in comparatively low-level cases will rarely come
to our attention because few defendants in such cases are ever exonerated. What Sergeant Watts
and his men were up to was well known in the neighborhoods in which they operated. In fact,
other officers had reported it to their superiors years before Watts was arrested by the FBI; they
were ignored and assigned to desk duty as punishment.
44
It would have been much harder,
perhaps impossible, to conceal an equally flagrant pattern of lying and planting evidence in
murder or sexual assault cases.
In drug cases, many defendants who are framed just suck it up, plead guilty, do their time and
try to move on. If the misconduct was part of a major scandal, maybe it will be discovered and, if
so, maybe someday they will be exonerated. But maybe not. The first Sergeant Watts
exoneration came nearly four years after he was arrested, and only because of the unusual
persistence of a single defendant—Ben Baker—who managed to get help from a major non-profit
innocence organization. Most cases will fall through that sort of sieve.
ii. Guilty pleas: Harris County, Texas
Since the beginning of 2006, 149 defendants have been exonerated in Harris County, Texas
(home to Houston), after pleading guilty to drug possession. In all of these cases, the defendants
pled guilty before the supposed drugs they possessed were tested in a crime lab, and were
exonerated weeks, months or years later after testing was done and no illegal drugs were found.
43
One of the 66 Watts group exonerations, that of Anthony McDaniels, was for illegal possession of a gun rather than
for possession or distribution of illegal drugs.
44
See infra Section XII.2.c.i(a); see also Jason Meisner, “Former Chicago Police Officer Sentenced for Stealing
Money from Drug Courier,” CHI. TRIB., Oct. 9, 2013.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 25 • National Registry of Exonerations • September 1, 2020
Why did these defendants plead guilty even though they possessed no controlled substances?
Some may have had powders or pills or smoking materials that they thought were illegal drugs
but were not. But as far as we can tell, most pled guilty to get out of jail.
In a typical case, the defendant had a criminal record and could not post the comparatively high
bail that was set for him. If he pled not guilty, he’d remain in jail for months before trial, and
then risk years in prison if convicted. It’s hardly surprising that an innocent defendant in that
situation would accept a deal to plead guilty and go home immediately or in a few days or weeks.
The only reason we know about these false guilty pleas is that the Harris County crime lab tests
materials seized from defendants even if they have already pled guilty. Few crimes labs do that,
which means that in many routine drug cases testing is never done, since more than 95% of drug
possession convictions are based on guilty pleas, many within days of arrest. Instead, dozens if
not hundreds of thousands of arrests each year are based on cheap and notoriously error-prone,
on-the-spot “presumptive” field tests for drugs—tests that are inadmissible in court but
sufficient for arrest—and nothing more is done before the inevitable guilty pleas.
45
Even in Harris County, where post-plea lab tests were run, they had little effect for years.
Sometimes the tests weren’t done until long after the defendants had served their sentences.
Often the paperwork notifying the prosecutor of the results was lost, misplaced or ignored. In
early 2014, the Harris County District Attorney’s Conviction Integrity Unit (CIU) took charge of
the process and realized the magnitude of the problem. It took them four years to clear the
backlog of false convictions stretching back to 2004; by then they had added 135 drug guilty plea
exonerations.
These exonerations did not cover all the innocent defendants who deserved to have their
convictions dismissed. As of March 2019, the Harris County DA's office had identified an
additional 150 defendants who pled guilty to drug crimes and who would have been exonerated
along with the rest because lab tests showed the suspected drugs contained no controlled
substances—but those defendants were not located, so no court action was taken.
46
The Harris County drug-possession guilty-plea exonerations are similar to the Sergeant Watts
group exoneration in Cook County in one respect: a common set of facts made proof of the
defendants’ innocence easy. In Harris County, very easy: once a cheap, routine lab test found no
illegal drugs and the results were made known to those concerned, innocence was a foregone
conclusion. On the other hand, unlike the Cook County cases, nothing about the common factual
basis for exoneration in Harris County suggests that the false convictions involved any sort of
official misconduct. Only 2% of the Harris County drug plea exonerations since 2014 included
any known misconduct—three cases in which the authorities (probably inadvertently) failed to
notify defendants that exculpatory test results had been obtained before they pled guilty.
45
Ryan Gabrielson & Topher Sanders, How a $2 Roadside Drug Test Sends Innocent People to Jail, N.Y. TIMES
MAG. (July 7, 2016). See also Ryan Gabrielson and Topher Sanders, Busted, Propublica (July 7, 2016).
46
Email from Harris County District Attorney’s Office to Maurice Possley, March 7, 2019.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 26 • National Registry of Exonerations • September 1, 2020
There may have been misconduct that we don’t know about in some of the Harris County drug
exonerations. Two possible types, however, are racial or ethnic discrimination in the decision to
stop, search or arrest suspects—“racial profiling”—and illegal searches that uncovered the
supposed drugs. Neither type counts as misconduct that contributes to a factually false
conviction, the type of misconduct that we are concerned with in this study.
47
As far as we know,
however, official misconduct was not a significant factor in these 149 Harris County cases—and
probably not in the many thousands of false convictions of other innocent drug possession
defendants who plead guilty without lab tests in the rest of the country, where post-plea lab tests
almost never happen.
d. White-collar Crimes
White-collar exonerations are the polar opposite of the clusters of drug cases we just discussed.
There are comparatively few cases—less than 3% of all known exonerations (63/2,400)—but,
with few exceptions, each was a big deal, both to prosecute and to obtain an exoneration.
In many respects, white-collar cases are more similar to murder prosecutions than to other non-
violent crimes. The crimes, of course, are extremely different from murder, and the penalties are
far lighter, but the two types of cases are at least equally complex and highly important. Perhaps
as a result, the rate of official misconduct among white-collar exonerations is second only to
murder at 62% (39/63).
White-collar exonerations are also unusual in another respect: nearly three quarters were based
on convictions in federal courts (46/63), compared to 3% of all other exonerations (66/2337).
As a result, we discuss white-collar cases in more detail in Section IX, which is devoted to official
misconduct in federal exonerations.
e. Misdemeanors
Only 4% of exonerations in the Registry are for misdemeanors (89/2,400), which make up at
least 80% of all criminal convictions in the United States, a vast under-representation. Two-
thirds of misdemeanor exonerations come from among the 149 drug possession guilty-plea
exonerations in Harris County, Texas (58/89). None of them include known official misconduct,
for reasons we just discussed.
The remaining 31 exonerations were drawn from all misdemeanor convictions in the country
other than drug possession cases in a single county. This is a tiny number, just over 1% of all
known exonerations. Clearly, only an infinitesimal fraction of false misdemeanor convictions are
ever pursued to the point of exoneration—unless routine but unexpected laboratory tests happen
to show up and conclusively prove innocence.
47
It’s possible that in some of the cases that we do count, where the materials seized included no controlled
substances, police officers lied about the quantities of the substances they seized, which might have contributed to
the defendants’ decisions to plead guilty and hence their convictions. It’s also possible, as we mentioned above,
Section III.3.c.ii, that some officers lied in their reports about evidence that the defendants knowingly possessed the
supposed drugs that were seized.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 27 • National Registry of Exonerations • September 1, 2020
On the other hand, 58% of these few remaining misdemeanor exonerations do include known
misconduct, a higher rate of misconduct than for most types of non-homicidal felonies (18/31).
Here’s one:
On September 5, 2013, police in Perth Amboy, New Jersey were called to the
home of 18-year-old Edwin Rodriguez by neighbors who complained that
someone was riding a mini-motorcycle up and down the street. What happened
next was recorded in two videos.
A police officer asked Rodriguez for identification. Rodriguez asked the officer to
step outside so he could close the door while he got his ID; the officer refused.
Rodriguez complained about that and began to walk up a flight of stairs; the
officer tackled him and slammed him onto the steps; other officers joined in and
helped subdue Rodriguez, who was pepper-sprayed, handcuffed and dragged out
of the house by his ankles.
Rodriguez suffered a broken clavicle. He was charged with two counts of
obstruction, resisting arrest, disorderly conduct and possession of a kitchen knife,
and convicted on the last two of those counts. Rodriguez was exonerated at a
retrial six months later because the judge at the first trial had excluded a video
that showed, among other things, that Rodriguez was not holding a knife, and
was attacked as he was trying to obtain his identification.
The few cases we have suggest that serious official misconduct—that sort we see in Edwin
Rodriguez’s case, or Wassillie Gregory’s
48
—helps generate the attention and outrage that, on
rare occasions, can lead to the exoneration of an innocent misdemeanor defendant.
49
4. RACIAL PATTERNS
Official misconduct is more common among exonerations with Black and Hispanic defendants
than those with white defendants, but only by several percentage points, as we have noted. That
racial gap, however, is considerably larger for murder and drug convictions, as we see in Table 6.
48
See supra Section I.
49
See Samuel R. Gross, Errors in Misdemeanor Adjudication, 98 B.U. L. REV. 999 (2018).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 28 • National Registry of Exonerations • September 1, 2020
Table 6: Proportions of Exonerations with Official Misconduct by Race of Exoneree
White
Exonerees
Black
Exonerees
Black & Hispanic
Exonerees
ALL
EXONEREES
Murder (908)
64%
78%
77%
72%
Death Sentences (121)
68%
87%
87%
78%
Other Murder
Convictions (787)
63%
77%
75%
71%
Drug Crimes (317)
22%
47%
44%
39%
ALL CRIMES (2,400)
52%
57%
56%
54%
Blacks are about 13% of the population of the United States, but account for 48% of all known
exonerations (1,158/2,400), 52% of murder exonerations (468/908) and 63% of drug crime
exonerations (200/317) (and an even higher proportion of group exonerations based on drug
crime frame-ups).
50
There is no doubt that race plays a role in the conviction of innocent
defendants in America, an issue we explored in detail in a report in 2017.
51
One of the mechanisms that leads to the disproportionate conviction of innocent Black
defendants is official misconduct. For murder exonerations, 78% of exonerated Black
defendants were the victims of official misconduct, compared to 64% of white defendants;
among innocent murder defendants who were sentenced to death, that gap is 87% for African
Americans and 68% for whites. For drug crime exonerations, the difference is even starker.
Black drug crime exonerees were more than twice as likely as whites to have been convicted in
whole or in part by official misconduct, 47% compared to 22%.
At least three comparatively small sets of exonerations with high rates of misconduct are racially
homogeneous, or nearly so: Ninety-three percent of the child sex abuse exonerations that
followed the child sex abuse hysteria (CSH) epidemic in the 1980s and 1990s
52
included official
misconduct (54/58); 81% of the exonerees in those cases were white (47/58) and only 3% were
Black (2/58). By contrast, all of the 66 drug exonerees in the one group exoneration included in
the Registry so far, the Sergeant Watts exonerations in Chicago, were Black
53
—and all were the
victims of official misconduct. And in a very different context, the overall rate of official
50
See supra Section III.3.c.i.
51
See Samuel Gross et al, Nat’l Registry of Exonerations, Race and Wrongful Convictions in the United States
(2017).
52
See infra , Section IV.4.
53
See supra Section III.3.c.i. As we’ve mentioned, one of the exonerees in the Watts group, Anthony McDaniels, was
convicted for illegal possession of a gun rather than for possession or distribution of illegal drugs.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 29 • National Registry of Exonerations • September 1, 2020
misconduct among white-collar crime exonerations is 62% (39/63), second only to murder; 81%
of white-collar crime exonerees were white (51/63) and only 8% were Black (5/63).
54
As used in Table 6—and in the Registry, and throughout this report—the categories “White” and
“Black” do not include individuals who identify themselves as Hispanic or Latino. We do not
discuss patterns in exonerations of Hispanic or Latino defendants (except in passing) because
the data at our disposal are unreliable. We study convictions that occurred from the late 1960s
through 2018, and use records that employ inconsistent and sometimes ambiguous definitions
of this ethnic group, or ignore it entirely.
55
The main problem with the data at our disposal is
that we can’t tell whether Hispanic defendants with African ancestry were classified as Hispanic
or as Black. We can circumvent that problem by combining these two minority groups, as we do
in Table 6 under the heading “Black and Hispanic Exonerees.” The rates of official misconduct
for the combined group are similar to those for Black exonerees alone.
5. CATEGORIES OF MISCONDUCT
We found many types of misconduct across these cases (and many cases with more than one
type) but the great majority fall into five general categories. They are, in an order that roughly
corresponds to the chronological order of criminal prosecution: Witness Tampering, Misconduct
in Interrogations, Fabricating Evidence, Concealing Exculpatory Evidence, and Misconduct at
Trial.
In the sections that follow, we discuss these forms of misconduct in that order. The remainder of
this section is a quick overview, starting with Table 7, which displays the proportions of cases
with each sort of misconduct, by crime. Murder cases, as we already saw, have the highest
overall rate of misconduct, and also the highest rate for each subcategory, except the
comparatively uncommon category of Fabricating Evidence.
54
Racial differences in the rates of official misconduct are smaller for the other crimes that appear most frequently
among known exonerations: Sexual Assault—Black exonerees 38% (72/191), white exonerees 44% (46/105);
Robbery—Black exonerees 39% (29/74), white exonerees 36% (10/28); “Other Violent Crimes”—Black exonerees
55% (55/100), white exonerees 52% (64/122); Child Sex Abuse exonerations excluding CHS cases—Black
exonerees 29% (20/68), white exonerees 35% (43/122).
55
See Samuel Gross et al, Nat’l Registry of Exonerations, Race and Wrongful Convictions in the United States
(2017), at p. 1.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 30 • National Registry of Exonerations • September 1, 2020
Table 7: Proportions of Exonerations with Misconduct by Category and by Crime
Witness
Tampering
Misconduct in
Interrogations
Fabricating
Evidence
(estimates)*
Concealing
Exculpatory
Evidence
Misconduct
at Trial
ALL OFFICIAL
MISCONDUCT
Murder (908)
23%
14%
5%
61%
32%
72%
Child Sex
Abuse (270)
28%
3%
3%
27%
14%
44%
Sexual
Assault (320)
12%
4%
9%
32%
17%
39%
Robbery (122)
11%
1%
0
28%
14%
38%
Other Violent
Crimes (270)
18%
5%
13%
43%
26%
55%
Drug Crimes
(317)
1%
1%
28%
37%
13%
39%
White-collar
Crimes (63)
14%
0
0
48%
30%
62%
Other Non-
Violent
Crimes (130)
5%
0
4%
28%
15%
32%
ALL CRIMES
(2,400)
17%
7%
10%
44%
23%
54%
* Unlike the types of misconduct reported in the other columns of this table, the percentages of cases with Fabricated
Official Evidence are estimates combining information from two separate analyses of the Registry data: our analysis
of all 2,400 exonerations covered in this report, and more detailed information from a study of the first 1,361 cases
posted among those 2,400. See the Methodological Appendix for a full description of how these estimates were
obtained.
a. Witness Tampering
In 17% of the prosecutions that ultimately led to exoneration, government officials improperly
induced witnesses to provide false evidence, or to withhold accurate evidence (409/2,400).
Usually this was done by police officers, but prosecutors and child welfare workers participated
in a substantial minority of the cases. There are many forms of witness tampering, but they all
fall into two basic types of misconduct:
Threats. In 5% of exonerations, government officials used threats to coerce witnesses to
change their testimony (130/2,400). Witnesses were told that if they did not cooperate,
they or close relatives would be charged with crimes, sentenced to prison or to death,
lose custody of their children, and so forth. The witnesses who succumbed to these
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 31 • National Registry of Exonerations • September 1, 2020
threats all knew that they were deceiving the fact finders, either by lying or by
deliberately withholding evidence favorable to the defense.
Manipulation. In 13% of exonerations, officials distorted witness testimony without
using threats (305/2,400). Some witnesses were bribed to change their testimony by
promises of lenience in their own criminal cases, or received other benefits: privileges in
jail, release from custody, drugs, money, etc. Others were deceived by false evidence of
the defendants’ guilt or tricked into thinking they saw things that did not happen. Some
were simply asked to lie (or to keep the truth to themselves) and agreed to do it. Some
witnesses who were manipulated into giving false testimony did not realize it; others did.
Like most categories of official misconduct, witness tampering is most common among
the worst violent crimes that appear frequently in exonerations—murder (23%) and child
sex abuse (28%)—followed by other categories of violent crimes, and is least common
among non-violent crimes.
b. Misconduct in Interrogations
Twelve percent of exonerees falsely confessed to the crimes they were convicted of (292/2,400),
usually under pressure from police. Some types of pressure and deception that would be
misconduct in dealing with ordinary witnesses are permitted in interrogations of criminal
suspects, presumably on the theory that because guilty suspects have obvious and overwhelming
motives to lie about their guilt, police must be allowed—within limits—to scare, trick and deceive
them in order to get the truth. But police sometimes go beyond those limits and commit
misconduct in interrogations. They did so (that we know of) in 57% of exonerations with false
confessions (165/292); three-quarters of those exonerations were murder cases (126/165).
In nearly two-thirds of cases in which police misconduct led to false confessions, the police used
or threatened to use physical violence (105/165). Half of all exonerations in the United States in
which confessions were obtained by violence occurred in Chicago, where violence was used to
obtain more than two-thirds of false confessions in exonerations (52/75); for the rest of the
country, that rate was less than a quarter (53/217).
c. Fabricated Official Evidence
56
In some criminal cases,, police and other law enforcement officers are themselves critical
witnesses to the alleged crimes or the facts surrounding them, or claim to be. In that situation,
officers don’t need to tamper with witnesses in order to present false evidence against
defendants—they are witnesses and can do so directly. We see this sort of misconduct in about
10% of exonerations, in three forms:
• Forensic Fraud. In about 3% of exonerations (75/2,400) state forensic analysts (often
police officers) deliberately presented false forensic evidence of the defendants’ guilt. In
some cases they described tests or observations they never made; in others they claimed
56
See supra Table 7, and infra Methodological Appendix.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 32 • National Registry of Exonerations • September 1, 2020
without justification that the defendants’ hair or blood or semen “matched” items from
the crime; and in some they testified that the defendant could have been or was the
source of biological trace evidence after testing had already shown that was impossible.
• Fake Crimes. In about 4% of exonerations, police officers planted incriminating physical
evidence on suspects or at the scene of the crime—usually drugs—or booked the evidence
and falsely claimed it was found on the suspects. In about 1% of additional cases, police
officers falsely testified that they themselves had been assaulted by the defendants,
usually to cover up their own violence.
• Fictitious Confessions. In close to 2% of exonerations, police officers testified that
defendants had confessed when they had not (36/2,400).
Fabricating evidence is most common among exonerations for those crimes that are defined by
the types of evidence that are fabricated: (planted) drug crimes, and (fake) criminal assaults.
d. Concealing Exculpatory Evidence
In 44% of exonerations (1,064/2,400), law enforcement officials concealed evidence favorable to
the defendant. This is the most common type of official misconduct we found. It is done by
prosecutors and police officers alike. Two basic categories of exculpatory evidence were
concealed:
• Substantive evidence of innocence. In 30% of exonerations, law enforcement officials
concealed substantive evidence that would have supported the defendants’ claims of
innocence (709/2,400). The hidden evidence included alibi evidence for the defendant,
evidence about alternative suspects (some of whom were later proven to be the real
criminals), forensic evidence that showed that the defendant was not the source of
semen or blood or fingerprints left at the scene of the crime, and so forth.
• Impeachment evidence. In an overlapping third of the cases, police and prosecutors
concealed evidence that would have undercut witnesses who testified to the defendants’
guilt (805/2,400). They hid statements in which prosecution witnesses said the opposite
of what they testified to in court, attempts by those witnesses to retract their accusations
or testify that the defendants were innocent, known histories of deception and crime by
prosecution witnesses, money or favors received by the witnesses or deals that saved
them years in prison in return for nailing the defendants, and so on.
57
Concealing exculpatory evidence is widespread across the board. It occurred in 27% to 48% of
exonerations for every category of non-homicidal crime. But it is uniquely prevalent in
exonerations for murder, where it occurred in 61% of all cases.
57
Nineteen percent of the cases included both types of concealed exculpatory evidence (450/2,400).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 33 • National Registry of Exonerations • September 1, 2020
e. Misconduct at Trial
Criminal convictions in the United States are overwhelmingly obtained by guilty pleas, but 80%
of exonerated defendants were convicted at trial (1,928/2,400). About 28% of those trials
(546/1928)—23% of all exonerations (546/2,400)—included official misconduct in court, in the
course of the trial.
• Police Perjury. Police and other law enforcement officers testify in almost all criminal
trials, and sometimes they lie. That happens—that we know of—in more than 14% of
exonerations after conviction at trial, or 13% of all exonerations.
Some of those cases were discussed above under Fabricating Official Evidence; they
involved forensic fraud or officers who falsely claimed to have witnessed or been
victimized by crimes by the exonerees. However, in three quarters of trials with police
perjury, officers lied about the conduct of the investigations, or about statements by
other witnesses rather than their own observations.
• Trial misconduct by prosecutors. Prosecutors committed misconduct at trial in more
than 14% of all exonerations (334/2,400). The great majority of this courtroom
misconduct fell into three categories:
58
o Knowingly permitting perjury. In 8% percent of exonerations, prosecutors
knowingly permitted witnesses to commit perjury at trial without notifying the
court (186/2,400).
o Lying in court. In about 4% of exonerations a prosecutor lied to the court about
the facts of the crime the defendant was charged with, or about the investigation
and prosecution of the case (94/2,400)—usually in closing argument.
o Improper statements in closing argument or cross-examination. In more than
3% of exonerations prosecutors made improper, often unconstitutional closing
arguments to the juries or judges that convicted the defendants (77/2,400)—not
counting the cases in which prosecutors lied directly in argument, which are
covered by lying in court. The most common improper arguments (other than
outright lies) were factual claims that had no basis in the evidence presented, and
arguments that a defendant who did not testify in her own defense must be guilty
because she did not deny it in court. In about 1% of cases (that we know of)
prosecutors made equally improper statements in the guise of questions on cross-
examination (26/2,400).
58
More than a third of cases with official misconduct at trial involved multiple forms of misconduct (121/334).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 34 • National Registry of Exonerations • September 1, 2020
IV. Witness Tampering
Police and prosecutors do many things that shape and alter witness testimony. Most are benign
or even necessary. They ask questions, remind witnesses of past events, present them with
documents and pictures, and help them prepare to testify in court. In the process, they
sometimes distort the witness’s testimony. This can have disastrous results—a suggestive lineup,
for example, can lead to an eyewitness misidentification and decades in prison—but mere
suggestiveness is not misconduct, not even if the investigatory practice is incompetent.
Witness tampering—deliberate and successful efforts to get witnesses to give false evidence or
to withhold true evidence—is misconduct that contributes to wrongful convictions. It was
witness tampering for a police officer in Juan Johnson’s case to threaten to charge an eyewitness
with murder if he refused to identify Johnson at his 1991 murder trial in Chicago. It was also
witness tampering when an officer, preparing for Debra Brown’s murder trial in Utah in 1995,
talked to a witness who saw a man shortly after the killing with a gun and lot of cash, and told
that witness to keep quiet “if he knew what was good for him.”
It’s not necessary for the officer to know that the witness is testifying falsely. It’s misconduct to
trick, persuade or force a witness to testify to something that she might have seen, if the officer
does not care whether she actually did. For example, in Sami Leka’s case, police officers
pressured a reluctant couple to identify Leka in a lineup, telling both that they knew Leka was
the killer, and telling the wife that her husband had already identified Leka. It’s also misconduct
to use irresponsible, coercive techniques to obtain testimony, even if the official believes the
crucial facts are true. For example, in Ricky Lynn Pitts’s case, a prosecutor (who probably
believed Pitts was guilty) told a young girl that she would never see her parents again unless she
“admitted” that Pitts had sexually assaulted her.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 35 • National Registry of Exonerations • September 1, 2020
1. GENERAL PATTERNS IN WITNESS TAMPERING
As we’ve noted,
59
17% of the exonerations in the Registry include witness tampering
(409/2,400), by threats (5%) or manipulation (13%) or occasionally both (1%). Table 8 displays
the rates of witness tampering by crime and category—threats or manipulation.
Table 8: Witness Tampering by Category and Crime*
Murder
(908)
Sexual
Assault
(320)
Child Sex
Abuse
(270)
Other
Crimes
(902)
ALL
CRIMES
(2,400)
Threats
10%
1%
9%
1%
5%
Manipulation
16%
11%
20%
8%
13%
ALL WITNESS
TAMPERING
23%
12%
28%
9%
17%
* Some cases include both witness tampering with and without threats.
Witness tampering by the authorities is unevenly distributed across crimes. The rates are
highest for murder, 23%, and child sex abuse, 28%, compared to 12% for sexual assaults and 9%
for all other crimes. Almost 80% of all exonerations with witness tampering are murder, sexual
assault or child sex abuse cases (325/409). Witness tampering by threats is even more
concentrated. More than 89% of exonerations with this form of tampering are for murder or
child sex abuse (116/130), which may reflect heavy commitments of time and attention to
alleged victims of child sex abuse, and to witnesses of all sorts in murder investigations.
In Table 9, we see that witness tampering is primarily a form of police misconduct. For this table
and Table 10, we rely on data from a detailed study of official misconduct among the 1,361
exonerations posted on the Registry by May 16, 2014, that includes information on the positions
of the officials who committed each type of misconduct, data that are not available for the entire
2,400 case database.
60
Police officers tampered with witnesses in 80% of the exonerations in
which that happened (206/259), whether by threats or by manipulation—75% more than the
total for prosecutors and child welfare workers combined. This is no surprise; police officers are
the main investigators who find and interview witnesses, and take statements from them.
59
Supra Section III.5.a.
60
See supra Section II.2.a.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 36 • National Registry of Exonerations • September 1, 2020
Table 9: Witness Tampering by Category and Responsible Party* (n=1,361)
Police Officer
Prosecutor
Child Welfare Worker
ANY STATE OFFICIAL
Threats
5% (67)
2% (27)
1% (14)
6% (80)
Manipulation
11% (144)
4% (57)
2% (23)
14% (189)
ALL WITNESS
TAMPERING
15% (206)
6% (80)
3% (37)
19% (259)
* Some cases include misconduct by more than one type of official.
We have identified three common types of witness tampering, one or more of which occurred in
78% of exonerations that include this category of misconduct (319/409). All three can be
accomplished by threats or by manipulation or both:
• Procuring false testimony is the practice of inducing a civilian witness to testify to facts
the officer or prosecutor knows the witness did not perceive.
• A tainted identification occurs when police deliberately induce a witness to identify a
suspect during a lineup or other identification procedure, whether the witness recognizes
that suspect or not.
• Improper questioning of a child victim is repeated, insistent and suggestive questioning
of a child by a government official who will not allow the child to deny that he or she was
a victim of sex abuse.
We discuss these types of tampering in more detail in the sections that follow—after a quick
review of a couple of general patterns:
Table 10: Proportions of Cases with Witness Tampering by Responsible Party* (N=1,361)
Police
Officer
Prosecutor
Child Welfare
Worker
ANY STATE
OFFICIAL
Procuring False
Testimony
3% (41)
3% (41)
0% (0)
6% (77)
Tainted
Identification
7% (91)
0.1% (2)
0% (0)
7% (91)
Improper
Questioning
of a Child Victim
3% (42)
1% (16)
3% (37)
4% (50)
* Some cases include misconduct by more than one type of official.
Table 10 displays the frequency of each type of witness tampering by the type of government
official who did it, and shows how police led the field:
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 37 • National Registry of Exonerations • September 1, 2020
• Police and prosecutors procured false testimony in equal numbers of cases, but child
welfare workers never did so.
• Police and child welfare workers engaged in improper questioning of children at about
the same rate, but prosecutors did so much less frequently.
• Police conducted all tainted identifications, with prosecutors also participating in two of
them.
Table 11 shows that these subspecies of witness tampering are unevenly distributed across
categories of crime. Procuring false testimony was far more common in exonerations for murder
than for lesser crimes; abusive questioning of children occurred exclusively in child sex abuse
exonerations; and tainted identifications were most common in exonerations for sexual assault
followed closely by murder exonerations.
Table 11: Proportions of Cases with Witness Tampering by Category and Crime
Murder
(908)
Sexual
Assault
(320)
Child Sex
Abuse
(270)
Other
Crimes
(902)
ALL CRIMES
(2,400)
Procuring False
Testimony
11%
1%
1%
5%
6%
Tainted Identification
9%
9%
1%
4%
6%
Improper Questioning
of a Child Victim
0%
0%
24%
0%
3%
ALL WITNESS
TAMPERING
23%
12%
28%
9%
17%
2. PROCURING FALSE TESTIMONY
Procuring perjury is an extreme form of misconduct. It only occurs in about 6% of all
exonerations; two-thirds of them are murder cases—the most serious cases, where the pressure
is greatest and the stakes highest. The Randall Dale Adams exoneration is a clear example:
Adams’s car ran out of gas on November 27, 1976. He was walking along the side
of the road when 16-year-old David Harris offered him a ride. Later that night
(when Adams was no longer with him), Harris shot and killed a police officer.
Once Harris became a suspect, he claimed that Adams was driving at the time
and shot the officer with Harris’s gun, while Harris himself was in the front
passenger seat.
There was no evidence connecting Adams to the shooting other than Harris’s
testimony, and Harris had been a suspect himself, so the police needed credible
eyewitness testimony to make the case. One eyewitness described the shooter as
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 38 • National Registry of Exonerations • September 1, 2020
Black or Mexican, and failed to identify Adams (who was white) in a lineup. The
police offered to drop pending robbery charges against the witness’s daughter
if she would change her statement and make the identification. She agreed. Her
original description and her failure to identify Adams were concealed.
At trial, that witness testified that she identified Adams unassisted, a claim that
the prosecution knew was a lie. Adams was convicted and sentenced to death in
1977.
Almost ten years later a filmmaker who was making a documentary about Adams'
case discovered the hidden history of this eyewitness identification, and learned
about Harris’s subsequent criminal history. Among other crimes, Harris had
committed another murder for which he too had been sentenced to death.
At a hearing in 1988, Harris recanted his testimony. Adams was exonerated in
1989. Harris was executed in 2004.
In some cases, the process of procuring false testimony is simpler but no less damaging. In 2012,
for example, the Texas Court of Criminal Appeals concluded that Richard Miles had proven his
innocence of the murder for which he was convicted in Dallas in 1995. Part of the basis for that
decision was new evidence that before Miles’s trial, the key eyewitness “was unable to identify
the shooter, and when he told the State of this fact, the prosecutor directed him to identify
[Miles] by showing him where [Miles] would be seated in the courtroom.”
61
And in Clarence
Brandley’s case in Montgomery County, Texas, in 1981, the officer investigating the murder of a
high school student told the two custodians who found the body, “One of you is going to have to
hang for this” and, turning to Brandley, added, “Since you’re the nigger, you’re elected.” To
make that promise stick, that officer later told a witness who saw another suspect near the scene
of the crime that he would be arrested if he said anything inconsistent with Brandley’s guilt.
In other cases, the perjury is the culmination of a long process of abuse and corruption.
In 1968 Peter Limone, Louis Greco and Henry Tameleo were convicted of
murdering Edward Deegan, a small-time criminal, three years earlier in an
organized crime hit in Boston. Their conviction was based on testimony from a
gangster named Joseph “The Animal” Barboza, who testified that he participated
in the murder, and that it was ordered by Limone, approved by Tameleo and
carried out by Greco. All three were sentenced to death, later reduced to life
imprisonment. A fourth defendant, Joseph Salvati, was convicted as an accessory
and sentenced to life.
In 2000, a special prosecutor came across numerous documents that painted the
real picture: Barboza was an FBI informant in 1965; FBI agents knew that he
planned to kill Deegan and allowed the murder to take place in order to protect
his position as an informant; the FBI knew that Greco, Limone, Tameleo or
61
Ex Parte Miles, 359 S.W.3d 647, 661 (Tex. Crim. App. 2012).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 39 • National Registry of Exonerations • September 1, 2020
Salvati had nothing to do with that murder; nonetheless, the FBI used Barboza’s
perjury to convict Limone, Greco, Tameleo and Salvati. All four were exonerated
in 2001; by then Greco and Tameleo had died in prison.
3. TAINTED IDENTIFICATIONS
Many identification procedures are poorly designed. They can lead to terrible mistakes and
cause false convictions, but these are not usually “tainted” identifications. In Timothy Cole’s
rape investigation in Lubbock, Texas in 1985, for example, police showed the victim six color
photographs: the five “fillers” were standard police mugshots, all in profile—but no mugshot
was available for Cole, who had never been arrested, so an officer took a Polaroid picture of him,
facing the camera.
Suggestive identification procedures like Cole’s are common.
62
They are bad police practice, and
(as a wealth of research has shown
63
) they cause false convictions. In this case, the victim picked
out Cole immediately. Cole died in prison 13 years later, 10 years before he was exonerated,
posthumously, by DNA testing and a confession from the actual rapist.
Despite such outcomes, using a suggestive identification procedure is not misconduct unless the
police intentionally structured the identification to induce the witness to identify the exoneree.
We have no evidence of that sort of deliberate behavior in Cole’s case. What the officers did may
have been just as harmful all the same, but it was incompetence not misconduct.
Police (and in a few rare cases, prosecutors) cross the line when they tell witnesses in one form
or another who to identify. Sometimes they do it indirectly:
Thomas Doswell was identified by a rape victim in Pittsburgh who was shown
photographs of eight men, one of which—Thomas’s—had the letter "R" (for
rapist) written on it. He was exonerated by DNA 19 years later.
In 1985, Nathson Fields, a member of the El Rukn street gang in Chicago, was
misidentified by members of a rival gang as one of two men who gunned down
two their comrades the previous year. Fields was the only person in the lineup
wearing short sleeves, which showed an El Rukn tattoo on his forearm.
62
We don’t know how common suggestive procedures are in the investigations of exonerated defendants because in
most cases we have don’t know enough about police initiated identifications. If we had video and audio recordings of
the process, and copies of any pictures shown to victims and witnesses, we might be able to assess the
suggestiveness of the procedures, but usually we don’t. Without that sort of record, we are limited to the cases in
which suggestiveness was noticed and raised in court. Unfortunately, the defense attorneys who might raise the
issue often have no better information than we do, and some are too ineffective to act on any information they do
have.
63
See National Research Council of the National Academy of Sciences, Identifying the Culprit: Assessing Eyewitness
Identification (2014); National Registry of Exonerations, Tainted Identifications (2016); Steven P. Grossman,
Suggestive Identifications: The Supreme Court’s Due Process Test Fails to Meet Its Own Criteria, 11 U. Balt. L. Rev.
53 (1981).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 40 • National Registry of Exonerations • September 1, 2020
In other cases, the police are more straightforward. Michael VonAllmen was identified by a rape
victim in Louisville, Kentucky after a police officer told her that VonAllmen had already been
identified by five or six other rape victims but had not been charged because those women were
too scared to testify. And in Jerry Lee Evans’s case, Dallas police simply told the victim which
picture to choose.
Tainted identifications were used twice as frequently in exonerations with minority defendants
as in those with white defendants: Black defendants, 7% (86/1158); Black or Hispanic
defendants,
64
8% (113/1,439); white defendants, 4% (34/906). This contributes to the high rate
of false sexual assault convictions of innocent Black men who were charged with sexual assaults
on white women.
65
Many victims who are directly told who to identify, and comply, know that they are lying—that
they did not themselves recognize the defendants—but they may also believe that the men they
accuse are, in fact, guilty. On the other hand, when police officers indicate the suspect indirectly,
some eyewitnesses may think they are identifying a person they saw and recognize. Walter
Snyder, for example, was not identified by a rape victim in a photographic lineup, but she did
identify him in a live lineup after Alexandria, Virginia police asked her to come to the police
station, and arranged for her to see Snyder waiting by himself in the lobby.
66
Three-quarters of the tainted identifications that we know about occurred in murder and sexual
assault cases (112/149). The rate of tainted identifications is the same for both crimes—9% (see
Table 11)—but the underlying behavior is entirely different.
In nearly 80% of murder cases with tainted IDs, one or more witnesses deliberately
misidentified the exonerees (65/82), while only half included any witnesses who mistakenly
identified them (41/82). In 35% of those cases, police got witnesses to identify the defendants by
threatening rather than manipulating them (29/82); all but one included deliberate false
identifications of the exonerees. For example:
A year after Charles Wilhite was convicted of murder in Springfield,
Massachusetts in 2010, the critical eyewitness testified that a detective repeatedly
threatened to charge him as an accessory to the murder if he didn’t identify
Wilhite—who was acquitted at a retrial in 2013.
When Mubarez Ahmed was convicted of murder in Detroit in 2001, the critical
eyewitness was Izora Clark. In closing argument, the prosecutor told the jury, “If
you don’t believe Ms. (Clark), you have to find him not guilty.” Fifteen years later,
Clark told defense attorneys and later the prosecutor’s office that she had not
64
See supra, Section III.4, Table 6.
65
See Samuel Gross et al, Nat’l Registry of Exonerations, Race and Wrongful Convictions in the United States
(2017), at pp. 11-12 (“In half of all sexual assault exonerations with eyewitness misidentifications, black men were
convicted of raping white women, a racial combination that appears in less than 11% of sexual assaults in the United
States.”).
66
Barry Scheck, Peter Neufeld & Jim Dwyer, Actual Innocence, 53-100 (2001).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 41 • National Registry of Exonerations • September 1, 2020
seen Ahmed at the shooting, and that she only identified him because she was
intimidated by the police and afraid that if she didn’t do what they said they
would retaliate against her son, who was incarcerated at the time.
In sexual assault exonerations the victims always survive the crimes because if a victim is killed,
the case becomes a murder. Unlike murder prosecutions, the critical eyewitness in a sexual
assault case is almost always the victim. Perhaps for that reason, none of the tainted
identifications in sexual assault exonerations include threats by the police, and all but one
produced mistaken witness identifications (28/29).
Police may trick or persuade some rape victims to identify suspects they did not in fact see, but,
as best we can tell, they don’t force them to lie about whether they were attacked or by whom. In
Garry Diamond’s case, for example, the victim failed to pick Diamond’s picture from a mug book
of police photographs—and when her three-year-old son pointed to him, said the child was
mistaken because Diamond had the wrong hair color, eye color and complexion. She later failed
to identify Diamond from a photograph again. Three months later, however, she did identify
Diamond in person—when a detective brought her to court to see Diamond on trial for an
unrelated sexual assault. Diamond was convicted in 1976 in Virginia, and exonerated by DNA 36
years later.
While we know of no adult sexual assault exonerations in which police or prosecutors
threatened victims to get them to identify the defendants,
67
they have done just that to the
supposed victims in some child sex abuse cases—as we see in the next section.
4. IMPROPER QUESTIONING OF A CHILD VICTIM
In some child sex abuse investigations, police officers, child welfare workers and occasionally
prosecutors have engaged in prolonged, improper and in many cases abusive questioning of the
children who were the supposed victims of the alleged crimes. This happened, that we know of,
in nearly a quarter of child sex abuse exonerations (see Table 11). Nearly 80% of exonerations
with this type of misconduct were part of an epidemic child sex abuse hysteria that swept across
the country from the early 1980s to the late 1990s (49/64). Most of those cases were based on
accusations against child daycare providers and many included allegations of satanic rituals.
Eventually, at least 70 innocent defendants were convicted of child molestation and other
serious felonies, many others were arrested and charged.
68
Many of the accusations the children made in response to this sort of questioning were bizarre if
not impossible on their face. Children at the Little Rascals Day Care Center in Edenton, North
Carolina, said that they had seen babies killed, children taken out on boats and thrown
overboard to feed sharks, and children taken to outer space in a hot air balloon.
69
In Kern
67
In a few exonerations that included sexual assaults as well as other violent crimes, police threatened witnesses
other than the rape victims and forced them to identify the exonerees.
68
See Samuel R. Gross & Michael Shaffer, Exonerations in the United States,1989 – 2012, at 75-78, National
Registry of Exonerations (2012).
69
Innocence Lost, Frontline, Public Broadcasting Service (1997).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 42 • National Registry of Exonerations • September 1, 2020
County, California, children described mass orgies with as many as 14 adults who forced groups
of children to inhale 18-inch lines of cocaine or heroin, gave them injections with syringes that
left large bruises, and hung the children from hooks as the adults repeatedly sodomized them.
70
Needless to say, no physical evidence ever corroborated any of these claims.
And in Wenatchee, Washington,
71
in 1994, police arrested 43 people on charges that they had
molested 60 children 29,726 times over a six-year period. That’s an average of more than 100
sexual assaults a year for each abuser, and nearly 500 apiece for each of the 60 children—but
somehow nobody in a town of 55,000 residents noticed anything at the time.
In other cases, the accusations were merely implausible. Either way, they appear to have been
generated by over-eager police officers and child welfare workers who insisted that the children
tell them that they had been molested, and would not take No for an answer.
The Registry includes 58 defendants who were exonerated in child sex abuse hysteria cases
across the country, 84% of which included improper questioning of children (49/58).
72
All of
these exonerees were convicted between 1984 and 1998 (and all but a handful, by 1995).
And then the epidemic passed. We know of no more recent convictions of that sort in which
claims of innocence are pending. A case from Minnesota that did not result in a conviction
illustrates the change.
73
In 1984, a psychiatrist in Minnetonka, Minnesota, interviewed four-year-old Aubree
LaBois about her parents, Edward and Karri LaBois, who ran a daycare center. The
psychiatrist asked Aubree leading questions about sex abuse at the center and showed
her “anatomically correct” dolls to illustrate the questions; Aubree eventually agreed that
she had been sexually molested by her parents.
At first Edward and Karri LaBois denied allegations of sex abuse at their daycare center.
But when they heard that the authorities were planning to take Aubree away, they took
her and fled. In 2003, 19 years later, they were found in Salt Lake City and arrested—and
then all charges were dismissed within two weeks.
Aubree—by then 23 and a mother herself—denied she had ever been molested, and said
she had been misled and confused by the psychiatrist who interviewed her when she was
four. Her statements from 1984 were now seen as “conflicting” and unreliable, and fresh
70
Debbie Nathan & Michael Snedeker, Satan’s Silence: Ritual Abuse And The Making of a Modern American Witch
Hunt (1995).
71
See here for all 11 Wenatchee exonerations.
72
Dozens of others who had been sentenced to years or decades in prison were released without exonerations,
typically after pleading guilty to lesser crimes.
73
J. Adams and M. Zack, Once fugitives, couple are freed; charges of child sex abuse against the couple dating from
1984 were dropped, Minneapolis Star Tribune, Nov. 26, 2003 at 1A; Matt Canham & Rhonda Hailes Maylett, Life on
the Lam, The Salt Lake Tribune, Dec. 14, 2003, Page A.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 43 • National Registry of Exonerations • September 1, 2020
interviews with other children from the daycare center (now grown up) and with their
parents produced no evidence of abuse. As the Minnetonka police chief explained,
interviews with children were “less sophisticated in 1984” when interrogators often used
leading questions and adult sex-abuse terminology.
The children in these cases were usually questioned by police (45/53) or child welfare workers,
typically social workers (37/53), or both. See Table 10. How did they get the children to make
such false and often fantastical accusations? In 2005, a 30-year-old man who had testified as a
nine year-old in a massive child sex abuse hysteria case that ultimately ended with no
convictions, told the story of his role:
“It was an ordeal. I remember thinking to myself, ‘I'm not going to get out of here unless
I tell them what they want to hear.’ …
“I remember telling [the investigators] nothing happened to me. I remember them
almost giggling and laughing, saying, ‘Oh, we know these things happened to you. Why
don't you just go ahead and tell us? Use these dolls if you're scared.'
“Anytime I would give them an answer that they didn't like, they would ask again and
encourage me to give them the answer they were looking for. It was really obvious what
they wanted. . . .
“Maybe some things did happen. Maybe some kids made up stories about things that
didn't really happen, and eventually started believing they were telling the truth. . . . But
I never forgot I was lying.”
74
We don’t doubt that the officers and social workers who conducted these investigations believed
that the sexual assaults they charged had actually occurred, even if some of the reported facts
were fanciful. They believed they were rooting out terrible hidden crimes. They had been trained
and told that many victims of child sex abuse need to be helped, if not pushed, to admit that
these shameful scary things ever happened to them. They accepted a then-prevalent dogma that
while children often deny the occurrence of sexual assaults that did happen, they never make up
ones that have no basis in fact.
We classify the questioning of a child witness as misconduct if it violates contemporary norms—
even though, as the Minnetonka police chief explained, practice has changed radically from the
“less sophisticated” interviews with children that were routinely conducted 30 years ago.
About a fifth of exonerations with improper questioning of children are not child sex abuse
hysteria cases (15/64)—including several in which the improper questioning occurred well after
that epidemic was over. For example, Michael Washburn was convicted in 2002 of raping a
three-year-old girl in Massachusetts nine years earlier. At trial, the girl—by then 12—admitted
that a police officer had promised her “presents” if she named the person who touched her, and
74
Kyle Zirpolo, as told to Debbie Nathan, I’m Sorry; A Long-Delayed Apology from One of the Accusers in the
Notorious McMartin Pre-School Molestation Case, Los Angeles Times, Oct. 30, 2005, at MM10.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 44 • National Registry of Exonerations • September 1, 2020
said she would never see her father again if she didn’t. Washburn was exonerated in 2003 by
medical evidence that undermined the claim that the girl had been raped.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 45 • National Registry of Exonerations • September 1, 2020
V. Misconduct in
Interrogations
1. BACKGROUND
In November of 1988, a 74-year-old woman was shot and beaten to death in
Rochester, New York. Frank Sterling became a suspect and voluntarily submitted
to questioning by the police, repeatedly, and provided an alibi that his co-workers
corroborated.
The investigation languished until 1991, when a new team questioned Sterling
again. He was hypnotized, shown photos of the crime scene to “help him
remember,” given a polygraph, and told he was justified in hurting the victim
because she “deserved” it. Officers lied and told Sterling that his brother had
implicated him. After eight hours, Sterling confessed, although many of the
details he gave were incorrect. He was convicted of murder.
Twenty-two years later, Sterling was exonerated by DNA tests that implicated
another man, who gave a detailed confession that matched the facts of the crime.
True or false, confessing to a crime is almost never in a suspect’s interests. A few suspects
volunteer confessions anyway, but usually police have to convince, trick, manipulate, or force
suspects to say things that may send them to prison for years, or even lead to their execution.
Courts acknowledge that, and strike a balance of sorts. They permit interrogating officers to use
some types of lies, tricks, manipulative techniques and moderate forms of coercion, but prohibit
other types of lies, promises and threats—and any use or threat of physical violence.
In this Section, we focus on violence and other types of misbehavior that are classified as
misconduct in interrogations. That misconduct produced 165 false confessions by defendants
who were later exonerated—7% of all exonerations (165/2,400), 57% of those with false
confessions (165/292). We also discuss how manipulative and coercive techniques that are
permitted by courts also contribute to false confessions. Last, we discuss 110 cases in which
misconduct in the interrogation of codefendants of exonerees produced confessions in which
those codefendants falsely implicated the innocent exonerees.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 46 • National Registry of Exonerations • September 1, 2020
In Section VI, on fabricated evidence, we discuss 36 false confessions by exonerees that were
obtained by a different type of misconduct: making things up. These are cases in which police
reported “confessions” the exonerees never made; instead police officers lied and said they did.
Fourteen of those cases include misconduct in interrogation. We include those cases in this
Section as well because (as we explain in Section VI) misconduct in interrogation and
fabrication cannot always be neatly separated. In some cases the coercive interrogation was part
of the process of fabrication, and in a few the police obtained actual false confessions and
“improved” upon them by adding fabricated ones as well.
a. Misconduct and permissible interrogation techniques
Frank Sterling did not want to confess. He had maintained his innocence through several
interrogations. So, three years after the crime, Sterling’s interrogators lied to him about the
evidence against him, administered a lie detector test that was rigged to make him appear to fail,
showed him pictures of the body and crime scene, and fed him details of a killing he knew
nothing about. They videotaped the 20-minute-long confession Sterling eventually gave, but
they did not record the interrogation that led to that confession. These tactics are known to
increase the risk of false confessions,
75
but they are not considered misconduct.
In August 2006, a 70-year-old man was found shot dead in his car in an apparent
robbery in New Haven, Connecticut. A suspect was arrested, and claimed he
committed the crime with 16-year-old Bobby Johnson. Johnson was picked up,
told police he had been at a pharmacy at the time of the shooting, and was
released. A few days later, detective Clarence Willoughby, who claimed a “100%
success rate” in solving homicides, joined the investigation, and Johnson was
brought back for extended questioning, without his parents.
Willoughby threatened Johnson with the death penalty if he didn’t confess, and
promised him probation if he did. He told Johnson that unless he confessed he
would never see his family again, and that the police had physical evidence tying
him to the crime—which was a lie. Johnson ultimately gave a tape-recorded
confession, which he practiced with the officers before they turned on the
recorder. A few days later, ballistics tests showed that Johnson’s confession was
inconsistent with physical evidence, so officers told him he had to “correct” his
statement or his “deal” for probation would be revoked—and Johnson made a
second “revised” taped confession.
Johnson pled guilty to murder and was sentenced to 38 years. He was exonerated
in 2015, after Detective Willoughby was caught extracting other false confessions
75
See Saul M. Kassin et al., Police-Induced Confessions: Risk Factors and Recommendations, 34 Law & Hum.
Behav. 13 (2010); Richard A. Leo, False Confessions: Causes, Consequences, and Implications, 37 J. Am. Acad.
Psychiatry & L. 332, 332–34 (2009). See generally Trainum, James L. How the Police Generate False Confessions:
An Inside Look at the Interrogation Room. Rowman & Littlefield Publishers; Jerome H. Skolnick & Richard A. Leo,
The Ethics of Deceptive Interrogation, 11 Crim. Just. Ethics 3, 5–8 (1992).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 47 • National Registry of Exonerations • September 1, 2020
and false witness statements, and new evidence implicated another man in the
murder.
Bobby Johnson’s interrogation—unlike Frank Sterling’s—was rife with misconduct. The police
pretended to “plea bargain” with Johnson (which only prosecutors may do), they lied about the
law that governs the case (Johnson, as a juvenile, was not eligible for the death penalty
76
), and
they threatened to never let him see his family again. These tactics are all prohibited and can (at
least in theory) lead to sanctions against police officers and exclusion of confessions from
evidence.
b. The frequency of false confessions, in Chicago and elsewhere
False confessions are not easy to obtain. That’s true of all confessions that are not volunteered:
persuading and manipulating a suspect to confess to serious crimes is a time-consuming and
“expensive procedure [that] is generally reserved for the most serious cases where there is no
other evidence sufficient to convict—which usually means a murder with no surviving
eyewitnesses.”
77
It works. Many reluctant defendants confess to crimes they committed. But it
also produces a steady trickle of false confessions. They occurred in 12% of known exonerations
(292/2,400), mostly in murder cases, where the rate of false confessions is nearly four times
that of other exonerations, 22% vs. 6%. Overall, almost 70% of false confessions in known
exonerations are for murder (204/292). See Table 12.
Table 12: Proportion of Exonerations with False Confessions, in Chicago and Elsewhere
Exonerations
in Chicago
Exonerations
Elsewhere
ALL
JURISDICTIONS
Murder
54%
(64/118)
18%
(140/790)
22%
(204/908)
All other Crimes
10%
(11/112)
6%
(77/1380)
6%
(88/1492)
ALL CASES
33%
(75/230)
10%
(217/2,170)
12%
(292/2,400)
Table 12 also displays another basic fact about false confessions in the United States: the rate
among exonerations in Chicago (Cook County), Illinois, is more than three times the rate
elsewhere. Among murder exonerations in Chicago, it’s an astounding 54%. Chicago produced a
quarter of all false confessions in exonerations in the nation (75/292); false convictions for
76
Roper v. Simmons, 543 U.S. 551 (2005).
77
Samuel R. Gross et al., Exonerations in the United States, 1989 Through 2003, 95 J. Crim. L & Criminology 523,
544–45 (2005).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 48 • National Registry of Exonerations • September 1, 2020
murder in Chicago account for 22% of all false confessions we know of. As Peter Neufeld put it,
“what Cooperstown is to baseball, Chicago is to false confessions. It’s the Hall of Fame.”
78
2. WHAT COUNTS AS COERCIVE MISCONDUCT IN AN INTERROGATION?
From the 1940s through 1960s the Supreme Court developed the doctrine that police behavior
that is so “coercive” that it makes a confession “involuntary” violates the due process clause of
the Fourteenth Amendment and requires the exclusion of the confession from evidence.
79
Over
the decades, the Supreme Court and lower courts have identified several specific types of
misconduct that may make confessions inadmissible because they are deemed “involuntary.”
In 1953, in Stein v. New York, for example, the Court made clear that the use of violence, or the
threat of imminent violence, automatically requires exclusion of any resulting confession.
80
For
other types of misconduct, courts weigh the “totality of [the] circumstances”
81
and determine on
a case-by-case basis whether the confession was “involuntary” enough that its use in evidence
against the suspect violates due process.
82
For example, in Rogers v. Richmond the Court weighed the impact of the officers’ threats to
arrest the suspect’s sick wife, concluded that such threats to a relative of the suspect may have
produced an involuntary confession, and remanded the case to a lower court to decide that
issue.
83
In other cases, courts have condemned sham plea bargaining by police officers who pretend
that they can and will obtain lenient sentences, or a reduction or outright dismissal of charges,
in return for confessions.
84
Sham plea bargaining is a form of lying about the law—in that
context, lying about the legal authority of the police, who claim to be able to do things only
prosecutors or judges can do. Lying about the law is not permitted and may contribute to a
finding that a confession is involuntary,
85
but vague promises of better treatment if a suspect
78
60 Minutes: Chicago: The False Confession Capital (CBS television broadcast Dec. 19, 2012).
79
See Rogers v. Richmond, 365 U.S. 534, 540–41 (1961).
80
Stein v. New York, 346 U.S. 156, 182 (1953). On the other hand, scaring a suspect by telling him that he is likely to
be a victim of sexual assault in the future if he is imprisoned does not automatically make a confession inadmissible,
but may contribute to a finding of involuntariness. See e.g., Little v. United States, 125 A.3d 1119, 1127–28 (D.C.
2015).
81
See, e.g., Haynes v. Washington, 373 U.S. 503, 513–14 (1963).
82
Id. at 514.
83
Richmond, 365 U.S. at 543–44.
84
See State v. Howard, 825 N.W.2d 32, 41 (Iowa 2012); see also Commonwealth v. Magee, 668 N.E.2d 339, 344–45
(Mass. 1996).
85
State v. Walker, 493 N.W.2d 329, 334–35 (Neb. 1992).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 49 • National Registry of Exonerations • September 1, 2020
confesses are not misconduct,
86
and lying about the evidence the police have against the suspect
is permitted and routine.
87
Finally, permitted pressure tactics can become impermissible if they’re carried too far. Isolation,
extended interrogation, uncomfortable conditions of confinement and prolonged aggressive
questioning by multiple interrogators, are all common practices that are widely used and
tolerated. But if they’re sufficiently severe, and if a court is so disposed, they may contribute to a
finding that a confession was involuntary under the totality of the circumstances.
88
So far we have not mentioned Miranda v. Arizona,
89
easily the best known Supreme Court case
on interrogations. Miranda, as it was originally decided, provided that a suspect in police
custody may not be questioned unless she has been informed that she has the right to remain
silent, to talk to a lawyer (free if necessary), and to have that lawyer present at the questioning;
and unless she then voluntarily waived those rights. Since Miranda was decided, most litigation
on confessions has focused on its requirements.
Over the past 20 years, however, Miranda has been severely limited by several Supreme Court
cases. Among other developments, the Court has held that statements obtained in violation of
Miranda may be used against other criminal defendants—and to rebut the defendants who
made them if they testify at trial; and that officers may not be sued for damages for violating
Miranda.
90
In light of these cases, the best interpretation of current Supreme Court law is that a Miranda
violation is not in itself “misconduct” that may produce an “involuntary” confession, but rather
an event that creates a limited opportunity for a defendant to prevent the state from using some
statements by that defendant in its case in chief at a trial of that defendant.
Accordingly, we do
not classify violations of Miranda as official misconduct.
3. COERCIVE MISCONDUCT IN INTERROGATIONS THAT PRODUCE FALSE
CONFESSIONS
Most false confessions in convictions that ended in exoneration were obtained by coercive
misconduct in interrogations, 57% overall. As with the frequency of false confessions
themselves, Chicago leads the pack in the rate of misconduct among false confessions, 77%
compared to 49% elsewhere. It appears that Chicago has a uniquely high rate of false
confessions in large part because of the prevalence of misconduct that produces most of those
false confessions. See Table 13:
86
See, e.g., Commonwealth v. Mandile, 397 Mass. 410, 414 (1986).
87
Frazier v. Cupp, 394 U.S. 731, 739 (1969). See also Laurie Magid, “Deceptive Police Interrogation Practices: How
Far Is Too Far?” 99 Mich. L. Rev. 1168, 1169 (2001).
88
See, e.g., Davis v. North Carolina, 384 U.S. 737, 752–53 (1966); Ashcraft v. Tennessee, 322 U.S. 143, 153–55
(1944).
89
Miranda v. Arizona, 384 U.S. 436, 467–473 (1966).
90
Chavez v. Martinez, 538 U.S. 760 (2003).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 50 • National Registry of Exonerations • September 1, 2020
Table 13: Proportions of Exonerations with False Confessions after Coercive Misconduct
in Interrogations, in Chicago and Elsewhere*
Exonerations
in Chicago
(75)
Exonerations
Elsewhere
(217)
ALL
EXONERATIONS
(292)
Specific
Factors
Actual or Threatened
Violence
69% (52/75)
24% (53/217)
36% (105/292)
Sham Plea Bargaining and
other Lies about the Law
23% (17/75)
18% (40/217)
20% (57/292)
Threats to Relatives
and other Third Parties
7% (5/75)
8% (18/217)
8% (23/292)
Coercion Found by a Court,
Without Specific Factors
3% (2/75)
3% (6/217)
3% (8/292)
Other Cases with Coercive
Misconduct in Interrogations
3% (2/75)
6% (14/217)
5% (16/292)
ALL CASES WITH MISCONDUCT
IN INTERROGATIONS
77% (58/75)
49% (107/217)
57% (165/292)
* Some cases include more than one type of misconduct.
In the great majority of cases with misconduct that produced false confessions—87%
(143/165)—that misconduct included one or more of the three specific categories we’ve
discussed: violence, sham plea bargaining and other lies about the law, and threats to third
parties. In the remaining cases, the misconduct was more variable, but typically consisted of
aggravated forms of prolonged questioning, isolation, sleep deprivation, and other practices that
are tolerated in milder doses.
a. Violence
Physical violence, or the threat of imminent violence, is the most common type of misconduct in
interrogations that produce false confessions. It was used in 64% of interrogations with
misconduct (105/165). The actual conduct involved varied greatly. Sheriff’s deputies in Lake
County, Illinois, for example, threatened to “beat the [expletive] out of” Jason Strong, while
police officers in Philadelphia chained Edward Baker to a chair and actually beat him with sticks
and a telephone book for hours. Both are criminal acts—assault in one case, battery in the
other—and both are subject to criminal, civil, and professional sanctions.
i. Torture in Chicago
Threatened or actual violence contributed to more than a third of false confessions in the
Registry. See Table 13. Here again, Chicago is in a class of its own. Sixty-nine percent of false
confessions in exonerations in Chicago were obtained by violence or threats of violence, nearly
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 51 • National Registry of Exonerations • September 1, 2020
three times the rate for the rest of the country. Half of all known cases of violence in obtaining
false confessions from exonerated defendants in the country occurred in Chicago (52/105).
Some other large cities have elevated rates of violence in interrogations, but they lag far behind.
New York City is a telling comparison.
New York City has a much larger population than Chicago, 8.6 million compared to 2.7 million,
but about 25% fewer exonerations, 173 vs. 230. Only 13% of exonerations in New York City
included false confessions (22/173)—about the same rate as the country as a whole—compared
to 33% in Chicago. And while the rate of violence in the interrogations that led to those false
confessions is comparatively high in New York City—41% (9/22), nearly double the rate for the
rest of the country, excluding Chicago and New York City (44/195)—it’s much lower than the
rate of violence in exonerations with false confessions in Chicago, 69%. The net result is that
Chicago, with one-third the population of New York City, has almost six times as many
exonerations with false confessions that were extracted by violence, 52 compared to 9.
A major part of the reason for Chicago’s unique status is that from the early 1970s through the
1980s hundreds of suspects—almost all Black men—were tortured by a group of Chicago police
detectives under the command of Lieutenant (later Commander) Jon Burge. (We discuss the
Burge torture regime and its causes in greater detail below.
91
)
In 1986, a couple was stabbed and killed in their apartment on the south side of
Chicago. A few days later, a 15-year-old girl told police that Aaron Patterson
admitted to committing the murder, and a neighbor said he had seen Eric Caine
in the vicinity. Patterson and Caine were arrested and taken for interrogation to
the Area 2 police station, under the command of Lieutenant Jon Burge.
Caine, the son of a Chicago police officer, was told that Patterson had confessed
that he and Caine went to the house to find weapons and ended up killing the
couple. After Caine refused to confess, Detectives beat him and took him to see
Patterson, who had been beaten so badly he could barely speak. Caine signed a
confession.
Patterson, left alone in the interrogation room with a confession to sign, used a
paperclip to scratch a message on a metal bench: “Police threaten me with
violence. Slapped and suffocated me with plastic. No lawyer or dad. Sign false
statement to murders.”
Patterson and Caine were convicted of murder in 1989. Patterson was sentenced
to death; he received a pardon based on innocence in 2003. Caine was sentenced
to life in prison, and exonerated in 2011.
91
See infra , Section XII.1.b.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 52 • National Registry of Exonerations • September 1, 2020
Caine and Patterson were exonerated because of litigation by other torture victims. In 1989, a
man name Andrew Wilson sued the city for torture at the hands of Burge and his officers.
92
That
case led to an investigation of Burge and his subordinates by the Chicago Police Department,
and a report in 1990 that concluded that “the number of incidents in which an Area 2 command
member is identified as an accused [in torturing suspects] can lead to only one conclusion.
Particular command members were aware of the systematic abuse and perpetuated it either by
actively participating in [the] same or failing to take any action to bring it to an end.”
93
The Burge torture cases have been in the public eye ever since. In 2009, the State of Illinois
created a “torture commission” to review his actions and the cases he oversaw.
94
Ultimately, 19
defendants who confessed under torture by Burge and his men were exonerated and are listed in
the Registry.
95
If their cases had not been part of a well-established pattern of torture, many
would still be in prison.
No eyewitness or physical evidence placed Caine at the scene of the crime. Without Caine’s
confession (and Patterson’s) there was no evidence against him, but such is the power of a
confession—“the queen of evidence”
96
—that nothing else was needed. The trial pitted Caine—a
terrified Black teenager, trying to avoid prison if not execution by convincing the jury that he
was tortured—against experienced white police officers who testified that he confessed
voluntarily. Naturally, Caine lost.
After conviction, Caine faced the added burden of convincing a court to reject the jury’s findings.
Paradoxically, the absence of corroborating evidence made that task harder. There were no
eyewitnesses to recant, no biological evidence to test, and no witness bias to uncover. All he
could do was try, once again, to pit his credibility against that of the police.
Of the 53 exonerees outside of Chicago who falsely confessed due to violence or threats of
violence, only three were exonerated solely by evidence that their confessions were coerced.
97
The other 50 all relied on other types of evidence that did not exist in Caine’s case: DNA tests,
witness recantations, alibi witnesses, non-DNA forensic evidence, or information pointing to the
real criminal.
92
Chicago Tribune staff, Jon Burge and Chicago's legacy of police torture, Chicago Tribune (Sept. 19, 2018).
93
Chi. Police Dep’t Office of Prof’l Standards, Special Report (1990).
94
TIRC Home: Mission and Procedures Statement, St. of Ill. Torture Inquiry and Relief Commission).
95
The 19 exonerees in the first 2,400 exonerations who confessed after torture by Burge and his subordinated are:
James Andrews, Kevin Bailey, Corey Batchelor, David Bates, Eric Caine, Arnold Day, David Fauntleroy, Madison
Hobley, Stanley Howard, Anthony Jakes, Melvin Jones, Ronald Kitchen, Leroy Orange, Aaron Patterson, Marvin
Reeves, Alonzo Smith, Michael Tillman, Shawn Whirl, and Stanley Wrice. Since we completed the 2,400 exoneree
dataset for this report, we have added five additional exonerations of defendants who gave false confessions after
abuse by Jon Burge and his subordinates, for a current total of 24: Gregory Banks, James Gibson, Victor Safforld,
Demond Westin, and Keith Walker.
96
In Latin: confessio est regina probationum. See Peter Brooks, Troubling Confessions: Speaking Guilt in Law and
Literature, University of Chicago Press (2000) at p. 93.
97
Robert Coney, William Oakes, James Simmons.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 53 • National Registry of Exonerations • September 1, 2020
In Chicago, however, after courts found that dozens of other men had been tortured by Burge
and his men, Caine’s words did not stand alone. That’s why he was freed after 22 years in prison.
We have no idea how many Eric Caines in other cities have never been identified because
officers who abused them did not do it as part of a concerted pattern of torture that, eventually,
could not be ignored.
(To be clear, torture in Chicago was not a one-man show. Thirty-three of the 52 Chicago
exonerations with false confession extracted by violence did not involve Jon Burge in any
capacity; in 13, the interrogations happened after he retired.)
ii. Violence in interrogations of suspects with mental disabilities
It’s well known that people with mental illness or diminished intellectual capacity are far more
likely than others to confess falsely.
98
We see that in these cases. Seventy percent of exonerees
with mental disabilities falsely confessed (103/147)—including 81% of mentally-disabled murder
exonerees (77/95)—compared to 8% of exonerees without such disabilities (189/2,253).
Mentally-disabled exonerees who falsely confessed were less likely than others to be victims of
violence in interrogation, 27% (28/103) compared to 41% (77/189)—and half of the mentally-
challenged exonerees who confessed after some use of violence by police experienced
comparatively mild versions: slapping, grabbing, or threatening. None was subjected to the sort
of torture that was regularly used by Jon Burge and his crew.
The likely explanation is that it’s easier to get a confession without resorting to violence if the
suspect has a mental disability. Manipulation and lesser forms of coercion usually do the trick,
and if violence is used, relatively mild forms are generally enough.
b. Sham Plea Bargaining and other lies about the law
Detective Willoughby lied when he offered Bobby Johnson probation in exchange for a
confession. He had no power to make that deal and could not enforce it if he tried. Willoughby
also told Johnson he’d be sentenced to the death penalty if he did not confess. That too was a lie.
By 2005, the death penalty could not constitutionally be imposed on a defendant as young as
Johnson.
Deception about the law—unlike deception about the facts of the case at hand—is misconduct.
Police officers are seen by civilians as authorities on the law; suspects are likely to believe police
lies about the law because they have no independent knowledge of their own—and that can
make a resulting confession “involuntary” and inadmissible in court. Judging from
exonerations, the most common form of deception about the law in interrogations is sham plea
bargaining by police.
98
See Steven A. Drizin & Richard A. Leo, The Problem of False Confessions in a Post-DNA World, 82 N.C. L. Rev.
891, 970–71 (2004); Nat’l Registry of Exonerations, Age and Mental Status of Exonerated Defendants Who
Confessed.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 54 • National Registry of Exonerations • September 1, 2020
The power to decide whether to prosecute, and on what charges, rests entirely with prosecutors.
If a suspect confesses because the prosecutor offered in return to reduce the charge from murder
to manslaughter, that creates an enforceable deal. Plea offers by prosecutors can put extreme
pressure on suspects to confess, truthfully or not. Kenneth Kagonyera, for example, was facing
the death penalty in North Carolina when he confessed to a murder he had no part in, and pled
guilty, in return for a sentence of 12 to 15 years. That sort of pressure is legal, accepted and a
central element of our system of criminal procedure.
99
Since he had no authority to strike a deal with Johnson, Detective Willoughby’s offer of
probation in return for a confession amounted to a lie about criminal procedure. In 18% of the
false confession cases in the Registry, an officer purported to offer a plea arrangement to a
suspect in return for a confession (54/292). These unauthorized and unenforceable offers
included probation, diversion from criminal prosecution, lenient sentencing, reduced charges,
and psychiatric treatment in lieu of prison.
In about half of those cases, officers threatened the suspects with the death penalty if they did
not confess (25/54). The police, of course, had no authority to make good on the implied or
explicit promise that the exoneree would be spared the death penalty if he did confess; Damon
Thibodeaux, for example, falsely confessed after he was threatened with the death penalty in
Louisiana if he refused to do so; he was sentenced to death anyway.
c. Threats to Third Parties
On March 8, 1987, Richard Lapointe and his wife, Karen, went to the house of
Karen’s grandmother, Bernice Martin, in Manchester, Connecticut, and visited
with her for a few hours. They left at 4 p.m. At 8 p.m. they received a call from
Bernice’s daughter (Karen’s aunt) asking Lapointe to check on Bernice. Twenty-
seven minutes later, Lapointe called 911 to report a fire at Bernice’s house.
Bernice was removed from the house, dead from stab wounds.
The murder investigation stalled for two years until a new detective took over and
focused on Lapointe. During questioning, police told Lapointe that his wife had
implicated him in the murder—and said they would arrest her and take his son
away if he did not confess.
After nine hours of interrogation, Lapointe (who suffered from a congenital
neurological disease) signed a confession that was inconsistent with forensic
evidence on many points, and left out other key facts. Lapointe was convicted in
1992, largely based on his confession. He was exonerated in 2015 after it was
discovered that a fire investigator had concluded that the fire began while
Lapointe was still at home, and DNA tests excluded him as the source of the
biological evidence found at the crime scene.
99
See Brady v. United States, 397 U.S. 742 (1970); Bordenkircher v. Hayes, 434 U.S. 357, 362–64 (1978); Missouri
v. Frye, 566 U.S. 134, 144 (2012) (“ [P]lea bargaining … . is not some adjunct to the criminal justice system; it is the
criminal justice system.”); Lafler v. Cooper, 566 U.S. 156 (2012).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 55 • National Registry of Exonerations • September 1, 2020
The threat to arrest Lapointe’s wife and take his son into state custody was credible. The police
may have had enough evidence to arrest Karen, and if they did, they would have called child
protective services to take their son. In fact, they never arrested her and probably never
intended to. Their goal was to get Lapointe to confess, and it worked. For that purpose, the
threat was misconduct.
Police are allowed to tell a suspect that if he doesn’t talk they will arrest him, assuming they have
probable cause to do so. Apparently, courts believe that suspects will not confess to crimes they
did not commit simply to avoid arrest. But similar threats against a suspect’s spouse or child
change the equation. They are thought to be intrinsically unfair and more likely to lead to false
confessions by suspects who are anxious to protect those they love.
In 23 cases in the Registry—8% of exonerations with false confessions—the exoneree falsely
confessed after an officer threatened to arrest a member of the exoneree’s family, place one or
more children in the exoneree’s family in state custody, or otherwise put family members in
harm’s way if the exoneree did not confess. See Table 13.
4. PERMITTED INTERROGATION PRACTICES THAT LEAD TO FALSE CONFESSIONS
Frank Sterling, as we saw, confessed to a murder he did not commit after police lied about
evidence against him and convinced him he had no real choice but to confess. There was no
misconduct in that interrogation. Lying about the facts of a crime or the evidence the police have
obtained is not misconduct, and may be effective at securing true confessions—but, as we also
noted, it increases the risk of false confessions.
100
Several interrogation tactics fit this description: police are allowed to use them to maneuver
reluctant suspects to confess, and they do it because it works—but in the process, they generate
false confessions from innocent suspects. This applies to the overall structure of many American
murder interrogations: prolonged aggressive questioning of a suspect who is uncomfortable,
tired, and isolated from all other contacts. As we’ve mentioned, that practice is permitted by the
courts, except in rare cases in which they say it went too far.
We don’t have sufficiently detailed data to detect and describe many of the features of the
interrogations that produced false confessions. But we are able to identify a few practices of this
sort: lying about the investigation; making promises and threats that are not categorically
prohibited; feeding details of the crime to the suspect; and interrogating a juvenile with no
parent present.
a. Lying about the investigation
Lying about the law in interrogations is prohibited, as we’ve mentioned, perhaps because
suspects are likely to regard police as experts on the law and will therefore be easily misled.
Lying about evidence is another matter; it’s routinely permitted.
101
Courts seem to believe that
100
Supra Section V.1.a.
101
See, e.g., Frazier v. Cupp, 394 U.S. 731, 739 (1969).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 56 • National Registry of Exonerations • September 1, 2020
while suspects who are guilty may be tricked into confessing by false claims that their guilt is
already proven, those who know they are innocent will not be shaken by lies about witnesses or
physical evidence. That is not always true in practice.
In 2004, police in Maricopa County, Arizona, questioned 20-year-old John
Watkins because he resembled a woman’s description of the stranger who raped
her. They told Watkins that they had recovered his fingerprints from the scene,
and that the victim had identified him. Watkins agreed to take a voice-stress
analysis test and was told that he failed. In the face of this “evidence,” Watkins
confessed. It was all lies. There were no fingerprints and no identification by the
victim, and he had not failed the voice test. In 2004, Watkins pled guilty to rape;
in 2010, he was exonerated by DNA evidence that proved he was not the rapist.
Watkins, like Frank Sterling, confessed because police officers convinced him (falsely) that they
had him dead to rights without a confession. Police get a lot of convictions that way—including
in 23% of cases in which innocent suspects confessed, were convicted, and were later
exonerated. Often, as in the Watkins’ case, they lie about physical evidence that supposedly
linked the exoneree to the crime. In four cases, police gave suspects real or fake polygraph tests,
and then falsely told them they had failed.
Another common deceptive tactic is to claim that an eyewitness identified the suspect, or that an
alibi witness failed to support their defense. John Horton, for example, told police he was with
his brother at the time of the murder they were investigating. He falsely confessed (and was
wrongly convicted) after police lied and said that his brother had repeatedly denied seeing him
on the night of the murder.
b. Permissible promises and threats
As we’ve discussed, some types of promises are off limits in interrogations. In particular, police
are forbidden to promise specific legal outcomes if the suspect confesses—dismissal of charges,
probation, whatever—because they don’t have the legal authority to make good on such
promises. Police are also forbidden to make some threats, especially threats of violence and
threats to take action against children, spouses, and other third parties who are dear to the
suspect.
But there are many types of promises and threats that police are allowed to make in
interrogations, whether or not they ever intend to keep them. Vague promises or threats are
generally fine, even though a terrified young suspect that might not notice the difference
between “Things will get very tough if you don’t tell us what happened, right now” (permitted)
and “You’ll be charged with aggravated murder unless you tell us what happened, right now”
(prohibited).
The basic test is this: Are the police promising or threatening to do things that they have the
power to do? If so, it’s OK. Police may make recommendations to prosecutors and judges, so
they may promise (or imply) that they will do so, even if they have no such intention. They may
also threaten to arrest a suspect, or promise to refrain from making an arrest—both are within
their power. Probably their most common promise is to allow suspects to go home if they
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 57 • National Registry of Exonerations • September 1, 2020
confess. Indeed, some exonerees—including Bobby Johnson—were allowed to go home after
they confessed. They just didn’t get to stay there.
Some promises were more specific:
In March 1997, 16-year-old Fancy Figueroa was raped by a stranger after she got
home from school in Queens, New York. She was taken to a hospital for a rape
kit, but when it was discovered she was pregnant, police concluded that she lied
about the rape to cover up the real cause of her pregnancy. They told her that if
she wrote down on a piece of paper that she was lying, they would help her look
for the rapist. She complied, and was charged with and convicted of filing a false
police report. The officers didn’t follow up on the rape—they considered the
matter closed by Figueroa’s admission that she’d lied. She was exonerated seven
years later when the DNA from her rape kit was matched to a serial rapist in a
routine search of a DNA database.
The officers’ promise to investigate Figueroa’s report if she admitted it was false made no sense,
and they had no intention of keeping it—but it was a permissible promise. Some permissible
promise or threat, or a combination, contributed to 18% of false confessions in exonerations.
c. Feeding the suspect details of the crime
In 1993, the United States Court of Appeals for the Fourth Circuit reviewed Earl Washington’s
1984 rape-murder conviction in Virginia, and concluded that Washington’s confession proved
his guilt beyond doubt:
“Washington had supplied without prompting details of the crime that were
corroborated by evidence taken from the scene. ... He had confessed to the crime… as
one who was familiar with the minutiae of its execution.”
102
In other words, his confession was persuasive because he volunteered facts that only the
criminal could know. But Washington was innocent. He was exonerated in 2000 after two sets
of DNA tests proved he was not the rapist who killed the victim.
How did Washington know facts “that only the criminal could know”? He learned them from
the only other people who know such facts: the officers who investigated the case. Washington
had a learning disability; his IQ was 69. He handled his disability by politely agreeing with
people in authority, such as police officers. When asked whether he committed the crime, he
confessed, but in the first version of his confession he didn’t know the race of the victim, her
address, or that she had been raped. Eventually, with a great deal of help, he managed to
produce a confession that was coherent enough for the police to use—and not too inconsistent
with known facts.
102
Washington v. Murray, 4 F.3d 1285, 1292 (4th Cir. 1993).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 58 • National Registry of Exonerations • September 1, 2020
Washington’s case is hardly unique. Professor Brandon Garrett studied 40 DNA exonerations in
which the defendants falsely confessed. He reports:
“Thirty-six of the thirty-eight cases for which transcripts were obtained had confessions
that reportedly included specific details about how the crime occurred. … At trial, law
enforcement testified that the suspect had volunteered specific details about how the
crime occurred, typically details corroborated by expert evidence or crime scene
evidence. … Almost all exonerees were reported to have provided detailed statements
that included facts likely to be known only by the culprit.”
103
It is not misconduct for police to tell a suspect the details of a crime. The most common way they
do it is to ask leading questions (“You stabbed her with the butcher knife, didn’t you?”)—which
was certainly done in Washington’s interrogation. Officers may take the suspect to the scene of
the crime, which was also done with Earl Washington, after he himself was unable to locate the
apartment where the crime occurred. In other cases—for example, in Frank Sterling’s murder
interrogation—police show the suspect pictures of the crime scene or the victim. In some cases,
officers divulge critical facts without realizing it. Former District of Columbia police detective
James Trainum, for example, has described a case in which he did so inadvertently by showing
the suspect photographs that contained information he didn’t notice, and asking leading
questions he forgot.
104
Hearing a detailed description of a violent crime—or seeing pictures, or being taken to the
scene—can be very troubling. It puts pressure on the suspect to confess and end the ordeal;
that’s one reason it’s done. Like other pressure tactics, it can contribute to confessions from
innocent defendants, and it can make those confessions seem more credible than they are.
“Telling details” in a confession make it believable and powerful if the suspect knew those
details. If they were supplied by the police, knowingly or unawares, they may convince a court
that an innocent person is guilty.
The simple cure for feeding details to the suspect is to studiously avoid doing it—but that can be
difficult, especially since it can happen unintentionally. Not surprisingly, feeding details of the
crime to the suspect is very common; it occurred in 56% of false confessions by exonerated
defendants. The next best solution—and the one more likely to work—is to make video
recordings of interrogations so investigators, lawyers and courts can tell whether details of the
crime were supplied by the police. Detective Trainum, for example, was only able to figure out
how he inadvertently passed such information to an innocent suspect by carefully reviewing a
video of the interrogation. Recording interrogations in full has become more common in recent
103
Brandon L. Garrett, The Substance of False Confessions, 62 Stan. L. Rev. 1051, 1066 (2010). See also Brandon
Garrett, Judging Innocence, 108 Columbia L. Rev. 55, 89 (2008).
104
James Trainum and Diana Havlin, “A False Confession to Murder in Washington, DC,” in Criminal Investigative
Failures, ed. D. Kim Rossmo (Boca Raton, Florida: CRC Press, 2009), 205–217.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 59 • National Registry of Exonerations • September 1, 2020
years,
105
but was only done in eight interrogations that led to false confessions by defendants
who were later exonerated.
d. Interrogating a juvenile with no parent present
Juveniles, like suspects with mental disabilities, are prone to false confessions. Thirty-six
percent of exonerees who were under 18 at the time of the crime with which they were charged
falsely confessed, compared to 10% of exonerees who were 18 or older.
Most states have rules that specify whether and when a parent or guardian must be present at
the interrogation of a juvenile suspect, but those rules vary widely. Some states require parents
to be present if the child requests it;
106
others require the police to notify a parent who can waive
the right to be present.
107
Some states require parental presence for any child under 18;
108
some
only require it for those under 16.
109
The Supreme Court has not spoken on the issue.
Given the wide range of rules on parental presence, we cannot say that interrogating a minor
without a parent is misconduct—but it certainly appears to be dangerous. We know that almost
two-thirds of exonerees who falsely confessed as juveniles did so without a parent present
(48/76); there may be others where that fact was not mentioned in the records we could obtain.
Part of the reason may be that police committed misconduct in 67% of interrogations of
juveniles without parents, but only 46% of such interrogations with a parent present.
e. General patterns
In Table 14 we display the rates of the permitted interrogation practices we discussed among
exonerations with and without misconduct in interrogations. Each of these permitted practices
that contribute to false confessions was more common in interrogations with misconduct than
in those without misconduct. Overall, 79% of interrogations with misconduct included at least
one of these dangerous practices, compared to 57% of interrogations without misconduct.
110
105
See infra Section XII.2.b.ii(a).
106
See, e.g., Ala. Code § 12-15-202 (2008).
107
See, e.g., Me. Rev. Stat. Ann. tit. 15, § 3203-A (2011).
108
See, e.g., Colorado Revised Statutes Title 19. Children's Code § 19-2-511.
109
See, e.g., North Carolina General Statutes Chapter 7B. Juvenile Code § 7B-2101.
110
The contrast become slightly sharper if we focus on interrogations with violence, the most severe form of
misconduct: 81% of interrogations with violence or threats of violence included one or more of these tactics (85/105),
compared to 64% of interrogations with no suggestion of violence (119/186).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 60 • National Registry of Exonerations • September 1, 2020
Table 14: Proportions of Exonerations with Permitted Interrogation Practices that
Contribute to False Confessions, With and Without Misconduct in Interrogation*
False Confessions
With Misconduct
In Interrogation
False Confessions
Without Misconduct
In Interrogation
ALL
EXONERATIONS
WITH FALSE
CONFESSIONS
Lying about the
Investigation
24%
(40/165)
21%
(27/127)
23%
(67/292)
Permissible Promises
and Threats
20%
(33/165)
16%
(20/127)
18%
(53/292)
Feeding the Suspect
Details of Crime
69%
(114/165)
39%
(49/127)
56%
(163/292)
Interrogating a Juvenile
With No Parent Present
19%
(32/165)
13%
(16/127)
16%
(48/292)
AT LEAST ONE IDENTIFIED
CONTRIBUTING PRACTICE
79%
(131/165)
57%
(73/127)
70%
(204/292)
* Some cases include more than one type of contributing practice.
Violence in interrogations is in a category of its own, but the other tactics we discussed,
permitted and prohibited alike, all aim to exhaust, trick, scare and manipulate suspects into
confessing—without physical abuse. It’s no surprise that interrogators who work close to the line
often step over it, and that those who break the rules also make aggressive use of legitimate
moves. Misconduct and permitted interrogation practices differ from each other in degree, not
in kind.
With the possible exception of interrogating a juvenile without a parent, all these practices—
torture, threats to third parties, lying about the investigation, feeding details to the suspect, and
so forth—undercut the power of any resulting confession, if they are observed. They are much
less likely to happen if interrogations are recorded, and if they do occur they can be considered
by juries in evaluating the resulting confessions, and by judges in deciding on their admissibility.
There’s a hint of the possible impact in our data. Five of the eight confessions that were obtained
in interrogations that were recorded in their entirety—63%—were ruled inadmissible in evidence
by judges who saw or heard the recordings.
5. MISCONDUCT IN THE INTERROGATION OF CODEFENDANTS
Confessions by the defendants are not the only false confessions that contribute to the
conviction of innocent people. In 13% of exonerations, suspects who were, or might have been
charged as codefendants gave confessions that implicated the exonerees as well as themselves.
In fact, codefendant confessions that implicated exonerees were slightly more common than
false confessions by the exonerees themselves, 315 to 292.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 61 • National Registry of Exonerations • September 1, 2020
In some cases, the confessing codefendants avoided charges altogether. Randall Dale Adams, for
example, was sentenced to death in 1977 for killing a police officer because David Harris, the
actual killer, testified that Adams—a hitchhiker who had been in his car earlier in the day—shot
the officer from Harris’s car with Harris’s gun. Based on his testimony, Harris could easily have
been charged as an accomplice or an accessory to murder. Instead, he became the star witness at
Adams’s trial, got off scot-free, and went on to commit another murder for which he was
executed 27 years later.
111
Other confessing codefendants did not fare as well. In Austin, Texas, for example, police
threatened Christopher Ochoa with execution if he did not confess, and with rape in prison, and
with beating by the officers themselves. In 1989, Ochoa confessed, pled guilty to murder and
went on to testify at the trial of his codefendant, Richard Danziger, that they raped and
murdered the victim together. He avoided the death penalty, as promised, but he, like Danziger,
was sentenced to life in prison. Both were exonerated by DNA in 2002.
A third of exonerations with codefendant confessions also include false confessions by the
exonerees (107/315), and two-thirds do not. The net effect is that false confessions—counting
those by codefendants as well as by exonerees themselves—contributed to the convictions of 21%
of all exonerees (500/2400).
Nearly three-quarters of exonerations with incriminating codefendant confessions were murder
cases (233/315), and a disproportionate number occurred in Chicago, where a slight majority of
all murder exonerations included codefendant confessions. See Table 15. These patterns, of
course, are familiar; we saw almost exactly the same thing in Table 12, for false confessions by
exonerees.
Table 15: Proportion of Exonerations with Confessions by Codefendants, in Chicago and
Elsewhere
Exonerations
in Chicago
Exonerations
Elsewhere
ALL
JURISDICTIONS
Murder
51%
(60/118)
22%
(173/790)
26%
(233/908)
All other Crimes
3%
(3/112)
6%
(79/1380)
5%
(82/1492)
ALL CASES
27%
(63/230)
12%
(252/2,170)
13%
(315/2,400)
Codefendants who confess are usually suspects themselves. That means that police are
permitted to interrogate them using the aggressive and frequently coercive techniques that are
111
See supra Section IV.2.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 62 • National Registry of Exonerations • September 1, 2020
permitted in interrogations of all suspects, including lying about physical evidence and
statements by witnesses, and making some types of threats and promises. That sort of
questioning might be misconduct—witness tampering—if applied to ordinary witnesses, but not
when used in the interrogation of suspects. We only count the process of obtaining codefendant
confessions as misconduct if it includes actions that we would count as misconduct in
interrogations of the exonerees themselves, as suspects.
Most codefendant confessions that were used to convict exonerees were obtained without
misconduct. There was no misconduct, for example, in the interrogation of David Harris. Once
he was told that ballistic evidence confirmed that the victim was shot with his gun, he jumped at
the chance to shift the blame to Adams—and managed to get away with murder, literally. That
impulse is common, if not always as successful. Many confessing codefendants were guilty of the
crimes they confessed to, and falsely implicated others to avoid conviction or minimize the
punishment.
On the other hand, it took extreme misconduct to persuade Ralph Myers to confess that he
drove Walter McMillian to the scene of a robbery murder in Monroeville, Alabama, heard shots,
and saw McMillian standing over and robbing the body of the dead victim. Myers, a career
criminal, was suspected of a murder in a nearby county. But instead of pursuing that case, the
police told Myers that they had witnesses who would get him convicted and executed for the
Monroeville murder—unless he implicated McMillian in that crime. McMillian was sentenced to
death; Myers pled guilty and got 30 years. Five years later, defense attorneys stumbled on a
hidden tape recording in which Myers complained bitterly that he was being forced to implicate
McMillian—whom he had never met—and to confess to participating in a crime that neither of
them had any role in.
Thirty-five percent of codefendant confessions in exonerations were obtained by official
misconduct (110/315), 5% of all exonerations (110/2,400).
In 44% of exonerations with codefendant confessions that were obtained by misconduct, the
exonerees themselves also falsely confessed (48/110). For example, in 1993, Daniel Villegas, a
16-year-old with learning disabilities, confessed to a double drive-by murder in El Paso, Texas
after an interrogation in which he was handcuffed to a chair for hours and threatened with
violence. In addition, two other teenagers confessed after their own abusive interrogations, and
implicated Villegas; they were both charged with murder but charges were later dismissed.
Villegas was convicted in 1995 and sentenced to life in prison, based on his own confession as
well as those of the other two suspects. (Yet another youth had falsely confessed earlier to the
same crime, after a similar interrogation, but was never charged.) Villegas was exonerated in
2018, after the nature of the interrogations in the case was exposed, and the likely real criminals
were identified.
As with all codefendant confessions in exonerations, those obtained by misconduct were
concentrated among murder cases, which make up three-quarters of the total (83/110), and in
Chicago. More than a quarter of all murder exonerations in Chicago included codefendant
confessions obtained by misconduct (31/118). See Table 16.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 63 • National Registry of Exonerations • September 1, 2020
Table 16: Proportion of Exonerations with Codefendant Confessions Obtained by
Misconduct, in Chicago and Elsewhere
Exonerations
in Chicago
Exonerations
Elsewhere
ALL
JURISDICTIONS
Murder
26%
(31/118)
7%
(52/790)
9%
(83/908)
All other Crimes
1%
(1/112)
2%
(26/1,380))
2%
(27/1,492)
ALL CASES
14%
(32/230)
4%
(78/2,170)
5%
(110/2,400)
Three-quarters of cases with coerced codefendants confessions were multi-defendant
exonerations (82/110). The prototypical exoneration with a codefendant confession that was
obtained by official misconduct is a multi-defendant murder exoneration in which the
codefendant was convicted along with the exonerees they falsely implicated. The Ford Heights
Four exonerations are a good example:
In May 1978, a young white couple were abducted in a Chicago suburb. Their
bodies were found the following day; both had been shot, and woman had been
gang-raped. Acting on a tip, the police arrested four Black men— Verneal
Jimerson, Dennis Williams, Kenneth Adams, and Willie Rainge. They also
questioned Paula Gray, a 17-year-old with a mild intellectual disability.
After two days of questioning, Gray told a grand jury that she held a disposable
cigarette lighter burning while the four men raped the female victim repeatedly,
and that she saw Williams shoot both victims with a .38-caliber pistol. A month
later, she recanted her story at a preliminary hearing, and testified that she had
been drugged and that the police walked her around the crime scene and told her
what to say.
Gray herself was charged with murder and perjury and brought to trial jointly
with three of the male defendants—Adams, Rainge, and Williams. All four were
convicted; Williams was sentenced to death, Rainge to life, Adams to 75 years and
Gray to 50 years.
After Williams and Rainge won new trials in 1982, prosecutors cut a deal with
Gray under which she was released in exchange for testifying against Williams
and Rainge at their retrial, and against Jimerson, who had not yet been tried. All
three were convicted; Williams was again sentenced to death, as was Jimerson,
and Rainge to life.
Jimerson, Williams, Adams and Rainge—the Ford Heights Four—were all
exonerated in 1996, after DNA evidence proved they had not raped the female
victim, the real criminals were identified and confessed, and Gray once more
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 64 • National Registry of Exonerations • September 1, 2020
recanted her confession and testimony against them. Gray herself was exonerated
in 2002 when Illinois Governor George Ryan pardoned her and several other
defendants who had falsely confessed in Cook County.
The overwhelming evidence of innocence that exonerated the Ford Heights Four in 1996 also
proved Paula Gray’s innocence of murder, and made it equally clear that she had committed
perjury. And yet she was not exonerated for another six years—perhaps because nobody was
motivated enough to help a defendant whose lies, however coerced, led to several false
convictions, and who had already been released from prison. Ralph Myers—the codefendant
whose false confession put Walter McMillian on death row—fared worse. He served his 30-year
sentence and was never exonerated, even though he could not possibly be guilty of helping
McMillian commit a crime McMillian had nothing to do with.
In addition to Myers, we know of two other codefendants who falsely confessed and implicated
an exoneree, were charged and convicted themselves, and were never exonerated despite clear
evidence of innocence.
112
There may be others; we don’t know what happened to many of the
confessing codefendants. On the other hand, in 62 exonerations, codefendants like Christopher
Ochoa, who implicated the exonerees in a confession obtained by misconduct, were themselves
exonerated.
112
Timothy Brown and David Gladden.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 65 • National Registry of Exonerations • September 1, 2020
VI. Fabricated Official Evidence
Police and other law enforcement officers are the most important sources of evidence in
criminal prosecutions. Usually they write reports or testify about investigations that gathered
evidence from other sources, but in a minority of cases they actually witnessed the crimes, or
discovered damning evidence against the defendant—or claimed to have done so.
Lies by law enforcement officers contributed to many of the false convictions in the Registry.
Often, they lied about the conduct of the investigation—who they talked to, when, what was said
and under what circumstances, and so forth. In about 10% of exonerations, however, police and
other officials made up incriminating evidence by lying about their own observations. These
cases come in three groups:
• Law enforcement officials—police officers or forensic analysts who work for police
agencies—sometimes testify as experts in forensic science and explain the significance of
trace evidence and other physical evidence connected to the crime. In about 3% of
exonerations (75/2,400), they deliberately presented false evidence against the
defendants or concealed or distorted true evidence that might have cleared them.
• In 5% of exonerations, officers lied as ordinary lay witnesses and said that they saw the
defendants commit crimes that never happened—possess illegal drugs or assault the
officers who testified, when in fact the drugs were planted and the assaults were made
up.
• In about 2% of exonerations, police made up confessions by the defendants (36/2,400).
1. FORENSIC FRAUD
“Forensic evidence” is a collection of different types of expert evidence that are used in criminal
cases for a variety of purposes: to identify people using biological samples or traces they left
behind (DNA, fingerprints, blood type, hair, bitemarks); to identify objects from traces or from
marks they left (tool marks, fibers, tire treads, shoe prints); to describe the composition of
substances (chemical analyses of drugs or poisons); to describe pathologies (illness, injury,
trauma); or to assign causes to events (deaths, injuries, fires).
False or misleading forensic evidence contributed to the false convictions of a quarter of known
exonerees (590/2,400). In most of those cases, forensic experts did not engage in misconduct
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 66 • National Registry of Exonerations • September 1, 2020
but rather made mistakes or relied on scientific procedures or beliefs that had been, or later
were discredited. In other cases they did commit misconduct, but not fraud; most often, they
concealed forensic information that might have helped the defendant.
Forensic fraud, as we use the term, is an aggravated form of intentional misconduct by state
actors: the deliberate falsification of forensic evidence by a law enforcement officer in order to
help convict a criminal defendant. We do not count fraud by privately employed forensic
experts, nor cases in which misstatements about forensic evidence might have been deliberate
but we don’t know enough to be confident.
We know of forensic fraud in 3% of exonerations (75/2,400). In 83% of those cases the fraud
was committed by a forensic analyst, including two cases in which a prosecutor also participated
in the fraud (62/75). In 16% of the cases, forensic fraud was committed by a police officer
(12/75), also including two cases in which a prosecutor participated. And in one case, a
prosecutor committed forensic fraud on his own.
Reviewing these cases, several themes emerged:
• False forensic “matches.” In more than a third of the forensic fraud cases analysts and
other witnesses falsely linked the defendants to crime scene evidence (27/75), most often
by claiming that their hair “matched” hair from the crime scene (10/27). For example, at
Glen Woodall’s 1987 rape trial, West Virginia State Police Trooper Fred Zain testified
that it was "highly unlikely" a hair found in the victim’s car could have come from any
source but Woodall’s blond beard—despite the fact that in a concealed report he wrote
three months earlier, Zain described that sample as a pubic hair.
In several other cases, the forensic experts falsely reported that blood or semen from the
crime scene had the same blood type as the defendants’ blood (7/27), or that crime scene
DNA matched the defendant (3/27). At Bernard Webster’s 1982 rape trial in Maryland,
for example, the forensic analyst testified that the rapist had type A blood, as did
Webster; in fact, there was no way to determine the blood type of the rapist from the
mixed sample that was available—as that same expert had testified earlier.
In a few cases forensic analysts testified falsely that bitemarks on the victims matched
the defendants’ teeth (4/27), or that tools owned by the defendant damaged the victim’s
property (2/27), or that a fiber found near the crime was linked to the defendant (1/27).
• Concealed tests that excluded the defendant from suspicion. In about a quarter of the
cases, forensic witnesses falsely reported that the defendants might have been the source
of crime-scene blood, semen or fingerprints, while concealing forensic tests that had
already shown that was impossible (17/75). For example, in 1988 Calvin and Larry
Ollins, Omar Saunders and Marcellius Bradford, ages 14 through 18, were convicted of
rape and murder in Chicago. At their trial, forensic analyst Pamela Fish testified that
semen found on the victim’s body and undergarments could have come from three of the
four defendants. In fact, she knew from blood tests that she had conducted and hidden
that none of the defendants could have been the source of the semen.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 67 • National Registry of Exonerations • September 1, 2020
• Planted evidence. In four exonerations, police planted the evidence that was tested by
forensic analysts. In three cases from upstate New York, state troopers planted
defendants’ fingerprints on objects from the crime scene, and then “collected” and
“analyzed” those prints themselves;
113
and in Desmond Ricks’ case in Detroit, officers
substituted a slug fired from a gun found in Desmond’s house for the slug that killed the
victim in order to lead another officer and an independent forensic analyst to falsely
conclude that Desmond’s gun fired the fatal shot.
• Various lies. The remaining third of forensic fraud cases are a mixed bag (27/75). In
three cases, for example, the forensic analyst lied and said that there was insufficient
trace evidence for forensic testing when in fact there was sufficient evidence to test and,
when tests were eventually conducted, they cleared the defendant. And in a two-
defendant murder case, an examiner testified falsely that there was no blood spatter on
the pants of the primary alternate suspect; testing 18 years later found that there was a
spatter of blood on the pants—the victim’s blood.
At William Dillon’s murder trial in Florida in 1981, a dog handler who was later labeled
“a charlatan” by the Arizona Supreme Court testified to a rigged “dog sniff identification”
that appeared to implicate Dillon. And in Adolph Munson’s 1984 murder case in
Oklahoma, Medical Examiner Ralph Erdmann first testified that the victim was killed by
a large caliber bullet; then the prosecutors told him that they needed it to be a small .22
caliber bullet, so he revised his testimony and said that maggots had enlarged the .22
caliber holes.
A third of known incidents of forensic fraud involved notorious bad actors (25/75). Dr.
Erdmann, for example, eventually lost his medical license and pled guilty to several felonies
after it was discovered that he had filed reports on hundreds if not thousands of autopsies that
he never conducted. Pamela Fish—who is responsible for 10 of the forensic fraud cases we list—
was removed from criminal case work by the Illinois State Police after several cases in which her
perjury was exposed ended in exoneration. Fred Zain, who presented fraudulent evidence in
seven trials that led to exonerations, was the subject of major investigations for massive patterns
of forensic fraud in both West Virginia and Texas, and was eventually indicted for fraud. And
Joyce Gilchrist, who committed forensic fraud in three exonerations (and provided false or
misleading evidence in three others) was ultimately investigated by the FBI and fired.
114
113
In the first case, State Police Trooper David Harding testified at Shirley Kinge’s arson trial in Tompkins County,
New York in 1990, that he found Kinge’s fingerprints on a gasoline can at the site of the fire. In 1992, Kinge was
exonerated after Harding pled guilty to perjury and admitted planting fingerprint evidence in Kinge’s case and others;
he was sentenced to 4 ½ years in prison. Two years later, William Labolt Jr. was exonerated after an investigation
revealed that a Lieutenant in Harding’s unit had planted a fingerprint that led to his conviction for burglary; and in
1995, Mark Prentice was exonerated when Trooper Harding admitted to planting the fingerprint that led to his
conviction for robbery.
114
For additional references on Pamela Fish, Fred Zain and Joyce Gilchrist, See infra Sections X.2.c, XII.1.c,
XII.2.b.i, and XII.2.c.ii(b).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 68 • National Registry of Exonerations • September 1, 2020
No doubt repeat offenders contribute more than their share of misconduct in this setting, as in
others. But we are also more likely to learn about incidents of forensic fraud by serial perjurers
who eventually get caught—after which their entire work history is likely to be reexamined—
than about similar acts by colleagues with more discretion or better luck.
2. FAKE CRIMES
As ordinary lay witnesses, police and other law enforcement officers manufactured false
evidence against innocent defendants in two ways: (i) In about 4% of exonerations they planted
evidence at the scene of the crime and claimed to have found it there; in all but a few of those
cases, they planted evidence of crimes that never occurred. (ii) In about 1% of the cases officers
falsely claimed that the defendants assaulted them, usually to cover up for their own violence
against the same defendants.
a. Planted evidence
Planting evidence is a devastating form of misconduct. It occurred in hundreds of group
exonerations in which corrupt police officers planted drugs on innocent defendants. As we’ve
discussed, most group exonerations are not included in the Registry, although we do include the
Sergeant Watts group exonerations in Chicago—66 so far—and expect to add others.
115
In addition, police officers planted evidence (usually drugs) in about 1% of non-group
exonerations that we know of (18/1,361). For example, at Maurice James’s 1990 drug-sale trial
in Rochester, New York, a police officer testified that he found a marked $20 bill on James,
which was supposedly used to pay for drugs. James spent two years in prison; he was
exonerated after a federal grand jury indicted the officers involved for misconduct in several
cases, and one of the officers admitted that he planted the marked bill. In other cases, officers
planted illegal drugs on the defendants, or in their cars or homes, or simply booked the drugs in
evidence and falsely claimed that they found them in the defendants’ possession.
We know of only four non-drug exonerations in which police planted evidence, all cases in which
real crimes did occur. In three, New York State troopers planted the exonerees’ fingerprints. We
discussed them in more detail above, as forensic fraud.
116
And Marvin Thomas’s 1987 murder
conviction in West Virginia was based in part on a human bloodstain and hair that were planted
in the exoneree’s car.
Planted evidence is very hard to detect. If an officer says “I found this bag of white powder (or
this hair, or $20 bill) on the defendant,” how would we ever know that he actually found it in the
trunk of his own car? Certainly few judges or prosecutors (nor most defense lawyers) will believe
the defendant’s claim that he was framed.
Almost all the planted evidence cases we know about occurred in the context of drug crime
enforcement. Illegal drug distribution in the United States is a major business that police
115
See supra Section III.3.i.
116
See supra Section VI.1.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 69 • National Registry of Exonerations • September 1, 2020
attempt to prevent, or at least control, with little or no success. Sometimes they succumb to the
temptation to simply arrest people they believe sell drugs and make up the evidence. It’s easily
done. And sometimes—as with Sergeant. Watts and his crew—they decide to make money
themselves by extorting or stealing from suspects, or selling drugs they seize. Along the way they
may find it useful to frame dozens or hundreds of innocent defendants.
If a widespread pattern of corruption of that sort goes on long enough, it may eventually be
detected and the dishonest officers may be brought to justice—after which, many of the
convictions they obtained may be reinvestigated, and innocent defendants exonerated. That’s
what happened after Sergeant Watts was convicted in federal court in Chicago, and in other
group exonerations elsewhere.
On a smaller scale, that’s what happened in the New York State trooper fingerprint planting
cases, which involved dozens of defendants in addition to the three who were exonerated. A
special prosecutor who investigated the scandal wrote:
“Some members of the Identification Unit were so careless with their fabrications they
left … 'practice' fabrications behind in the actual case files in which evidence had been
fabricated and used in criminal prosecutions. . . . [This] strongly suggests that the
individuals fabricating evidence on a routine basis had no fear of discovery and … took
few steps to cover their tracks.”
117
On the other hand, if an officer lies about where he found a gun or a jacket or a quantity of drugs
in one or several unrelated investigations, we’ll probably never know. We only know about the
planted bloodstain and hair in Marvin Thomas’s case because they appeared miraculously after
the car had been searched twice, by local police and then by the FBI, and nothing of value was
found. If they had been planted before the first search, Thomas would probably still be in prison.
b. Phony assaults
In an important minority of criminal cases, police officers are essential fact witnesses because
they are the victims of violent crimes—or claim to be, as in several exonerations in the Registry.
For example, Wassillie Gregory, as we saw,
118
was exonerated in Alaska in 2014 when a
surveillance video showed that Gregory was the victim of an assault by the police officer who
arrested him, rather than vice versa.
In 1970, Malcolm Emory was walking from the library to his dormitory at Northeastern
University in Boston when he passed by an anti-Vietnam war demonstration. He was arrested by
an officer who later testified that Emory had a brick in one hand and a concrete block in the
other, and struck the officer in the chest with the brick. Emory was convicted of assault and
battery with a dangerous weapon on a police officer. He was exonerated 20 years later, after he
117
Nelson E. Roth, "The New York State Police Evidence Tampering Investigation," confidential report to the
governor of New York (Ithaca, Jan. 20, 1997) 289. See also Richard Perez-Pena, Supervision of Troopers Faulted In
Evidence-Tampering Scandal, The New York Times, Feb. 4, 1997, page B.1.
118
See supra Section I.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 70 • National Registry of Exonerations • September 1, 2020
managed to find an unpublished newspaper photograph of the demonstration that clearly shows
Emory, as he testified, being beaten and wrestled to the ground by police as he clutches several
books, with no brick or concrete block in sight.
In 2006, Daniel Gonzalez, Jonathan Dominguez and Jeffery Funes were convicted in Los
Angeles for supposedly throwing a beer bottle that shattered and injured a police officer. After a
defense attorney discovered that the bottle in question was still intact, the trial judge declared “I
have been flat out lied to,” vacated the convictions and dismissed the charges.
In the cases we’ve mentioned, the defendant was exonerated by physical evidence that
contradicted the officer/victim’s false story. That seems to be necessary in order to win an
exoneration in a case where the supposed victim is a police officer: the undamaged beer bottle in
the Gonzalez, Dominguez and Funes case; photographs or videos in the others. The only
exception is the case of Jeffrey Santos, who was exonerated in 2004 after the officer he was
convicted of assaulting—a captain at the main New York City jail—was investigated and
disciplined for a pattern of severe assaults on several prisoners, all similar to the one Santos
described at his trial.
Judging from these cases, if an innocent defendant is convicted because a police officer who beat
him up lies and says that the defendant was the attacker, he will only be exonerated if he can
produce physical evidence that contradicts the police version of events—testimony by civilians
won’t do it—unless the officer is caught doing the same thing repeatedly.
3. FABRICATED CONFESSIONS
Twelve percent of the exonerations in the Registry include false confessions (292/2,400.) We
discuss many of these exonerations in Section V, those in which the innocent defendants really
did confess to crimes they did not commit, usually under severe pressure from police officers,
ranging up to torture. But we define a “confession” as any statement that is treated as such by
law enforcement, and in 12% of exonerations with false confessions the defendants did not
confess at all (36/292). Instead, the police fabricated confessions. Needless to say, none of these
“confessions” were recorded or confirmed by any other means.
In several cases the exonerees signed “confessions” but did not know what they were signing.
Omar Aguirre, for example, who was interrogated in a murder investigation in Chicago in 1997,
did not speak or read English. He was told that he had to sign a document in English in order to
go home. It was not read to him, he signed it—and it was introduced at trial as his confession.
Meredith Town, on the other hand, did know that he was signing a confession—if not the details
of what it said—but he signed it anyway because he was threatened with violence if he refused.
In the great majority of exonerations with fictional confessions, the police fabricated oral
confessions out of whole cloth (31/36). Robert Hays, for example, called a police detective in Las
Vegas in June 1992 because he had heard that his eight-year old daughter had told the police
that he had raped her. Hays denied the charge—which his daughter later recanted, many times,
and said her mother had forced her to make—but a detective testified at Hays’s trial that he
confessed to the crime. Hays was convicted and sentenced to life in prison. In 2007, a federal
judge found that the detective lied and made up the confession and that Hays was innocent, and
ordered that Hays be released immediately.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 71 • National Registry of Exonerations • September 1, 2020
Debra Milke’s case is probably the best-known fabricated confession in the Registry:
In 1990, James Styers and Roger Scott were convicted of murdering Debra
Milke’s four-year-old son, Christopher, in Phoenix, Arizona. Both admitted to
participating in the murder and both were sentenced to death. Debra Milke
herself was also tried for conspiring with Styers and Scott to kill her own son in
order to collect a $5,000 life insurance policy, and she too was convicted and
sentenced to death.
The only evidence connecting Milke to the murder was a confession that was
supposedly obtained by Detective Armando Saldate, Jr. Saldate testified that
Milke flashed her breasts, offered him sex if he would not arrest her—and then
admitted that she conspired to kill her own son for insurance money. Milke
denied that any of that happened. There was no written confession, nobody else
witnessed it, and, although Saldate’s supervisor specifically instructed him to
record the interrogation, he did not do so.
Milke was exonerated 25 years later, after her attorneys uncovered Detective
Saldate’s extensive history of deception and perjury. A 2003 opinion by the
United States Court of Appeals for the Ninth Circuit outlined that history: “[A]
five-day suspension for taking ‘liberties’ with a female motorist and then lying
about it to his supervisors; four court cases where judges tossed out confessions
or indictments because Saldate lied under oath; and four cases where judges
suppressed confessions or vacated convictions because Saldate had violated the
[Constitution] in the course of interrogation. … And it is far from clear that this
reflects a full account of Saldate’s misconduct as a police officer.”
119
Fabricated confessions are somewhat less common in cases with coercive misconduct in
interrogations than in those without, 8% vs. 17%, both in Chicago (12% to 29%)—which has
extraordinarily high rates of false confessions and misconduct in interrogations
120
—and in the
rest of the country (7% to 15%). See Table 17.
119
Milke v. Ryan, 711 F.3d 998, 1003 (9th Cir. 2013).
120
See supra Section V.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 72 • National Registry of Exonerations • September 1, 2020
Table 17: Proportions of Cases with Fabricated Confessions, by Misconduct in
Interrogation and Jurisdiction
Exonerations
in Chicago
Exonerations
Elsewhere
ALL
EXONERATIONS
Fabrication after Coercive
Misconduct in Interrogation
12%
(7/58)
7%
(7/107)
8%
(14/165)
Fabrication without Coercive
Misconduct in Interrogations
29%
(5/17)
15%
(17/110)
17%
(22/127)
ALL FABRICATED
CONFESSIONS
16%
(12/75)
11%
(24/217)
12%
(36/292)
Perhaps the reason for this difference is that coercion (especially violence) and fabrication are
usually alternative means to the same end: producing a “confession” to use in court. Officers
who are reluctant to use violence may be driven to make up confessions that never happened,
while those who are proficient at violence, or other types of coercion, usually get suspects to say
or sign what that want, and have no need to cook up confessions from scratch.
But when coercion—extreme coercion—fails to produce results, fabrication may follow:
In January of 1982, a prosecution witness in an upcoming murder trial in
Chicago was shot to death. Detectives arrested Melvin Jones while he was
babysitting his girlfriend’s child, and found a gun in a drawer in the child’s room.
Police Lt. Jon Burge and his officers abused Jones for days. They administered
electrical shocks to his feet, thighs and penis, and hit him on the head with a
stapler—but he didn’t confess. After four days, they gave up and charged Jones
with unlawful use of a weapon, for which he was acquitted.
A few months later, police arrested Jones for an unrelated triple homicide, but
never charged him with those murders. Instead, after another interrogation, they
charged him with the earlier murder. Detectives said Jones confessed to that
murder because he believed he had gotten away with it and could no longer be
charged.
There was no recording of the confession, and no written record. At one point,
the police called in a prosecutor to “take the confession,” but Jones told the
prosecutor he had not confessed. The officers later said that after the prosecutor
left, Jones confessed to them again, orally—but they did not record it.
Jones was convicted based on the detectives’ testimony that he confessed—which
he denied—and testimony from an officer that an eyewitness who had identified
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 73 • National Registry of Exonerations • September 1, 2020
Jones had passed a polygraph test.
121
Jones was acquitted on retrial after the
“eyewitness” testified she had not been polygraphed and had never identified
Jones. There was no evidence that she had—other than the officers’ words.
Jones is not alone. In more than 70% of exonerations with fabricated confessions that were
preceded by abusive but unsuccessful interrogations, those interrogations included actual or
threatened violence (10/14).
In other cases, misconduct in interrogation and fabrication are two parts of a single course of
action. Charles Johnson, for example, was interrogated for a murder he did not commit, in
handcuffs, for twelve hours. Chicago detectives told him that if he did not “tell the truth” he
would “never see his family again,” would get the death penalty, and would be raped in prison.
At the end of that ordeal, he signed—without reading—a few documents that were said to be
“release papers.” One was a confession written by the police. When an officer read it to him, he
said “I didn’t tell you that, I told you nothing”—but it sent him to prison for 19 years before he
was exonerated.
121
The eyewitness did not testify at that trial. The conviction was vacated in part because the police officer’s
testimony about her supposed identification was inadmissible hearsay.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 74 • National Registry of Exonerations • September 1, 2020
VII. Concealing Exculpatory
Evidence
In 1987, in Williamson County, Texas, Michael Morton was convicted of
murdering his wife, Christine, by bludgeoning her to death in their bed. He was
exonerated in 2011, 24 years later, when DNA tests were conducted on a bandana
found 100 yards from the Mortons’ house.
122
That testing found Christine
Morton’s blood on the bandana together with DNA from a convicted felon, who
later bludgeoned another woman to death in her bed in nearby Travis County in
1988, while Michael Morton was in prison. For years, the prosecution had
opposed testing the bandana.
In the process of reinvestigating the case, lawyers for Morton also found the
following items of information that were well known to police and prosecutors
but had been concealed from the defense at trial: (i) Neighbors told police that in
the days before the murder they saw a man park a green van, get out, and walk
into the woods behind the Mortons’ house. (ii) Days after the murder, while
Morton was in custody, someone attempted to use a credit card belonging to his
wife at a store in San Antonio, Texas. (iii) Weeks later, a $20 check written to
Christine that had been in her missing purse was cashed with a forged signature.
(iv) Christine’s mother reported to the police that the Mortons’ three-year-old
son, her grandson, had told her that “a monster” killed his mother when “Daddy”
was not there.
Michael Morton’s exoneration is one of the best known in the Registry. The misconduct was
blatant, and the consequences horrific. If the police had followed the leads they concealed rather
than hiding them in order to convict Morton, they might have identified the real killer before he
murdered another woman in a similar manner—and of course, spared Morton and his family
from the horror of his false conviction. Morton’s case has been the subject of at least one book,
123
122
See Annie Daniel & Johnathan Silver, From Michael Morton's Wrongful Conviction to Exoneration (Timeline), The
Tex. Trib. (Aug. 13, 2016).
123
Michael Morton, Getting Life: An Innocent Man's 25-Year Journey from Prison to Peace (2014).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 75 • National Registry of Exonerations • September 1, 2020
an award winning-documentary,
124
and countless podcasts and news stories. In 2013, the Texas
Legislature passed the “Michael Morton Act” which requires prosecutors in Texas to provide
extensive information about criminal investigations to the defense before trial.
125
Quedillis Ricardo Walker’s case, by contrast, is obscure.
In 1991, Walker went on trial for the murder of a former girlfriend who was found
dead in her apartment in Cupertino, California—bound, gagged and stabbed. The
main evidence against Walker came from a codefendant, Rahsson Bowers, whose
fingerprints were found on duct tape that was used to bind the victim. Bowers
pled guilty to second-degree murder during the trial, and testified that he helped
Walker kill the victim because he was afraid of him.
In addition, Sarah Dunbar, an acquaintance of Walker’s, testified that Walker had
threatened her with a knife and a gun, and that she had bought him a pair of
gloves that were similar to a piece of a glove found near the body. Walker was
convicted of first-degree murder.
Twelve years later, Dunbar admitted that she had lied in her testimony against
Walker in return for a deal on pending drug charges—a deal that was concealed at
trial. Walker was exonerated after several witnesses identified a different man as
Bowers’ accomplice, and that man’s DNA was found at the scene of the killing.
The concealed evidence that we know about is much more mundane in Walker’s case than in
Morton’s. There was nothing that directly indicated that someone else was the killer, just a
hidden reason to distrust a witness against Walker. But the rest of the case against Walker was
weak; it consisted of little more than the testimony of an admitted murderer who got years off
his sentence in return. Hiding the deal with Dunbar might well have made the difference
between conviction and acquittal. In any event, the outcome was the same: an innocent man was
convicted of murder.
Note the qualification in the previous paragraph: “The concealed evidence that we know
about….” The essence of this type of misconduct is the effort to conceal; we only know about it
when that effort eventually fails. Exculpatory evidence that we don’t know about may also have
been concealed at Quedillis Walker’s trial—and at Michael Morton’s—not to mention in many
other cases where no concealed evidence has come to light.
Even so, concealing exculpatory evidence is the most common type of official misconduct that
we report. It was done in 44% of the exonerations in the Registry (1,064/2,400)—that we know
about.
124
An Unreal Dream: The Michael Morton Story (Al Reinert) (2013).
125
Legislative Reference Library of Texas, SB 1611, 83rd Regular Session; see Brandi Grissom, Perry Signs Michael
Morton Act, The Texas Tribune, May 16, 2013.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 76 • National Registry of Exonerations • September 1, 2020
1. THE DUTY TO DISCLOSE EXCULPATORY EVIDENCE
In 1963, the Supreme Court held in Brady v. Maryland that “suppression by the prosecution of
evidence favorable to the accused … violate[s] due process where the evidence is material either
to guilt or to punishment.”
126
Brady was decided at a time when pre-trial discovery in criminal
cases in the United States was in its infancy. It quickly became—and has remained—the central
point of reference for discussions of the government’s duty to disclose exculpatory information
to criminal defendants. Failure to do so is routinely described as a “Brady violation” and the
information at stake is often described as “Brady material.”
That framework is unfortunate and misleading for two reasons: (i) The requirements for
disclosure under Brady are limited and incoherent. (ii) Other rules—including rules governing
discovery in criminal cases and rules of professional responsibility that govern the conduct of
prosecutors—require the disclosure of evidence favorable to criminal defendants in broader and
clearer terms.
a. Brady v. Maryland and the “Materiality” Requirement
Under Brady, the obligation to disclose is limited to information that is “material” to the
defendant’s guilt or punishment. “Materiality” is a term that means different things in different
legal contexts. As used in Brady itself, the most likely meaning was one that was current at that
time as a requirement for admissibility of evidence: An item of evidence is “material” if some
proposition that it tends to prove matters to the resolution of the case at hand; otherwise, the
evidence is inadmissible because it’s “immaterial.”
127
In 1985, however, in United States v. Bagley,
128
the Court introduced a new definition of
“materiality” as applied to Brady violations. It held that under Brady, exculpatory evidence “is
material only if there is a reasonable probability that, had the evidence been disclosed to the
defense, the result of the proceeding would have been different.” This sounds like a “harmless
error” rule, a judgment that sometimes a constitutional violation does not require reversal of a
conviction because the outcome of the case would have been the same without the violation. But
it’s not. If the evidence is “immaterial” there is no obligation to disclose it; failure to do so is not
a forgivable violation of the constitution rule but, no violation at all.
The materiality requirement of the Brady rule, as interpreted by Bagley, is incoherent.
129
First,
how can anyone know whether a jury would have decided a case differently if it had additional
evidence? This is a problem with real “harmless error” rules as well, which the Court has
attempted to address by setting a very high standard for a finding of harmlessness. In general, a
violation of the constitution is only “harmless” if a court reviewing all the evidence in the case
126
Brady v. Maryland, 373 U.S. 83, 87 (1963).
127
See United States v. Bagley, 473 U.S. 667, 703, n.5 (1985) (Marshall, J., dissenting). Under modern codified
rules of evidence, materiality in this sense has been absorbed into the more general concept of “relevance.” See,
e.g., Notes of Advisory Committee on Proposed Rules, Fed. R. Evid. 401.
128
United States v. Bagley, 473 U.S. 667 (1985).
129
See Bagley, 473 U.S. at 687-709 (Marshall, J., dissenting).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 77 • National Registry of Exonerations • September 1, 2020
concludes “beyond a reasonable doubt” that the violation did not affect the outcome.
130
Bagley
goes far in the opposite direction.
Failure to disclose exculpatory evidence is only subject to the rule if there is “a reasonable
probability” that disclosure would have changed the outcome. Under the best of circumstances,
this rule will lead to unpredictable and inconsistent outcomes—and circumstances are often far
from the best.
Judges don’t like to reverse criminal convictions. They believe the defendants are probably
guilty, and reversals lead to more work: retrials of the defendants (unless charges are dismissed
outright), more appeals by other convicted defendants who draw hope from the reversals, and (if
they’re consistent) more reversals in future cases. As Judge Kozinski of the United States Court
of Appeals for the Ninth Circuit pointed out, courts can avoid these dangers by requiring an
“impossibly high” standard for materiality before a conviction will be reversed.
131
Second, how can a prosecutor be expected to make that judgment?
132
It seems illogical, but
under Bagley the prosecutor is the person who must determine whether the exculpatory
evidence is “material.” If she concludes that there is no “reasonable probability” that the
evidence would change the outcome, then she need not tell anybody about it, and, very likely,
neither the defense nor the courts will ever learn that the evidence exists. But the prosecutor is
an adversary in the case; her job will always be easier if she decides there is no obligation to
disclose, and even if she tries her best to make that decision in good faith, she is biased. Trial
lawyers are generally optimistic about their prospects at trial. It’s a posture that helps them win
by projecting confidence—especially if they actually do feel confident that they will win, whether
or not they are right. But that same optimism may also lead a prosecutor to be confident that
exculpatory evidence that only she knows about would not affect the jury’s decision.
Third, how can a prosecutor—or anybody—possibly make that determination before trial? For
the disclosure required by Brady to be effective, it must be made before trial, or perhaps at the
very outset of a trial. At that point, the prosecutor cannot possibly know what evidence the jury
will hear. Her own witnesses may change their stories, or add new information, or be damaged
on cross-examination—or in some cases, fail to testify at all—and she will at best only be able to
guess whether the defendant himself will testify, or who else will testify for the defense, or what
the defense witnesses will say and how they will come off. She certainly can’t know how the
defense might use the exculpatory evidence if revealed: Would it open up new avenues of
investigation? Would it lead to new witnesses the prosecutor does not know about? In any case
130
Chapman v. California, 386 U.S. 18 (1967).
131
United States v. Olsen, 737 F.3d 625, 633 (9th Cir. 2013).
132
See David A. Sklansky, The Progressive Prosecutor’s Handbook, 50 U.C. Davis L. Rev. Online 25, 34 (2017).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 78 • National Registry of Exonerations • September 1, 2020
where guilt or innocence is truly in dispute, nobody can say at the outset of the trial how strong
the evidence of guilt will be at the end.
133
Unsurprisingly, the “materiality” requirement that Bagley and other cases applied to the
requirement of disclosure of exculpatory evidence under Brady has been the subject of extensive
criticism by legal scholars.
134
b. Other Legal Bases for the Duty to Disclose
i. Professional Responsibility
The Brady decision was not about misconduct by the prosecutor or any other government
official. The issue, rather, was whether the defendant received a fair trial. As the Supreme Court
explained in 1988, in Youngblood v. Arizona, “The Due Process Clause of the Fourteenth
Amendment, as interpreted in Brady, makes the good or bad faith of the State irrelevant when
the State fails to disclose to the defendant material exculpatory evidence.”
135
This does not mean
that failing to comply with Brady is not misconduct—it usually is—but rather that this
constitutional doctrine is not designed to address the misconduct but to remedy violations of
due process of law.
There are, however, other rules that regulate the conduct of prosecutors directly, and specifically
require them to disclose exculpatory evidence to the defense regardless of “materiality.”
Violating those rules is misconduct by any definition.
136
Rule 3.8 of the American Bar Association (“ABA”) Model Rules of Professional Responsibility
provides in part:
The prosecutor in a criminal case shall:
…
(d) make timely disclosure to the defense of all evidence or information known to the
prosecutor that tends to negate the guilt of the accused or mitigates the offense, and, in
connection with sentencing, disclose to the defense and to the tribunal all unprivileged
133
See John G. Douglass, Fatal Attraction? The Uneasy Courtship of Brady and Plea Bargaining, 50 Emory L. J. 437,
516 (2001) (“Brady suffers from a severe case of bad timing. Brady establishes a retrospective standard for defining a
prospective obligation.”)
134
E.g., David A. Sklansky, The Problems with Prosecutors, 1 Ann. Rev. Criminology 451-69 (2018) at 456-58; David
A. Sklansky, The Progressive Prosecutor’s Handbook, 50 U.C. Davis L. Rev. Online 25, 33-36 (2017); Bennett
Gershman, Litigating Brady v. Maryland: Games Prosecutors Play, 57 Case W. Res. L. Rev. 531–66 (2007); Alafair
Burke, Revisiting Prosecutorial Disclosure, 84 Ind. L .J. 481 (2009).
135
Arizona v. Youngblood, 488 U.S. 51, 57 (1988).
136
For an excellent summary of the professional and criminal procedure rules that require prosecutors to disclosed
exculpatory evidence to criminal defendants, See Marc Allen, Non-Brady Legal and Ethical Obligations on
Prosecutors to Disclose Exculpatory Evidence, National Registry of Exonerations (July, 2018).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 79 • National Registry of Exonerations • September 1, 2020
mitigating information known to the prosecutor, except when the prosecutor is relieved
of this responsibility by a protective order of the tribunal….
Every state has adopted section (d) of Model Rule 3.8, or a similar standard; none include any
requirement of “materiality.”
137
ii. Pretrial Discovery
The ABA has also issued a set of more specific Criminal Justice Standards for the Prosecution
Function that are intended to be read as best practices for prosecutors and are meant to
supplement the Model Rules for Professional Conduct as they bear on prosecutors.
138
Standard
3-5.4 deals with disclosure of exculpatory evidence, and explicitly rejects materiality as a
condition for disclosure: “Before trial of a criminal case, a prosecutor should make timely
disclosure to the defense of [exculpatory] information … that is known to the prosecutor,
regardless of whether the prosecutor believes it is likely to change the result of the
proceeding.”
139
The ABA Criminal Justice Standards also include a section on pre-trial discovery. Standard 11-
2.1 of the Standards on Discovery provides that:
a) The prosecution should, within a specified and reasonable time prior to trial, disclose
to the defense the following information…
(viii) Any material or information within the prosecutor’s possession or control which
tends to negate the guilt of the defendant as to the offense charged or which would tend
to reduce the punishment of the defendant.
State law on pre-trial discovery in criminal cases is uneven.
140
Thirty-four states have discovery
rules that require prosecutors to disclose exculpatory evidence to defendants in all cases or
alternatively upon motion by the defense. None of these rules include materiality as a condition
for this obligation, although Louisiana and New Hampshire refer to Brady in their definitions of
exculpatory evidence—which could be interpreted as incorporating Brady’s materiality
137
See id. at 5-7, Table 1.
138
Criminal Justice Standards for the Prosecution Function § 3-1.1(b) (Am. Bar Ass’n 2017).
139
Id. at § 3-5.4(c). Standard 3-5.4 includes other relevant language as well:
a. After charges are filed if not before, the prosecutor should diligently seek to identify all information
in the possession of the prosecution or its agents that tends to negate the guilt of the accused, mitigate the
offense charged, impeach the government’s witnesses or evidence, or reduce the likely punishment of the
accused if convicted.
b. The prosecutor should diligently advise other governmental agencies involved in the case of their
continuing duty to identify, preserve, and disclose to the prosecutor information described in (a) above.
Before trial of a criminal case, a prosecutor should make timely disclosure to the defense of information
described in (a) above that is known to the prosecutor, regardless of whether the prosecutor believes it is
likely to change the result of the proceeding, unless relieved of this responsibility by a court’s protective
order. …
140
See Allen, supra note 136.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 80 • National Registry of Exonerations • September 1, 2020
requirement—while West Virginia’s discovery statute, which also refers to Brady, explicitly
rejects materiality as limitation.
In short, these rules of professional responsibility and pre-trial discovery impose on prosecutors
a stronger, clearer, and more administrable obligation to disclose exculpatory evidence. Why
then does Brady dominate the discussion of the issue?
The answer is simple: The remedies. Brady, as we said, describes a requirement of due process.
As a result, a proven violation of the rule in Brady—however arbitrary and inconsistent that rule
may be in practice—requires reversal of a criminal conviction and can lead to civil liability for
violating the defendant’s constitutional rights. Violations of non-constitutional rules of
professional responsibility rarely have such consequences.
141
For this study, however, we are concerned with misconduct itself rather than the legal remedies
that are sometimes applied in response. For that purpose, we define the duty to disclose
exculpatory evidence as it is described in the prevailing rules of professional responsibility and
discovery, with no special requirement of “materiality.” These rules, as written, apply to
prosecutors, but most of the evidence at issue is generated by police and forensic analysts (and
in some cases, child welfare workers)—which implies that other law enforcement officers have a
duty to inform prosecutors of exculpatory evidence.
Concealing exculpatory evidence may be done intentionally or unintentionally. Unlike the duty
to speak the truth that is behind rules against perjury, the obligation here is in some respects
technical. Officers and prosecutors are more likely to misunderstand or forget it—or they may
misunderstand or forget reports or exhibits they saw, or fail to prepare and remain unaware of
evidence they should have noticed and acted on. These are not complete excuses. It is no defense
to a failure to disclose exculpatory evidence to say “I never read my own file”—but (if true) it
may mean that exculpatory evidence in that file was concealed by culpable neglect rather than
deliberate choice.
141
See Strickland v. Washington, 466 U.S. 668, 688-89 (1984).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 81 • National Registry of Exonerations • September 1, 2020
2. CONCEALING EXCULPATORY EVIDENCE, BY CRIME
The rate of concealing exculpatory evidence varies greatly by crime, as we see in Table 18.
Table 18: Rate of Concealing Exculpatory Evidence, by Crime
Murder (908)
61%
Child Sex Abuse (270)
27%
Sexual Assault (320)
32%
Robbery (122)
28%
Other Violent Crimes (270)
44%
Felonious Assault (82)
54%
Attempted Murder (50)
40%
Manslaughter (45)
42%
Drug Crimes (317)
37%
White-collar Crimes (63)
46%
Other Non-Violent Crimes (130)
28%
ALL CRIMES (2,400)
44%
As we mentioned, concealing exculpatory evidence is the most frequent type of official
misconduct among known exonerations. It’s most common in murder cases—like misconduct in
general. Exculpatory evidence was concealed in 82% of exonerations with any official
misconduct (1064/1296).
Concealing exculpatory evidence is not only the most frequent type of misconduct we study but
also the most complex. In the sections that follow we discuss who did the concealing, what they
concealed, and how they did it.
3. WHO DOES THE CONCEALING?
At Steven Crawford’s 1974 murder trial in Harrisburg, Pennsylvania, a forensic
analyst for the Pennsylvania State Police testified that Crawford left a bloody
palm print on a car at the scene of the crime because there was blood on his hand
before he touched the car—a critical conclusion that supported the prosecution’s
theory of the crime. Crawford was exonerated in 2002 after a copy of the original
report that the analyst submitted to the prosecutor was discovered. It indicated
that, as the defense claimed, the blood was already on the car when Crawford
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 82 • National Registry of Exonerations • September 1, 2020
touched it, not on his hand. The analyst had altered the report she presented in
court in order to help obtain a conviction.
142
The forensic analyst in Crawford’s case committed perjury and forensic fraud. Her misconduct
probably sent an innocent man to prison for decades. Nonetheless, as we categorize these cases,
she did not conceal exculpatory evidence. She told the prosecutor about the exculpatory
conclusion—it was in her original report to him. The prosecutor is the government official who is
responsible for all dealings with courts and with defense attorneys; informing them of
exculpatory evidence is his responsibility. As far as the duty to disclose goes, the forensic analyst
did her part by informing the prosecutor.
Given their role, it’s not surprising that prosecutors were responsible for concealing evidence in
73% of exonerations in which it happened (378/520). On the contrary, it’s surprising that police
were responsible for concealing exculpatory evidence in 33% of exonerations where it occurred
(172/520), forensic analysts in 6% (29/520), and about 11% had more than one type of official
participate in concealing evidence.
143
In each of those cases, police officers (and occasionally forensic analysts) either hid the
exculpatory evidence from the prosecution as well as the defense, or informed prosecutors
without leaving a record that we know of. Eric Robinson’s case is an extreme example:
In July 1993, Eric Robinson was arrested for a drive-by shooting in Los Angeles;
in June 1994 he was convicted of murder. Robinson was exonerated in 2006 after
his attorneys obtained the complete police file for the case and discovered that
within days of his arrest the police had excluded him as a suspect; that months
before his trial police Sergeant Mark Arneson learned the identity of the real
shooter; and that Sergeant Arneson threatened to arrest witnesses or beat them,
or both, if they did not identify Robinson. All of this was hidden from the
prosecution as well as the defense.
In some cases both the prosecutor and the police officer involved concealed different items of
exculpatory evidence. Dwight Love, for example, was convicted of murder in Detroit in 1982
primarily on the testimony of the surviving victim of an armed robbery in which one victim was
killed; he was exonerated in 2001 after the real killer confessed. In the process, defense
attorneys discovered that the prosecutor had concealed police reports with descriptions of a
killer who looked nothing like Love. In addition, a police officer had threatened to arrest Love’s
girlfriend if she did not contradict Love’s alibi, and hid earlier statements in which she
142
The briefcase was discovered by a stroke of unlikely luck. “In 2001 a discarded briefcase that had belonged to a
detective in the case was found by two youths rummaging through a dumpster. It contained information about the
case, including a copy of the original report prepared by Janice Roadcap, the state police chemist.” See Steven
Crawford.
143
This Section and Sections VII.4.a and VII.4.b do not include cases where the exculpatory evidence that was
concealed was evidence of other misconduct, which are discussed in Section VII.4.c. Except for Section VII.5.b.v, on
concealing evidence that no crime in fact occurred, all the proportions in the remainder of Section VII are based on
the analysis of the first 1362 exonerations pasted in the Registry; see Section II.2.a.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 83 • National Registry of Exonerations • September 1, 2020
corroborated that alibi. The prosecutor was unaware of the evidence concealed by the officer—as
far as we know.
4. WHAT WAS CONCEALED?
a. Objects vs. Information
In May 1989, James Richardson Jr. rescued a three-year-old from a burning
house next door to his own home in Cross Lanes, West Virginia, and called the
police, who found the girl’s mother bound, raped and beaten to death in the
house. They arrested Richardson for murder. Richardson was convicted based in
part on testimony from Fred Zain, a state police forensic analyst who later
became notorious for rampant fraud, perjury and incompetence.
144
In the
aftermath of Zain’s exposure, Richardson’s case was re-examined and the defense
discovered that the police had concealed a blood-covered flashlight they found at
the scene. DNA testing of the blood showed that it came from someone other
than Richardson or the victim. Richardson was exonerated in 1999.
In Richardson’s case, the police concealed a tangible object—the blood-covered flashlight. When
discovered, it cleared him. Objects seem like the obvious sort of things to conceal, but in fact
that only happened in 13% of exonerations in which exculpatory evidence was hidden (66/520).
Concealing information about what a person perceived, believed, knew, said or wrote happened
in almost all of those cases.
145
Concealing an object means hiding or destroying the object itself. In Richardson’s case, it was
the bloody flashlight, a physical entity that could have been subjected to tests that would have
cleared him.
By comparison, when Ulysses Charles was tried for sexual assault in Boston in 1984, a forensic
analyst testified that he found no seminal fluid on a vaginal swab of the victim, leading him to
conclude that the rapist had not ejaculated. That was a lie. In fact, that analyst had found
seminal fluid on the swab that came from a person with blood type O—which meant that
Charles, whose blood type was B, could not have committed the rape. The damage from that lie
was as catastrophic as the misconduct in Richardson’s case, but it consisted of concealing
information about an object, the results of the tests on the swab rather than the swab itself.
Nearly 40% of the few exculpatory objects that we know were concealed, altered or destroyed
(25/66) were directly associated with the crime, or were said to be. They included actual murder
weapons—or in one case (Michael Pardue), an old shotgun that was cleaned so it looked like it
might have been fired recently and therefore could have been the murder weapon; biological
trace evidence; clothing; stolen checks; and so forth. Another third (23/66) were photographs or
144
See supra Section IV.1, and infra Section X.2.c.
145
In this context, a witness statement on paper or in any electronic format is classified as evidence of the underlying
statement, a form of “information,” rather than as an “object”—although, of course, it is that too. A surreptitious video
or audio recording of a person who is not talking to a law enforcement officer—or who doesn’t know that the person
he or she is talking to is an officer—is not a witness statement and is classified as physical evidence.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 84 • National Registry of Exonerations • September 1, 2020
videos connected to the crime, or in a couple of cases, surveillance tapes that showed the
defendant at a location away from the crime scene at about the time the crime occurred.
Why do we know of so few cases in which exculpatory objects were concealed? Part of the
answer may be that most evidence in all trials, criminal and civil alike, is “testimonial”—what a
person says orally or in writing—rather than “real,” a physical object that’s connected to the
case. With more items of information to work with, there are more opportunities to conceal.
Another possibility is that concealed physical evidence is often impossible to detect because
many objects can easily be discarded or destroyed. If the flashlight in Richardson’s case had
been tossed in a river (or in a dumpster, most likely) the defense would never have known about
it, and James Richardson might still be in prison.
There are only two exonerations in which we know that items of physical evidence from the
crime scenes were successfully discarded or destroyed.
146
In one, it was hair from the crime
scene and a photograph of the victim’s body (Clarence Brandley); in the other, it was the hat and
jacket that the defendant was wearing when he was arrested (Christopher Harding). We have no
idea how many other troublesome objects disappeared without a trace in the investigations of
other defendants who were convicted and later exonerated.
Information, on the other hand, may be difficult to hide or destroy for at least three reasons:
(i) If the information is recorded—as a witness statement, in a report, in an invoice, as a
business record, and so forth—it’s common to create multiple copies, one of which may later
turn up. In Steven Crawford’s case it took a stroke of unlikely luck. Twenty-seven years after he
was convicted of murder, a discarded briefcase that had belonged to a detective in the case was
found by accident in a dumpster; it contained information about the case, including a copy of a
concealed forensic report by a state police chemist that undermined the prosecution’s case.
Usually, the copy is just found in somebody’s file.
Or on a computer drive. If a statement or report is entered in a computer, there may be
electronic copies in unreachable or unforeseen locations, especially if it was transmitted to other
computers. It may be impossible to delete all copies.
(ii) Recorded information in a criminal investigation is usually evidence of what a person said or
wrote at some past time and place. Even if that statement is successfully hidden, the witness
herself may still be accessible and willing to say what she knows. For example, before Brad
Childers was convicted of a string of robberies in Texas in 2004, the real robber, Jonathan Clark,
was arrested for other similar robberies and confessed to those that Childers was charged with
as well—but that confession was concealed. Ten years later, investigators hired by Childers’
family found Clark in prison, and he told them that he had committed the robberies Childers
146
In Madison Hobley’s case, a gas can that was said to be linked to the arson murder of which she was convicted
was destroyed without fingerprinting, but the existence of the can was not concealed; in fact, the can had been
subpoenaed by the defense—and a court later ruled the destruction was not done in bad faith, and therefore was not
misconduct.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 85 • National Registry of Exonerations • September 1, 2020
was imprisoned for, and that he had told that to the police in 2002. A bloody flashlight in a
landfill can’t do that.
(iii) Information is often shared, in two senses. Some facts are known by multiple people from
their inception—the five officers who worked with a particular detective in a given year, or the
three witnesses to a robbery. And some facts that are initially known to only one person—who
set fire to the barn—become more widely known if the original witness shares them with others
after the event.
b. Substantive Evidence vs. Impeachment
Evidence at trial can serve two distinct functions:
• Substantive evidence helps prove an issue at stake in a trial directly. In Michael Morton’s
case, for example, the concealed evidence that someone used his dead wife’s credit card
while he was in jail made it more likely that he was innocent because it meant that
someone else had access to that card after her death—the killer or someone who got the
card from the killer. That is substantive evidence of innocence.
Evidence that neighbors saw a strange man with a green van hanging around the Morton
house before the murder was also substantive evidence of Morton’s innocence because
that unknown man was a potential suspect. It’s weaker evidence than the credit card use,
but this classification does not turn on the power of an item of evidence but on what it
helps prove. The test is whether the evidence would have helped the defense at trial
regardless of who testified for the state. If so, it’s substantive
• Impeachment evidence challenges the credibility of a prosecution witness who has
already testified. At Quedillis Walker’s trial, for example, Sarah Dunbar testified that
Walker was a violent man, and that he had gloves similar to ones that might have been
used in the murder—two items of circumstantial evidence that suggest that he might be
guilty, if they are true. The fact that Dunbar made a deal with the prosecution would
have been relevant at trial not because it bears on Walker’s behavior, but because it casts
doubt on the truthfulness of Dunbar’s sworn testimony.
Evidence of Dunbar’s deal was only admissible at trial because she testified for the prosecution,
in order to show that she lied. If she had not testified, it would have no bearing on the case.
That’s the test for whether evidence is used solely for impeachment.
Some concealed evidence falls exclusively into one or another of these categories. For example,
the prosecutor in Kenneth Kagonyera’s case concealed a confession by the true perpetrator, a
DNA report that excluded Kagonyera, and taped over and destroyed part of a surveillance video
that could have cleared Kagonyera and his codefendants. Each of these items stands on its own
as substantive evidence of innocence. In Gerald Atlas’s case, on the other hand, the prosecution
concealed evidence that a critical eyewitness was on felony probation, and that he had a history
of drug addiction. Neither of those facts in themselves points to Gerald’s innocence, but they do
suggest that the prosecution witness may not have told the truth.
Not all concealed evidence fits neatly into a single category. In William Harris’s case in West
Virginia, for example, an eyewitness told the police, before she picked Harris out of a lineup,
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 86 • National Registry of Exonerations • September 1, 2020
that she knew Harris and that he wasn’t the attacker. If the defense had known about that
information, they could have used it as substantive evidence that Harris was innocent—and also
to impeach that eyewitness’s later identification of Harris at trial.
Almost all of the small number of concealed exculpatory objects that we know of—the
surveillance video in Kenneth Kagonyera’s case, for the example—were substantive evidence of
innocence (62/66). Only a few concealed objects were relevant to impeachment only. In Jesus
Ramirez and Alberto Sifuentes’ case in Littlefield, Texas, for example, the prosecution concealed
a video that impeached a critical eyewitness by showing her at the crime scene an hour and a
half before she said she was there.
The much larger set of concealed information is a mixed bag. Impeachment evidence was
hidden in 80% of exonerations with concealed exculpatory evidence (416/520); substantive
exculpatory evidence was concealed in 63% (325/520); and both types were concealed in 50%
(262/520).
147
Prosecutors were responsible for concealing information in about two-thirds of the
cases, both those with concealed impeachment (286/416) and those with concealed substantive
evidence of innocence (218/325).
c. Concealing Other Misconduct
At about 10 p.m. on August 21, 2003, Noris Hilde and his wife Sherl were shot in
their trailer at a campground in Umpqua National Forest, in Oregon. Noris was
killed; Sherl was critically wounded but survived. Samuel Lawson, who had been
at the campground that morning, was charged with murder and attempted
murder.
Sherl Hilde failed to identify Lawson several times while she was recuperating,
and told the police that she did not see the shooter, but at his trial in 2005 she
identified Lawson as the man who shot her and her husband. Asked if she had
any doubt, Sherl said, “Absolutely not. I’ll never forget his face as long as I live.”
She also said that she “always knew it was him.” No physical evidence and no
other witnesses tied Lawson to the crime, but he was convicted and sentenced to
life in prison.
In 2012, the Oregon Supreme Court reversed Lawson’s conviction because Sherl
Hilde’s identification of Lawson was obtained by unduly suggestive procedures.
Sherl had been repeatedly shown different photographic lineups, each of which
included a picture of Lawson but otherwise pictured different men than the other
lineups. In addition, a detective took Sherl to court to view Lawson in person
prior to trial, and, before testifying, she was given a single photograph of Lawson
wearing the same clothes he wore when Sherl saw him the morning before the
shooting. In 2014, charges were dismissed and Lawson was exonerated.
147
The substantive evidence includes substantive testimonial evidence, substantive forensic evidence, and
substantive physical evidence.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 87 • National Registry of Exonerations • September 1, 2020
The identification of Samuel Lawson was obtained by police misconduct—deliberately
suggestive procedures that were designed to induce Sherl Hilde to identify Lawson despite her
repeated failures to do so and her statement that she didn’t see the shooter. It was a tainted
identification, one of the varieties of false evidence produced by witness tampering, as we have
discussed.
148
The Oregon Supreme Court reversed Lawson’s conviction in part because of that
misconduct.
But the court also reversed because the police concealed the improper methods they used to
obtain the identification. According to the Oregon Supreme Court, “[t]hat kind of information is
essential to an accurate determination of the reliability of an eyewitness's identification and is
the kind of potentially exculpatory evidence that the state is constitutionally required to disclose
to a defendant.” In other words, it was a separate act of misconduct to conceal the misconduct
that surrounded the identification because that information might have led the jury to reject
Sherl Hilde’s testimony and acquit Lawson.
Obviously, a police officer who has done something wrong, perhaps even criminal, is not likely
to turn around and tell the world about it. We’re all always inclined to conceal our misdeeds, and
especially in this context. The goal of the witness tampering in the Hilde murder investigation
was to nail Samuel Lawson; telling the world how it was done would defeat the purpose.
Nonetheless, concealing the misconduct that led to the identification of Lawson was itself
misconduct, as the Oregon Supreme Court said—a form of derivative misconduct.
Derivative misconduct by concealing exculpatory evidence can occur any time the authorities
commit another type of misconduct which, if known, might undercut the case against the
defendant. That includes any cases in which law enforcement officers fabricated or planted
evidence, coerced false confessions from defendants by illegal means, or deliberately tampered
with and distorted evidence from victims or third-party witnesses.
The Oregon Supreme Court recognized that witness tampering can also entail a separate form of
misconduct—concealing impeachment evidence about the prosecution witness who was
tampered with. Most court opinions on official misconduct ignore the issue, understandably: it’s
easy to overlook since it generally has no consequences for deciding the case once the primary
misconduct has been identified.
The opinions we have that do discuss derivative concealing of exculpatory evidence usually focus
on concrete records of the misconduct that were concealed—a report or a recording. For
example, in 2013 Brandon Lewis was convicted of assaulting three police officers in Payson,
Arizona. He was exonerated in 2014 because of newly disclosed evidence that the officers beat
Lewis and then lied and said he had assaulted them. The trial judge dismissed the case because
the police and prosecutors did not disclose written reports that described the beating (and
refuted the officers’ sworn testimony): “It is an injustice that so many legally relevant documents
were not properly disclosed prior to trial. The failure of due process here is clear.” The officers
would have been equally guilty of concealing that information if it had not been reduced to
148
See Section IV.3.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 88 • National Registry of Exonerations • September 1, 2020
writing, but without those reports who would have believed that the police beat a suspect for no
reason and then lied to cover it up?
Exculpatory evidence of other official misconduct was concealed in 26% of all exonerations,
more than half of exonerations with concealed exculpatory evidence (353/660). Other,
“independent” forms of exculpatory evidence were concealed in 38% of exonerations—almost
80% of exonerations with any concealed exculpatory evidence (520/660), and more cases than
any other form of official misconduct.
149
See Table 19.
150
Table 19: Concealing Exculpatory Evidence: Rates of Independent and Derivative
Violations (N = 1,361)
Concealing Exculpatory Evidence (660)
48%
Independent (520)
38%
Derivative (353)
26%
ALL OFFICIAL MISCONDUCT (757)
56%
More than 60% of exonerations with derivative concealment also include separate independent
acts of concealing exculpatory evidence (213/353). For example:
The 13-year-old victim of the rape for which Peter Rose was falsely convicted in
California failed to identify Rose in a photo array. Police officers responded by
accusing her of lying and of being a prostitute, and bullied her into making the
identification. The police concealed the victim’s initial failure to identify Rose,
and they concealed the bullying that led to the identification she eventually made.
In addition, the prosecution concealed a forensic test on semen recovered from
the rape which revealed that the rapist had a different blood type than Rose.
In this case, the police improperly pressured and manipulated the victim to identify Rose, and,
predictably, concealed that misconduct—derivative concealment. But they also concealed the
fact that the victim initially failed to identify Rose, and the prosecutor concealed a forensic test
that cleared Rose entirely—independent concealment.
149
See supra, Table 7, Section III.5.
150
Counting derivative misconduct cannot change the overall proportion of exonerations with official misconduct
because, by definition, a case can only include derivative misconduct if it also includes some other known act or acts
of misconduct.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 89 • National Registry of Exonerations • September 1, 2020
5. CATEGORIES OF CONCEALED INFORMATION
a. Impeachment of prosecution witnesses
Any number of different types of impeachment evidence were concealed in these exonerations.
Some would probably have had modest impact, if known; some would have been explosive. In
Levon Jones’s murder trial in North Carolina in 1993, for example, the defense was not told that
a central witness for the prosecution had been paid $4,000 as a reward for information about
the crime. It’s less likely that Jones would have been convicted if that item had been disclosed,
but probably not by much. (By the time Jones was exonerated in 2007, however, that same
witness had recanted her testimony entirely and disclosed that she had been threatened with
prosecution herself if she did not testify against Jones.)
At the other end of the spectrum, the sole eyewitness at Alvena Jennette and Daryl Austin’s
murder trial in 1988 was also the sole eyewitness who identified other defendants in several
earlier unrelated murder trials that had been investigated by the same Brooklyn detective. That
single piece of hidden impeachment evidence undermined the prosecution’s entire case, and led
the Brooklyn District Attorney to dismiss Jennette’s and Austin’s convictions, and at least three
others as well.
A few categories of concealed impeachment evidence show up repeatedly:
i. Incentives to testify
In a fifth of all exonerations (281/1,361), the prosecution concealed an incentive to testify on the
part of one or more of their witnesses. Sometimes it was a promise. For example, in Bernard
Baran’s wrongful prosecution for child sex abuse in Massachusetts in 1985, a video of an
interview with a supposed victim was edited to remove the portion where the child—who’d been
pressured to accuse Baran— asks “Where’s my prize? You promised me a prize.”
In other cases, it was a threat. When Adam Miranda was prosecuted for the murder of Robert
Hosey in Los Angeles in 1983, for example, the main prosecution witness was told that unless he
testified against Miranda, he himself would be prosecuted for that crime.
And sometimes, as in Levon Jones’s case, it was both.
The most common concealed incentive to testify falsely is a deal with the prosecutor to dismiss
or reduce pending criminal charges against the witness. The witness might be a codefendant
charged for the same crime as the exoneree, or she may face unrelated charges.
Michael Hash’s 2001 murder trial in Virginia included both types of dishonest witnesses. A
codefendant falsely testified that Hash and a third defendant shot the victim; he had a secret
deal with the prosecution that reduced his own charge to second-degree murder and his
punishment to six years and eight months in prison. In addition, a jailhouse informant falsely
testified that Hash confessed to him while in jail; the defense was not told that the informant
had an agreement with the prosecutor under which his term in prison was reduced from 15 years
to five—or that he had given similar testimony against at least 20 other defendants.
And of course, there’s money, from several possible sources:
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 90 • National Registry of Exonerations • September 1, 2020
After Shawn Lawrence was convicted of murder in Amityville, New York in 2015, his
appellate lawyer discovered that the prosecution had paid $4000 in relocation expenses
to a crack addict who testified that he saw Lawrence shoot the victims.
At Perman Pitman’s murder trial in New Jersey in 2007, the prosecution hid evidence
that the only witness who identified Pitman had been paid to do so by the drug dealers
who were trying to muscle the murdered victim out of the drug business.
And at John Thompson’s 1985 murder trial in New Orleans, a witness testified that he
heard Thompson admit to the killing—but the prosecution concealed a $15,000 reward
the witness received for his testimony from the victim’s family.
ii. Inconsistent statements
In 14% of the prosecutions of exonerated defendants, the authorities concealed statements by
prosecution witnesses that contradicted their testimony at trial (185/1,361). For example:
Sandra Craig did not know that the six-year old girl who accused her of child sex
abuse in Maryland in 1987 had also identified numerous other people as her
abusers and had mistaken someone else for Craig. The prosecutor, who had
attended some of the interviews with the girl, did know about these statements
and hid them.
James Haley was convicted of murder in Boston in 1972 after the victim’s
girlfriend and the victim’s roommate both testified that they had seen Haley—the
roommate’s estranged husband—in the neighborhood around the time of the
crime. The victim’s girlfriend also testified that she saw Haley commit the
murder. Haley was sentenced to life imprisonment with no possibility of parole.
Thirty-four years later, a public records request by Haley revealed that the
prosecution had concealed reports that both women initially told the police they
had not seen Haley in over a month.
iii. Criminal records and histories of dishonesty
In about 4% of the cases, prosecutors (and occasionally, police officers) concealed evidence of
crimes or dishonest conduct by state witnesses (61/1,361). For example, Paula Hall’s murder
conviction was reversed in Missouri in 2011 in part because prosecutors failed to disclose that a
witness who claimed to have heard Hall confess to the crime had herself been convicted of
passing bad checks, forgery and violating probation, and had other charges pending at the time
of Hall’s trial. And Gerardo Sandoval-Gonzalez’s immigration law conviction was dismissed in
California in 2012 because the defense was not told that a federal agent who testified against
Sandoval-Gonzalez had forged his supervisor’s signature on a report in another case.
Some types of concealed impeachment evidence are rare or unique, at least as far as we know,
but just as important as more common types in the cases in which they were eventually revealed.
In several cases, police officers concealed sexual affairs with government witnesses. At Raymond
Carter’s 1988 murder trial in Philadelphia, for example, the sole eyewitness had been in a sexual
relationship with the arresting officer for years.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 91 • National Registry of Exonerations • September 1, 2020
In the most extraordinary of these cases, five Latino gang members—Anthony Adams, Luis
Davalos, Jesse Alvarez, Jorge Alvarez and Cesar Menendez—pled guilty to manslaughter in Los
Angeles in 1996 after they had been identified by 18-year-old Sonya Flores. Five years later, they
were exonerated when Flores admitted that she had lied. She did not see the defendants at the
shooting; she identified them from pictures that disgraced police officer Raphael Perez showed
her in a police book of gang members. They would probably never have been charged in the first
place if the prosecution or the defense had known that Flores had carried on a romantic affair
with Perez since she was 16.
In other cases, officers concealed evidence that would have undermined their own testimony
about the facts of crimes they claimed to have witnessed. For example, in 1998 Jeffrey Santos
was sentenced to six years in prison after a correctional captain testified that Santos punched
him in the face at the Rikers Island jail in New York City. He was released three years later and
eventually exonerated because the captain concealed the fact that he ordered a guard to punch
him in the face hard enough to draw blood in order to create false evidence that he had been
attacked by Santos, when in fact the captain and other guards had beat Santos without physical
provocation.
b. Substantive evidence of innocence
As with impeachment, the range of concealed substantive evidence of innocence is huge, and—as
with impeachment—several categories occur frequently:
i. Forensic tests
Some of the most disturbing cases of concealed exculpatory evidence involve hidden forensic
tests. In some of the cases, it’s hard to escape the conclusion that the officials who concealed the
tests knew that the defendants were innocent but sent them to prison all the same.
You will recall that in the prosecution of Michael Morton for the murder of his wife, prosecutors
not only concealed several items of powerful exculpatory evidence, but also resisted DNA testing
on a discarded bandana that ultimately identified the real killer 24 years after Morton was sent
to prison. In the murder prosecution of Kenneth Kagonyera, Robert Wilcoxson, Larry Williams,
Jr., Damian Mills, and Teddy Isbell in North Carolina in 2001, pre-trial DNA testing was done
on bandanas and gloves discarded by the criminals. The tests proved that none of the
defendants participated in the crime—but the prosecution concealed that evidence.
DNA testing was not yet available when four teenagers—Omar Saunders, Larry Ollins, Calvin
Ollins, and Marcellius Bradford—were charged with abduction, rape and murder in Chicago in
1988, but Pamela Fish, the forensic analyst who worked on the case, was able to determine the
blood types of the rapists from semen that was recovered after the crime. Those tests also
proved that the defendants were innocent; none of them had the same blood type as any of the
rapists. That should have been the end of the case against them, but Fish concealed those results
and then lied on the witness stand and said that semen found on the victim’s body and
undergarments could have come from three of the four defendants.
Not all concealed forensic evidence is equally powerful. At William Gregory’s 1993 rape trial in
Kentucky, a forensic analyst testified that several hairs found in a stocking cap left behind by the
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 92 • National Registry of Exonerations • September 1, 2020
assailant shared “unusual characteristics” with Gregory's hairs, and were “more than likely”
from Gregory. That testimony was shown to be false once DNA testing of hairs became available
seven years later, but it reflected forensic knowledge at the time. On the other hand, the analyst
also concealed the fact that several other hairs from same cap did not resemble Gregory’s by any
standard—evidence that did not establish his innocence, but would certainly have called the
analyst’s “match” into question.
Altogether, we know of 83 cases in which exculpatory forensic evidence was concealed, 6% of all
exonerations (83/1,361). In two-thirds of those cases the forensic evidence was concealed by
prosecutors (55/83), in a quarter by the forensic analysts themselves (22/83), and in about 9% it
was concealed by police (8/83).
ii. Alternative suspects
Exculpatory information about suspects other than the wrongfully convicted defendants was
concealed in 22% of exonerations with official misconduct (166/768), about 12% of all
exonerations, more than any other type of concealed substantive exculpatory evidence.
In some cases, the concealed evidence described people other than the defendants who
committed the crimes, or might have done so, but did not uniquely identify those suspects. By
the time Anthony Gray pled guilty to murder in Maryland in 1991, for example, the authorities
knew that an eyewitness said that a white man drove the victim’s car away from the crime scene
(Gray and his two codefendants were Black), that DNA tests on semen from the victim and
fingerprints from the crime scene did not match any of the three defendants, and that a hair
from the crime scene came from a white male. Gray and his attorneys knew none of this.
Similarly, after Douglas Dilosa, a white man, was convicted of murdering his wife in 1987 in
Louisiana, his attorneys obtained a police report that had been concealed. It revealed that a
witness saw two Black men leaving Dilosa’s condominium complex on the morning of the
murder; that at the time of the crime, police were investigating a similar break-in at a nearby
condominium; that a hair found on the rope used to strangle his wife came from a Black person;
and that unidentified fingerprints were found at the crime scene.
In other cases, the alternative suspect was identified by name but the authorities ignored and
concealed that information. By the time Roy Brown was arrested for murder in 1991 in upstate
New York, the police had collected statements from at least four witnesses describing suspicious
behavior by the real killer, Barry Bench—but Brown only found out years later, after he had been
convicted and sentenced to 25 years to life in prison. Brown was exonerated when Bench was
identified by DNA testing in 2007. He would never have been convicted at all if the authorities
had investigated Bench and obtained his DNA back in 1991.
And in some cases, prosecutors and police must have known exactly who committed the crime,
but proceeded to convict an innocent defendant all the same. In 1991, Darrell Cameron was
charged with a string of robberies in Chicago despite the fact that he was at least seven inches
taller and 90 pounds heavier than the robber witnesses described. After he was convicted in one
case, Cameron’s defense lawyer interviewed a man who had been convicted of several similar
robberies in the same area, and that man admitted that he had also committed the robberies
that Cameron was charged with. Previously concealed documents in the prosecution file
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 93 • National Registry of Exonerations • September 1, 2020
revealed that the real robber had already confessed to those robberies to an assistant state’s
attorney, months before Cameron was tried.
When evidence about alternative suspects was concealed, prosecutors were responsible two-
thirds of the time (111/165), more than twice as often as police officers. Police, of course, talk to
many more witnesses than prosecutors and have many more opportunities to conceal
exculpatory witness statements, on this or any other issue. If they do so, however, we are much
less likely to ever know.
We usually learn about concealed witness statements because they were recorded or
incorporated in a report. A police officer can hide a witness statement by simply doing nothing—
never writing it up, recording it or telling anyone else. If so, it will probably never be discovered.
Prosecutors as well as defendants usually learn about witness statements from reports by police
officers—which, as we’ve pointed out, are harder to conceal than unreported information. Once
those statements are known to prosecutors, they—not the police—are responsible for any failure
to disclose them to the defense.
Twenty percent of murder exonerations include concealed evidence of alternative suspects
(123/608), compared to 6% for all other exonerations (42/753). That probably reflects the fact
that investigations for murder are bigger deals than for other crimes: the police are likely to talk
to more witnesses, so there are more witness statements to conceal, and those statements are
more likely to be written or recorded.
iii. “I don’t see him” and “Not the guy”
In a smaller number of cases, authorities concealed evidence that the defendant did not commit
the crime, without reference to who did.
In 2% of exonerations the defendants were not told that some eyewitnesses who saw the
defendants in police identification procedures failed to identify them (28/1,361). For example, in
1977, Timothy Howard and Gary Lamar James were sentenced to death based entirely on
testimony from eyewitnesses who identified them as the men who robbed a bank in Columbus,
Ohio, and shot a security guard. They were exonerated in 2003 when their lawyers discovered
that their convictions were based on egregious misconduct by a police detective who was later
fired by the Columbus Police Department. Among other things, the detective suppressed the fact
that three eyewitnesses—the two with best views of the criminals and a third who identified
Howard in court—all failed to identify either defendant in a police lineup.
We suspect that this happens far more often than we know, including many cases in which there
is no other misconduct. Police officers focus on finding evidence that identifies and convicts
their suspects. Some don’t seem to think of a failure to identify the suspect as “evidence” at all.
But it is, of course, if the defendant is charged: evidence of innocence.
In about a quarter of these cases the witnesses went beyond failing to identify the defendant and
specifically said that the defendant was not the perpetrator (8/28). For example, in 2001, 11
years after Lathierial Boyd was convicted of shooting and killing a man outside a nightclub in
Chicago, journalists located an eyewitness who was standing three feet away from the shooting.
She said that after she failed to identify anyone in a lineup, she asked the detective to identify
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 94 • National Registry of Exonerations • September 1, 2020
the suspect—and when he pointed to Boyd, she said there was “no way in the world” Boyd could
be the gunman because he did not resemble the shooter. This statement was concealed; Boyd
was exonerated in 2013.
iv. Alibi evidence
Only 19 cases, about 1% of the total, included concealed alibi evidence. This makes sense: a
defendant with an alibi usually knows where she was and who witnessed her at that time and
place. In several cases, however, authorities concealed evidence uniquely available to the police
that would have confirmed alibis the defendants had already presented, and prevented the
conviction of the innocent defendants.
When Daniel Taylor was charged with a double murder in Chicago in 1992, he knew where he
was when the crime took place: in a Chicago police station lockup. He had a release slip with his
signature and time-stamped after the murders to prove it. The prosecution claimed that the
time on the slip was a mistake. Seventeen years after Taylor was convicted, concealed notes by
the prosecutor described pre-trial interviews with seven police officers who confirmed that
Taylor was still in the lockup at the time of the murders; two of them testified at Taylor’s trial
and said the opposite. If those interviews had been disclosed, neither Taylor nor his three
innocent codefendants would have been convicted.
In two cases, the authorities suppressed video evidence of alibis. Kian Khatibi was convicted of a
stabbing on a street in Pleasantville, New York, at 1:17 a.m. on January 11, 1998, because the
Pleasantville police suppressed a videotape of Khatibi entering their police station 5 minutes
earlier, at 1:12 a.m. And Claude McCullom was convicted of murder at the Lansing (Michigan)
Community College in 2006 because the prosecution did not disclose a report by a Michigan
State Police detective who analyzed security surveillance recordings and concluded that
McCullom was in a different campus building at the time of the crime.
At Sean Ennis’s trial, the concealed evidence did not concern the location of the defendant but
the other element of an alibi defense: the time of the crime. Several years after his conviction for
rape and kidnapping in Ohio in 1990, Ennis learned for the first time that the man who gave the
victim a ride to the police station shortly after the crime told the police that the attack had
occurred earlier than the victim said, at a time for which Ennis had a clear alibi.
v. No crime
In more than a third of known exonerations the defendant was not convicted of a crime that
someone else committed, but of a crime that did not occur at all (883/2,400). In about a sixth
of those cases, 6% of all exonerations, the authorities concealed evidence that the crime the
defendant was convicted of never happened.
151
151
This estimate is based on data from two sources: 46 exonerations with concealed evidence that no crime occurred
among the first 1,361 exonerations posted, or 3.4%, plus the 66 Sergeant Watts group exoneration that all occurred
after that set, or 2.8% of all 2,400 cases. See Methodological Appendix for additional details.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 95 • National Registry of Exonerations • September 1, 2020
In about 70% of exonerations with concealed evidence that no crime occurred, police planted
drugs on innocent defendants (80/112). The great majority of those cases were the Sergeant
Watts group exonerations in Chicago (66/80);
152
the rest were similar, but not part of any
known long-term pattern of planting evidence. In 1997, for example, members of the Manatee
County (Florida) Sheriff’s Department’s anti-narcotics trafficking unit, broke in the door of
Sarah Smith’s apartment and arrested her for possession of crack cocaine they found in a bottle
of Tylenol. At her trial, the deputy who arrested her concealed the fact that he had placed the
crack in the bottle. He admitted it two years later after he was arrested in a federal investigation
of corruption in that unit.
153
In five murder trials prosecutors concealed evidence of the cause of death. Beverly Monroe, for
example, might not have been convicted of the murder of her boyfriend in Virginia in 1992 if the
prosecution had disclosed a medical report that concluded that the death was a suicide; and Eric
Jackson-Knight might have been acquitted in Brooklyn in 1980 if the prosecutor had produced
an expert report that concluded that the deadly fire he was accused of setting was accidental.
In a few rape exonerations, the authorities concealed evidence that the complainants had a
history of making false rape accusations. And in at least a dozen child sex abuse cases, police,
prosecutors and child welfare workers concealed statements by the supposed victims that they
had not in fact been molested. For example, in 1982, Alvin McCuan, his wife Debbie McCuan,
and their friends Scott and Brenda Kniffen were convicted on bizarre charges of sexually abusing
the two couples’ four children in Kern County, California. They might not have been if the jury
had been told that the children had denied many times that any abuse had occurred.
In nine frame-up cases police officers were not merely witnesses against the exonerees, but
claimed to be victims of assaults by those innocent defendants. Adam Tatum’s case is probably
the most outrageous. In 2012, he was attacked by Chattanooga (Tennessee) police officers at a
re-entry facility for convicted felons, thrown to the ground and beaten so severely with fists and
metal batons that his right leg was fractured in six places and he lost so much blood that hospital
personnel thought he had been shot. Tatum was sentenced to two years in prison for assaulting
the officers. A year later, he was exonerated when a previously concealed video showed that he
was attacked without provocation.
152
See supra Section III.3.c.i. In one of these cases, the defendant was convicted of illegal possession of a planted
gun rather than planted drugs.
153
In one additional drug crime exoneration, that of Jose Luis Pena, officers concealed evidence that plants they
found in the defendant’s care were not marijuana, but did not plant drugs on him.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 96 • National Registry of Exonerations • September 1, 2020
VIII. Misconduct at Trial
About 95% of felony convictions in the United States—and an even higher proportion of
misdemeanor convictions—are obtained by guilty pleas rather than trial verdicts. Exonerations
are dramatically different: 80% of those in the Registry followed conviction at trial
(1,928/2,400). About 28% of those trials—23% of all exonerations (546/2,400)—included
official misconduct in court during the trial.
The most common form of official misconduct at trial was perjury by law enforcement officers,
usually police officers. In addition, there were several distinct types of misconduct by
prosecutors who violated rules that govern their conduct as officers of the court who represent
the state in criminal proceedings.
1. POLICE PERJURY
Law enforcement officers are the most common witnesses called by prosecutors in criminal
trials. Most of them no doubt tell the truth, but misconduct does occur.
154
We know law
enforcement officers committed perjury in 13% of all exonerations (304/2400), including in
more than 14% of exonerations after conviction at trial (284/1928).
155
Eighty-four percent of
cases with official perjury include perjury by police officers (256/304), about 18% include
perjury by forensic examiners, and a handful include perjury by both.
156
In Section VI we discussed one subgroup of cases with perjury by law enforcement officers:
those in which officers “fabricated evidence” by writing false reports or testifying falsely to
things they claimed to have seen or heard that powerfully incriminated the defendants. That
154
It may be—as many defendants and defense attorneys believe—that police are more likely to lie about the
circumstances and legal justifications for stops, searches and arrests, than about evidence of guilt. Much, perhaps
most of such testimony takes place at preliminary examinations and hearings on motions to suppress rather than at
trial. In any event, this research focuses exclusively on misconduct that bears on evidence of guilt or innocence. We
did not collect or analyze data on perjury in testimony that bears on the legality of searches and seizures.
155
In about 7% of the exonerations with known perjury by government officers, the convictions were based on guilty
pleas (20/304); in those cases the perjury occurred at grand jury hearings, preliminary examinations or hearings on
pre-trial motions.
156
In four cases we know of perjury by prosecutors; three of them also included perjury by police officers.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 97 • National Registry of Exonerations • September 1, 2020
category includes almost all known perjury by forensic examiners;
157
cases in which police
officers falsely testified that defendants had confessed when they had not; and cases in which
officers planted drugs on innocent suspects, or falsely claimed that they had been assaulted.
Each category includes cases in which the defendants pled guilty and spared the officers the
need to commit perjury at trial.
In more than 90% of exonerations with police perjury the defendants were convicted at trial
rather than by guilty plea (236/256). At 75% of those trials (178/236)
158
—7% of all
exonerations—police officers lied about the conduct of the investigations or statements by other
witnesses rather than about their own observations. Those are the cases we discuss in this
section.
In some of the cases, police lied to make their own observations or conduct appear consistent
with statements by other witnesses. For example:
At Steven Dewitt’s murder trial in Washington D.C. in 1992, a police officer lied
and said that an eyewitness told the police that the killer was driving a car with
temporary license plates with the number sequence “818”—as was Dewitt when
he was stopped. Ten years after Dewitt was convicted, defense attorneys
discovered that the eyewitness had actually said the temporary plates included
the numbers “829” and that police had stopped the real killer shortly after the
crime—driving a car with temporary plates that ended with “829”—and let him
go. Dewitt was exonerated in 2004.
Most police perjury, however, concerned the conduct of investigations.
At Christopher Roesser’s murder trial in Georgia in 2008, a detective denied that
a witness to the homicide ever told him that the deceased pointed a gun at
Roesser and demanded his money. At a retrial in 2011, after Roesser’s conviction
was reversed, a medical examiner’s investigator told the jury that the detective
had told him exactly that. Roesser was acquitted.
At a pretrial hearing before James Walker’s robbery-murder trial in Brooklyn in
1971, a detective testified that Walker had never been placed in a lineup. In fact,
there had been a lineup at which the surviving victim of the robbery picked a
police officer “filler” rather than Walker, but the jury did not know that when that
157
The exceptions we know about include a case in which the forensic examiner lied about his educational qualifications
as an expert (LeFever), and one in which a medical doctor misrepresented the contents of the extensive medical
records in the case (Eldridge).
158
This is an estimate. It’s derived by totaling the number of trials of exonerees in which police officers lied about their
own observation—trials of cases in which the officer planted drugs (or in one case, a gun) on the defendants (7 in the
Sergeant Watts group and 12 others), plus those in which officers falsely claimed that the exonerees assaulted them
(6), plus those in which police testified about fabricated confessions (33)—subtracting that total (58) from the number
of all exonerations after convictions at trial in cases with perjury by police officers (236), and dividing the result by the
total number of trials of exonerees with any police perjury (178/236).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 98 • National Registry of Exonerations • September 1, 2020
witness identified Walker at trial. Walker was exonerated 19 years later, after a
pro-bono lawyer managed to uncover the true course of the investigation.
One common aspect of criminal investigations that police lied about was the conduct of
interrogations at which innocent defendants confessed. For example:
Richard Danziger was convicted at a rape-murder trial in Texas in 1990 based on
a confession by his codefendant, Christopher Ochoa. At Danziger’s trial, the
officers who obtained the confession testified that they took “great pains” not to
tell Ochoa any of the facts of the crime, when in fact they had gone to great
lengths to “correct” Ochoa’s false confession to correspond to facts they already
knew. Danziger and Ochoa were both exonerated by DNA in 2002.
When William Oakes was convicted of murder in New York in 1985, the two
detectives who took his confession testified that they did not abuse him in any
manner. After he was convicted, however, two fellow detectives gave sworn
statements that those officers told them that during a “night of terror” that
“scared the hell out of Oakes” they beat Oakes, fired a gun three times near his
ear, pointed it at him and threatened to pull the trigger. Oakes was exonerated in
1992.
As we have pointed out, we undercount misconduct of all sorts because it’s usually hidden and
often never comes to light, even if an innocent defendant is exonerated. That general problem is
even worse for police perjury because of the limitations of our data.
We rarely have access to transcripts of the trials at which exonerees were falsely convicted. In
most cases, we rely on other legal documents—motions, briefs, opinions—and on news reports.
In many cases it is clear that the jury was misled by false evidence, but we don’t know which
police officers testified, or if they did, what they talked about.
We are most likely to be able to identify police perjury if defendants (or their lawyers) raised it
as a legal claim, or if we know that an officer testified in a manner that is plainly inconsistent
with facts that are now known. Otherwise, we rarely know.
2. TRIAL MISCONDUCT BY PROSECUTORS
a. Permitting Perjury
A lawyer may not knowingly use false evidence in court. That prohibition has two parts. First,
it’s misconduct for a lawyer to “offer evidence [in court] that the lawyer knows to be false.”
159
Second, “if a witness called by the lawyer has offered material evidence and the lawyer comes to
know of its falsity, the lawyer [must] take reasonable remedial measures….”
160
159
Am. Bar Ass’n, Model Rules of Prof’l Conduct, Rule.3.3(a)(3). Note that there is a possible exception for testimony
by a criminal defendant.
160
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 99 • National Registry of Exonerations • September 1, 2020
If the lawyer is a prosecutor in a criminal case, the “remedial measures” in response to
unanticipated lies by a witness must include disclosing the perjury to the court. That follows
from the prosecutor’s professional and constitutional obligation to disclose exculpatory
evidence—in this context, the fact that a prosecution witness has just lied—but it’s also an older
constitutional rule that the Supreme Court built on when it defined the duty to disclose
exculpatory evidence in Brady v. Maryland in 1963.
161
In 1959, the Supreme Court held in Napue v. Illinois
162
that “a conviction obtained through use
of false evidence, known to be such by representatives of the State, must fall under [the due
process clause of] the Fourteenth Amendment,” that this “result obtains when the State,
although not soliciting false evidence, allows it to go uncorrected when it appears,” and that “[i]t
is of no consequence that the falsehood bore upon the witness' credibility, rather than directly
upon defendant's guilt.”
163
In short, under Napue, a prosecutor has a constitutional obligation to
correct perjury by a state witness even if she did not herself offer the false testimony.
We know that prosecutors permitted perjury to go uncorrected in 8% of exonerations
(186/2,400). In almost all of these cases, by definition, the prosecution also concealed
exculpatory evidence.
164
For example, when a forensic analyst falsely testified at Ulysses
Charles’s rape trial that he found no seminal fluid on a vaginal swab of one from the victims, the
prosecutor knew that the analyst had in fact found seminal fluid on the swab that could not have
come from Charles, but did nothing to disclose that exonerating evidence or correct the
perjury.
165
Sometimes the perjury is related to other categories of misconduct. In a minority of cases, the
prosecutors themselves procured the perjury, presented it in court, and then, of course, did
nothing to correct it. In others, prosecutors failed to correct perjury by police or other law
enforcement officers. For example:
In June 2003, Abdel-Ila Elmardoudi, Karim Koubriti, and Ahmed Hannan were
convicted of supporting terrorism and related charges in federal district court in
Detroit, Michigan, for allegedly planning bomb attacks on targets around the
world. The prosecution focused heavily on drawings by the defendants that
allegedly depicted one such “target”—a hospital in Jordan. In fact, the prosecutor
knew that the drawings were among a set of harmless sketches of the entire
region, and the prosecutor concealed actual photographs of the hospital taken by
a State Department investigator.
161
Brady v. Maryland, 373 U.S. 83 (1963).
162
Napue v. Illinois, 360 U.S. 264 (1959).
163
Id. at 269.
164
In a handful of cases, it appears that the perjury by the state witness in question was or should have been known
to the defense.
165
See supra Section VI.4.a.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 100 • National Registry of Exonerations • September 1, 2020
When the investigator lied under oath and said he had no comparison photographs of
the hospital, the prosecutor let the perjury stand. All three defendants were exonerated
in 2004. In 2006, the prosecutor and the investigator were indicted for obstruction of
justice; they were later acquitted.
In most cases, however, the perjury was by civilian witnesses. The most common issue they lied
about is the one the Supreme Court addressed in Napue: the fact that they had received benefits
from the state in return for testifying—usually favorable treatment in pending criminal cases of
their own.
166
For example:
At Jabbar Collins’s murder trial in Brooklyn in 1994, three witnesses lied and said
that they made no deals with the prosecution in return for testifying against
Collins. The prosecutor not only failed to correct the record, but ridiculed the
claim that such deals had been made in his argument to the jury. In 2010, a
federal judge compelled the prosecution to produce withheld documents that
described how the prosecution had threatened and manipulated the witnesses,
and eventually struck deals with them. After the court vacated the conviction, the
District Attorney’s office agreed to dismiss the charges but reiterated its
"position, then and now… that we believe in this defendant's guilt." The judge
responded: “It’s really sad that the D.A.'s office persists in … saying they did
nothing wrong here.”
Lying about deals with the prosecution may be a common form of perjury by government
witnesses, but that is not the only reason we see it so frequently in cases in which prosecutors
have permitted perjury to stand uncorrected. In that context, the prosecutor’s knowledge of the
perjury by a witness is unusually easy to detect.
We only list a case as including this sort of misconduct if there is specific evidence—statements
or documents—that the prosecutor knew the testimony was false. If a witness lies about other
issues, we are unlikely to have enough information to conclude that the prosecutor was aware of
the perjury; but prosecutors will always know about deals with witnesses that they or their
offices entered into.
b. Lying in Court
In 1997, in Hampton County, Virginia, Ricky Cullipher was convicted of shooting
a friend who was left brain-damaged and in a coma. Cullipher’s attorney, George
L. Smith, asked for a delay because an eyewitness failed to appear in court, but
the trial proceeded after the prosecutor told the court that the witness had
166
Napue, 360 U.S. at 270.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 101 • National Registry of Exonerations • September 1, 2020
nothing to add to the testimony of other witnesses. In fact—as the prosecutor
knew—the missing eyewitness would have testified that the victim shot himself.
167
In 2009, Timothy Parkes was convicted of bank fraud in federal court in
Tennessee. In his closing argument, the prosecutor told the jury that if Parkes
were acquitted, he’d get to “keep the $4 million” that his company had at one
point owed the bank in question. This was false; by then most of the money had
been repaid. The jury would have known that if the same prosecutor had not
persuaded the trial judge to prevent Parkes’ lawyer from presenting evidence that
the debt had been paid off.
In both of these cases, prosecutors lied about important facts in trials at which they represented
the government. They weren’t speaking under oath, so they did not commit perjury, but lawyers
are prohibited from lying in court under any circumstances. Rule 3.3 of the American Bar
Association’s (ABA) Model Rules of Professional Conduct, for example, provides that “A lawyer
shall not knowingly… make a false statement of fact or law to a tribunal or fail to correct a false
statement of material fact or law previously made to the tribunal by the lawyer.”
168
Violating this
duty can subject a lawyer to punishment for contempt of court, even if not under oath.
169
This rule applies to prosecutors with special force. Standard 3-1.4 of the ABA’s Standards for the
Prosecution Function states that “the prosecutor has a heightened duty of candor to the
courts.”
170
Courts agree that they are “entitled to a higher degree of candor and professional
responsibility from government counsel,”
171
and should be able “to rely on the prosecutors' open
court, on the record representations, without the need of a formal oath.”
172
In practice, punishing prosecutors for lying is rare. In 2007, Mike Nifong—the District Attorney
of Durham County, North Carolina, who handled a notorious rape prosecution of several
members of the Duke University lacrosse team—was held in contempt and sent to jail for one
day for lying in court in that case.
173
And in 2013, Ken Anderson, who as District Attorney of
Williamson County, Texas, sent Michael Morton to prison for 25 years for a murder he did not
commit, was sentenced to 10 days in jail because he lied to a judge in 1987 when he said that he
167
It’s worth noting that Cullipher’s defense attorney provided a constitutionally inadequate defense: he had not read
the police reports and did not know that the prosecutor was lying. In fact, it was the defense counsel’s ineffective legal
assistance rather than the prosecutor’s misconduct that was the basis for the reversal of Cullipher’s conviction.
168
Am. Bar Ass’n, Model Rules of Prof’l Conduct, Rule.3.3(a)(1).
169
E.g., In re Aguilar, 97 P.3d 815, 820 (Cal. 2004) (“It is, of course, an extremely serious breach of an attorney's duty
to a court to lie in statements made to the court, and an intentionally false statement made by an attorney to a court
clearly constitutes a contempt of court.”) (Internal citations omitted.)
170
Criminal Justice Standards for the Prosecution Function § 3-1.4 (Am. Bar Ass’n 2017).
171
Miller v. Lehman, 603 F. Supp. 164, 166-67 n.3 (D.D.C. 1985).
172
People v. Hameed, 666 N.E.2d 1339, 1342 (N.Y. 1996).
173
Sahila Dewan, Duke Prosecutor Jailed; Students Seek Settlement, N.Y. Times, Sept. 8, 2007.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 102 • National Registry of Exonerations • September 1, 2020
had disclosed all the evidence in his possession that tended to show that Morton was innocent.
174
(Anderson actually served only four days.) As far as we know, these are the only two American
prosecutors who have ever been convicted of criminal contempt for lying in court, although a
few others may have been disbarred, as Nifong and Anderson were.
On the other hand, lying in court by prosecutors is reasonably common in cases in which
defendants were later exonerated. We know it occurred in 4% of exonerations (94/2,400), but
the real rate is probably considerably higher because we only classify a statement as a lie if we
have a clear indication that the prosecutor knew it was false.
About half of courtroom lies by prosecutors occurred in closing argument. This is not surprising:
closing argument is the context in which lawyers are permitted to discuss the case most freely.
Needless to say, closing arguments are not supposed to be occasions for lawyers to lie or make
up facts. Advocates are allowed to try to persuade the jury to accept their versions of the events
on trial—that’s the main purpose of argument—but they are limited to discussing evidence that
was presented in court. Lawyers are allowed to draw inferences from the evidence in closing
argument, and to try to persuade the jury to accept those inferences—for example, that a witness
whose testimony is disputed is telling the truth. But all lawyers—and especially prosecutors in
criminal trials—are prohibited from making such arguments based on factual assertions they
know to be false.
A common lie of this sort concerns deals between prosecutors and witnesses:
At Yancey Douglas and Paris Powell’s separate trials, in 1995 and 1997
respectively, for a drive-by murder in Oklahoma City, the surviving witness to the
shooting identified them as the shooters—and testified that he had received no
favors from the prosecution for testifying. In his closing argument, the prosecutor
repeated the claim that the victim had received no benefits for testifying. Powell
and Douglas were both sentenced to death. In 2009, all charges were dismissed
after the victim recanted and said that he had been drunk and high on the night
of the shooting and was unable to identify the shooters, but that he agreed to
name them in return for a deal with the prosecutor for a reduced sentence on
drug trafficking charges that he himself faced.
In other trials of exonerated defendants, Timothy Parkes’s for example,
175
the prosecutor lied in
closing argument by making factual claims that had no basis in any evidence in the record. And
in Michael Tillman’s murder trial in Chicago in 1986, the prosecutor told the judge who tried the
case without a jury that a pubic hair from Tillman had been found at the crime scene. In fact, as
the prosecutor knew, the only witness who testified on the issue—a crime analyst—had only said
the hair was “similar” to Tillman’s hair.
174
Molly Hennessy-Fiske, Ex-Prosecutor Punished For Withholding Evidence in Murder Case, L.A. Times, Nov. 8,
2013.
175
See supra Section VIII.2.b.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 103 • National Registry of Exonerations • September 1, 2020
When prosecutors lied in court outside of closing argument, they usually did it in proceedings
on procedural issues. In some trials they then went on to repeat the lies in closing argument.
At Daniel Villegas’s murder trial in El Paso, Texas in 1995, his defense attorney
agreed to stipulate to the prosecutor’s description of the autopsy report, including
a statement that the victim had been shot three times. In fact, the autopsy had
found only two bullet wounds, as the prosecutor well knew: he had said so
himself at an earlier trial that ended with a hung jury. In his closing argument,
the prosecutor used this misinformation to argue that because Villegas had shot
the victim three times, he must have intended to kill him.
In many of these cases, defense lawyers could have prevented or corrected the prosecutors’
misrepresentations. Michael Tillman’s and Daniel Villegas’s lawyers both had access to the
forensic reports that the prosecutor misrepresented at their trials, but let the lies stand. Ricky
Cullipher’s lawyer, as we noted, was no better.
176
That is a separate and equally important form
of misconduct by the lawyers who represented the exonerees. Taking advantage of the
incompetence or indifference of defense attorneys is no justification for government
misconduct.
c. Improper Statements in Closing Argument or Cross-examination
So far, the misconduct at trial that we’ve discussed concerns lies—lies by police, lies by state
witnesses that prosecutors fail to correct, lies by prosecutors themselves. But prosecutors may
also commit misconduct in court proceedings without lying, typically in statements in closing
arguments or questions on cross-examination.
Many rules restrict what lawyers can say in argument and cross-examination. Some are firm and
some are discretionary; some are clear and specific, others general and vague. Violations of
these rules—clear or ambiguous—are common.
In general, we consider these to be less severe forms of misconduct than the others we have
discussed—in part because the most serious subcategory, improper argument that includes
deliberate deception by the prosecutor, is discussed in the previous section on lying in court
rather than here.
177
Unlike lying or permitting perjury, these misdeeds do not involve concealing information from
the defense. As a result, they can (at least in theory) be brought to the attention of the judge and
addressed on the spot—if the defense attorney is prepared and vigilant (which is often not the
case), and lodges an objection on the spot. Otherwise, any objection will generally be considered
176
See supra Section VIII.2.b.
177
Strictly speaking, a question to a witness—even a strongly worded leading question on cross-examination, such as
“You’ve killed several people, haven’t you?”—is not a statement that can be deceptive but an inquiry that is negated if
the witness answers “No.” Such a question obviously implies that the witness killed people, and may well be improper,
but even if the implication is false, it is not, literally, a lie.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 104 • National Registry of Exonerations • September 1, 2020
to have been “waived”—which happens frequently because the lawyer is inattentive, unskilled or
unprepared.
There are also strong reasons for competent, attentive lawyers to forgo objections.
Claims of improper cross-examination or argument rarely succeed on appeal. If the defendant
did not object at trial, she is usually not allowed to raise the issue at all; if she did object and the
trial judge sustained the objection and told the jury to disregard the statement, an appellate
court will typically say that any problem was corrected by the lower court; and if the trial judge
overruled the objection, a higher court will generally agree with that outcome, or find that if the
trial judge “erred” the error was “harmless” and does not require reversing the judgment.
There is little incentive to make an objection that is likely to be ignored, and it may draw
attention to a statement that would otherwise go unnoticed, or suggest that the defendant has
something to hide.
In addition, this sort of misconduct occurs in open court, an arena—unlike criminal
investigation—in which the state and the defense operate on roughly equal terms. Defense
attorneys as well as prosecutors often tread close to or cross the foul lines in court. Many trial
lawyers fail to object to arguably improper questions and arguments by their opponents in order
to discourage the opposition from objecting to their own questions and arguments, and the
judge from granting objections that are made.
For all these reasons, defense attorneys frequently fail to object to misconduct in argument or
cross-examinations—and without an objection, we will rarely know that this sort of misconduct
occurred.
We only know about impermissible arguments and cross-examinations if they become the
subjects of significant disputes, typically on review after conviction. They occurred in about 4%
of exonerations (93/2,400), mostly, 3%, in closing argument (77/2,400), although 1% of cases
included impermissible cross-examination (26/2,400) and about half a percent had both
(10/2,400).
Closing argument and cross-examination are two contexts in which an attorney can talk about
the facts of the case in her own voice. In both, the ABA Standards for Criminal Justice provide
that a “prosecutor should not bring to the attention of the trier of fact matters that the
prosecutor knows to be inadmissible, whether by… asking legally objectionable questions, or
making impermissible comments or arguments.”
178
For argument, the nature of the problem is obvious. The main purpose of closing argument is to
permit the lawyer for a party to make her case to the jury—or the judge in a non-jury trial—by
persuasively organizing and summing up the evidence that supports her client’s position. That
creates an opportunity to tell the jury things that are false, unsupported by evidence, misleading
178
Criminal Justice Standards for the Prosecution Function § 3-6.6 (Am. Bar Ass’n 2017).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 105 • National Registry of Exonerations • September 1, 2020
or inflammatory. More often than not, such transgressions are overlooked—but they can do
serious damage in the context of a particular case.
In 1993, Kathryn Wilson was convicted on five felony counts of child sex abuse in
a notorious prosecution for satanic ritual child sex abuse at the Little Rascals
daycare center in Edenton, North Carolina. She was sentenced to life
imprisonment.
Two years later, the North Carolina Court of Appeals reversed the conviction for
prosecutorial misconduct, including “grossly improper” closing argument. During
the trial, the prosecutor was prohibited from impeaching Wilson’s credibility by
calling witnesses to testify to an alleged theft for which she was never charged, let
alone convicted. Such witnesses are generally excluded in American trials
because their evidence is considered insufficiently reliable and overly prejudicial
to the defendant. The prosecutor responded by telling the jury in closing
argument that he had witnesses in court ready to testify to “the truth”—that
Wilson had committed the theft—but “could not put those people on the witness
stand” to do so.
The Court of Appeals held that “This was a grossly improper argument... [T]he
prosecution accomplished during its closing argument precisely what it could not
during the trial.” The court went on to reverse the conviction because the
supposed victims, who had been heavily coached, “testified to events occurring
approximately three years before trial when they were only three or four years
old…. [T]here were no witnesses to the alleged abuse, and scant physical
evidence….Defendant testified on her own behalf and denied all of the allegation.
… [Her] credibility was critical to her defense.”
179
The prosecutor at Wilson’s trial made factual claims that had no basis in the evidence presented
in court. That’s a common type of improper closing argument that we see in trials of exonerated
defendants. In Wilson’s case, the misconduct was aggravated because the evidence the
prosecutor described in argument had been specifically excluded from consideration.
Other types of improper argument include:
• Statements that the prosecutor “knows” or is “certain” that the defendant is guilty, which
are prohibited by the general rule that at trial “A lawyer shall not… state a personal
opinion as to the justness of a cause… or the guilt or innocence of an accused.”
180
179
State v. Wilson, 456 S.E.2d 870, 874-875 (1995).
180
Am. Bar Ass’n, Model Rules of Prof’l Conduct, Rule.3.4.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 106 • National Registry of Exonerations • September 1, 2020
• Arguments that a defendant who did not testify or talk to the police must be guilty or she
would have said otherwise, which violate the defendant’s Fifth Amendment right to
remain silent.
181
• Appeals to racism, sexism or homophobia.
• Inflammatory arguments: the defendant is a monster who must be stopped; God, or
justice for the dead victim, or protection of the community, requires a conviction.
Marcella and Ricky Pitts’s child sex abuse convictions, for example, were reversed in part
because the appellate court found that the prosecutor had essentially told the jury that
Christ was a witness for the prosecution.
182
Cross-examination also permits a lawyer to speak directly to the judge and jury in a different
manner than in closing argument. Unlike direct examination, lawyers may use leading questions
freely in cross-examination of a witness called by the opposing party. That means they can
articulate the testimony in the questions they ask, with the witness limited to agreeing or
disagreeing. For example:
Q: When you first entered the conference room, there was only one other person there,
right? A: Yes.
Q. And that was your boss, Sheila Brooks? A: Correct.
Q: And she told you, didn’t she, while you were still alone, ‘Don’t say anything unless I
ask you a question’? A: Right
On direct examination, the lawyer might have been limited to “Was anybody in the conference
room when you entered?” and “What, if anything, did she say to you?”
The structure of cross-examination makes it possible for a prosecutor to phrase an assertion as a
question—“You embezzled more than $100,000 from your employer, didn’t you?”—and create
the impression that the defendant did something bad or incriminating even if the defendant,
predictably, denies it. That practice is frowned upon; in some cases, it amounts to serious
misconduct.
At Charles McClaugherty’s 2001 murder trial in Albuquerque, New Mexico, the
prosecutor asked him on cross-examination: "Sir, would it surprise you to hear
that your sister, Sarah Tucker, gave a statement to the police 6/19/99 … that said
that you admitted to her that you shot a gun?” The defendant answered that it
would surprise him. The prosecutor then asked a question about a similar
statement McClaugherty supposedly made to his roommate, Sherri Goen, and got
the same answer.
181
Miranda v. Arizona, 384 U.S. 436 (1966); Griffin v. California, 380 U.S. 609 (1965).
182
People v. Pitts, 223 Cal. App. 3d 606, 701 (1990).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 107 • National Registry of Exonerations • September 1, 2020
In 2003, the New Mexico Supreme Court vacated McClaugherty’s conviction and
ordered a new trial because “Defendant had no chance to prove that he never
made the statements to which the prosecutor referred during cross-examination.
Without such an opportunity, the jury was left to assume that Defendant actually
admitted that he shot a gun that night.”
183
Charges against McClaugherty were later dismissed because the misconduct was found to be
willful, in part because “[t]he actual police interviews conducted on June 19, 1999, do not
contain statements from either Tucker or Goen that Defendant confessed that he shot and then
bragged about it.”
184
183
State v. McClaugherty, 133 N.M. 459, 468 (2003).
184
State v. McClaugherty, 144 N.M. 483, 494 (2008).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 108 • National Registry of Exonerations • September 1, 2020
IX. Federal Cases
Federal cases are a small minority of all criminal prosecutions in the United States. In 2006, for
example, only 6% of felony convictions in the United States were in federal court,
185
and only
14% of prison inmates were in federal prisons.
186
Federal convictions are also systematically
different from state convictions. Eighteen percent of felony convictions in state courts in 2006
were for violent crimes,
187
and 54% of inmates in state prisons were convicted of such
crimes
188
—but fewer than 4% of federal convictions were for crimes of violence,
189
and only 9%
of federal prisoners were serving time for violent felonies.
190
Five percent of exonerations in the United States are from convictions for federal crimes
(112/2,400). Exonerations in federal cases are no more representative of all federal convictions
than state-court exonerations are of state convictions. Violent crimes are overrepresented
among federal exonerations, 28% (31/112) versus about 4% of convictions, but they’re still a
minority compared to more than 80% of state-court exonerations. Drug crimes are
underrepresented but only slightly, 22% of exonerations (25/112) and 36% of convictions, while
immigration and weapons crimes are greatly underrepresented—about 4% of exonerations each,
compared to 23% and 12% of convictions, respectively. On the other hand, white-collar crimes
account for 16% of federal convictions,
191
but 41% of federal exonerations. See Table 20.
185
See Bureau of Justice Statistics, Federal Justice Statistics 2006 Table 4.2 (2006) (hereinafter BJS, Federal
Justice Statistics 2006); see also Bureau of Justice Statistics, Felony Sentences in State Courts, 2006 Table 1.1
(Dec. 2009) (hereinafter BJS, Felony Sentences in State Courts, 2006) (dividing 73,804 by 1,206,094).
186
Bureau of Justice Statistics, Prisoners in 2006, Table 1 (Dec. 2007).
187
See BJS, Felony Sentences in State Courts, 2006, supra note 185, at Table 1.1.
188
Id. at Table 11.
189
See BJS, Federal Justice Statistics 2006, supra note 185, at Table 4.2.
190
Id. at Table 12.
191
“White-collar” exonerations include those for convictions for bribery, false report, forgery, fraud, obstruction of
justice, perjury, supporting terrorism, and tax evasion. “White-collar” crimes for federal convictions in 2006 include
convictions for Property offenses classified as “Fraudulent” and Regulatory offenses classified as “Public Order,”
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 109 • National Registry of Exonerations • September 1, 2020
Table 20: Proportions of all Federal Exonerations, and of Federal Convictions in 2006, by
Crime
Exonerations
(112)
Convictions
(73,804)
192
White-collar (46)
41%
16%
Violence (31)
28%
4%
Drugs (25)
22%
36%
Immigration (4)
4%
23%
Weapons (5)
4%
12%
Military Justice (1)
1%
-
Other Property &
Public Order Crimes
-
9%
The rate of official misconduct is somewhat higher among exonerations in federal cases than in
state cases, 61% compared to 54%. There is a much larger difference, however, between the rates
of misconduct in state and federal exonerations by prosecutors and by police officers. In state
cases, police officers are more likely to commit misconduct than prosecutors by a modest
amount, 36% (821/2,288) to 29% (671/2,288). In federal cases prosecutors commit misconduct
much more often than police, 52% (58/112) to 20% (22/112). All three rates—overall
misconduct, misconduct by prosecutors and misconduct by police—vary considerably among the
three categories of crimes that account for 91% of federal exonerations. See Table 21.
“Tax,” “Bribery,” “Perjury, Contempt & Intimidation,” “National Defense” and “Environmental.” See BJS, Federal
Justice Statistics 2006, supra note 185, at Table 4.2.
192
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 110 • National Registry of Exonerations • September 1, 2020
Table 21: Official Misconduct by Prosecutors and by Police in Federal Exonerations, by
Crime*
Misconduct by
Prosecutors
Misconduct
by Police
ALL OFFICIAL
MISCONDUCT
White-collar (46)
65%
9%
65%
Violent (31)
45%
39%
61%
Drugs (25)
36%
20%
48%
ALL FEDERAL
EXONERATIONS (112)
52%
20%
61%
* Some cases include misconduct by more than one type of official.
As we see in Tables 19 and 20 together, the distinctive patterns in federal exonerations overlap:
the largest concentration of federal exonerations is among white-collar crimes; the
overwhelming majority of misconduct is by prosecutors rather than police; and that skewed
pattern of misconduct—prosecutors rather than police—is strongest among white-collar cases.
1. WHITE-COLLAR CRIMES
In some respects, white-collar exonerations are to federal courts what murder exonerations are
to state courts: They’re the most common type of federal exonerations, 41% (compared to 40%
for state murder exonerations); they’re overrepresented compared to the rate of convictions
(although by a smaller margin than murder cases among state court exonerations); and they
have a high rate of official misconduct, 65% compared to 72% for state-court murders.
There are major differences, of course—especially from the point of view of the exonerees. Those
convicted of federal white-collar crimes received far lighter sentences than those convicted of
murder,
193
and they were exonerated much more quickly. The median time from conviction to
exoneration was 14 years for murder cases but only three years for federal white-collar crime
exonerations.
On the other hand, from the point of view of the prosecutors, there are important similarities
between these two sets of cases. Judging from exonerations, white-collar cases in federal court—
like murder cases in state court—are more likely than any others to be big-ticket prosecutions:
expensive, long-running, conspicuous. That suggests that, like murder cases in state court,
194
193
Thirteen percent of murder exonerees were sentenced to death (121/908) and 43% were sentenced to life
imprisonment (391/908); 2% were not sentenced before exonerations (14/908); and the remaining 42% received
prison sentences averaging 34 years. Among Federal white-collar exonerations, 11% received probation, 41% were
not sentenced, and 48% were sentenced to prison terms averaging 2.6 years.
194
See supra, Section III.3.a.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 111 • National Registry of Exonerations • September 1, 2020
they are unusually likely both to include official misconduct, and, for misconduct that occurs, to
be discovered. For example:
In July 2008, United States Senator Ted Stevens of Alaska was indicted in
Washington, D.C., for failing to report gifts totaling about $250,000, allegedly
the costs of renovations of his home that were paid by the VECO Corporation, an
oil pipeline service and construction company. Stevens, who was running for
reelection, demanded an immediate trial. He was convicted on October 27, 2008,
and lost the election eight days later.
Five months later, the charges against Stevens were dismissed at the request of
the Department of Justice. In between, a whistleblower affidavit from an FBI
agent led to the discovery of a raft of government misconduct and concealed
evidence, including:
• A female FBI agent had an inappropriate relationship with the
central witness against Stevens—Bill Allen, the former CEO of
VECO, who had pled guilty to bribing several Alaskan state
legislators and whose sentencing was deferred until after he
testified against Stevens. Allen had also done favors for other FBI
agents.
• Allen had told the government that the fair market value of the
improvements on Steven’s property was about $80,000—not the
$250,000 that he testified to at trial—and that Stevens probably
would have paid for them if asked.
• Another witness who was involved in the renovations told
prosecutors that he had told the senator that the bills Stevens
received—and promptly paid—included all of the work that was
done. In response, they sent him back to Alaska and concealed his
existence.
In November 2011, an independent investigation ordered by the judge concluded
that federal prosecutors in the Stevens case engaged in “significant, widespread
and, at times, intentional misconduct.” By then, one of those prosecutors had
hanged himself.
The Stevens case is extreme, disturbing and tragic. The judge who dismissed the charges said “In
nearly 25 years on the bench, I’ve never seen anything approaching the mishandling and
misconduct that I’ve seen in this case,” and he cited several prosecutors for criminal contempt.
But less notorious cases have disturbing similarities.
In 2014, Reddy Annappareddy was convicted on charges that he bilked Medicaid,
Medicare, and private insurers out of millions of dollars by submitting fraudulent
billings for prescriptions filled by pharmacies he owned in Maryland and
Washington, D.C. In 2017, the trial judge dismissed the charges and said that the
failure to disclose exculpatory evidence and “the history of late disclosures and
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 112 • National Registry of Exonerations • September 1, 2020
the promotion of false significant testimony in this case does shock the
conscience of this Court.”
Among other things, the prosecution had concealed an audit that indicated that
alleged losses were far smaller than the government claimed (a thorough
accounting eventually found none at all); and the FBI had repackaged expired
drugs from a clinic that the defendant’s company was holding for destruction,
and claimed in court that the pharmacies had obtained those drugs using
fraudulent prescriptions and never delivered them to patients.
2. MISCONDUCT BY PROSECUTORS
As we have noted, state-court exonerations include more misconduct by police than by
prosecutors. Even among murder exonerations, where the overall rate of official misconduct is
comparable to federal white-collar cases, misconduct by police is slightly more common than
misconduct by prosecutors, 48% to 44%. In federal cases generally, prosecutors commit
misconduct more than two-and-a-half times as often as police officers. This disproportion is
greatest among white-collar exonerations, where federal prosecutors committed misconduct
seven times as often as police. In fact, every single federal white-collar exoneration with official
misconduct includes misconduct by a prosecutor.
This striking pattern no doubt reflects the structure of white-collar prosecutions, at least in
federal court.
195
Judging from exonerations, they are overwhelmingly the products of lengthy
proactive investigations. Nearly 60% are multi-defendant exonerations (26/46), and, although
in general federal prosecutors often work with state and local police, at least two-thirds of these
cases included federal law enforcement agencies (30/46). It appears that federal prosecutors
were deeply involved in the investigations that led to these white-collar prosecutions, and
probably initiated many of them. By contrast, state prosecutors usually first hear about a violent
crime—even a murder—when police ask them to charge a defendant who has already been
identified, arrested, and, frequently, interrogated.
That level of prosecutorial control over criminal investigations means that federal prosecutors
are more likely to know what the police know, and to play a role in any official misconduct that
occurs—especially if the misconduct is concealing exculpatory evidence.
As we’ve noted, the primary duty to disclose exculpatory evidence falls on prosecutors.
196
If
police tell prosecutors about such evidence, they’ve done all they’re required to do. Therefore,
we only count the police as responsible for concealing evidence if there is no indication that
prosecutors knew about the evidence in question. Even so, police officers were responsible for
concealing exculpatory evidence in 46% of state murder exonerations where such evidence was
hidden (165/355), compared to 65% for prosecutors (230/355). But among the federal white-
collar cases in the first 1,361 exonerations posted in the Registry, prosecutors were responsible
195
Seventeen white-collar exonerations were in state courts, 27% of the total. Of these, 53% included official
misconduct (9/17)—35% misconduct by prosecutors (6/17) and 18% by police (3/17).
196
See supra Section VI.3
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 113 • National Registry of Exonerations • September 1, 2020
for withholding exculpatory evidence in every single case in which such evidence was concealed
(12/12), and police in none.
On the other hand, witness tampering is a form of misconduct that’s overwhelmingly committed
by police officers.
197
It occurred in 32% of all state murder exonerations (286/904), and 88% of
those cases included misconduct by police (252/286). Witness tampering only occurred in 17%
of federal white-collar exonerations (8/46), and in every case it was done by prosecutors; we
know of no police misconduct in any of those cases.
Some of these discrepancies may in part reflect differences in record keeping and in the
structure of the cases. Federal prosecutors may keep more complete records than state
prosecutors because they have more resources at their disposal, or are subject to more
demanding departmental rules. If so, some of the difference in assignment of responsibility for
concealing exculpatory evidence may simply mean that better record keeping reveals
prosecutorial knowledge of the concealed evidence that might otherwise not be noticed.
Similarly, witness tampering by police may be rare in part because the document-heavy
investigations of federal white-collar crimes are less likely to turn on the type of testimony that
police officers obtain by tampering with witnesses in violent crime cases. For example, two-
thirds of all exonerations include mistaken or intentionally false identifications of the
defendants (1,778/2,400), compared to one-third of federal white-collar exonerations (15/46).
It’s hard, however, to think of an innocent explanation for the rate of federal prosecutorial
misconduct in an arena that is always their domain: trial. Prosecutors knowingly permitted
witnesses to commit perjury in 26% of the trials of exonerated federal white-collar defendants
(12/46)—almost twice as often as state prosecutors at the trials of murder exonerees (134/904);
and they committed some form of misconduct in 39% of those trials (18/46), compared to 22%
for state murder prosecutors (196/904).
The picture that emerges suggests that prosecutors dominate federal white-collar prosecutions
in a way that largely determines the extent and nature of all official misconduct, for better and
for worse. Prosecutorial control over these investigations may well discourage and reduce
misconduct by law enforcement officers. It’s possible that it reduces the overall rate of
misconduct in such cases; we don’t have the sort of data that would permit even an educated
guess on the overall frequency of official misconduct in any category of criminal cases, as we’ve
noted.
198
However, when there is official misconduct in federal white-collar exonerations, prosecutors are
at the heart of it. They may be able to prevent police misconduct more effectively than state
prosecutors, but they don’t always do it. The FBI, for example, committed serious misconduct in
the investigations of Ted Stevens and of Reddy Annappareddy, but in both cases—and every
other similar federal case with police misconduct—so did the prosecutors.
197
See supra Section IV.1.
198
See supra Section II; infra Methodological Appendix.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 114 • National Registry of Exonerations • September 1, 2020
Why do federal prosecutors commit—and permit—misconduct in white-collar prosecutions?
Probably for the same combination of reasons that produces the high rate of official misconduct
in murder exonerations: They are strongly motivated to prosecute serious crimes that are
difficult to prove; they are convinced of the defendants’ guilt but afraid they might not have
enough evidence to convince a judge and jury; and, unlike almost all criminal prosecutions,
these cases receive a lot of public attention, so failure is humiliating but success can advance a
professional or political career. Most prosecutors undoubtedly respond to these pressures by
keeping their heads down and doing their job—but some, as we see, cut corners, jump to
conclusions, hide evidence that suggests innocence or manufacture evidence of guilt.
It’s easy to see how it can happen.
Fraud in the provision of prescription medicines under Medicare and Medicaid is a huge
national problem, but difficult and costly to prove. What should a prosecutor do after years of
investigating what she believes is a major prescription drug fraud mill, when the evidence just
isn’t there? Drop the case, of course—and ideally reconsider her belief that the defendant is
guilty. But the temptation to conceal and deceive must be strong if that prosecutor remains
convinced of the defendant’s guilt, and has committed herself to that position in public.
By the time Alaska Senator Ted Stevens was charged with receiving illegal gifts, there had been
years of stories and rumors that he had used his public position improperly to amass personal
wealth.
199
But he had never been charged with a crime or other actionable misconduct. The
prosecutors who did charge him were probably convinced of his guilt, but as the evidence came
in, they had no case. So, at a trial in the glare of the national media, they concealed evidence that
would have led to an acquittal if not a dismissal.
We know of dozens of other federal white-collar exonerations in which prosecutorial misconduct
led to false convictions. We do not know how often that has happened in cases of innocent
defendants whose convictions still stand.
199
See Chuck Neubauer & Richard T. Cooper, Senator’s Way to Wealth Was Paved with Favors, Los Angeles Times
(Dec. 17, 2003).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 115 • National Registry of Exonerations • September 1, 2020
X. Discipline
1. IN GENERAL
How many of the prosecutors, police officers, forensic analysts and child welfare workers are
disciplined for misconduct that leads to false convictions? It happens, but not often. We know
that some discipline, broadly defined, was imposed in 17% of exonerations with known official
misconduct (219/1,295), but not in the remaining 83%.
Misconduct often comes in bunches. In most cases in which discipline for misconduct was
imposed, the officials who were disciplined committed misconduct of the same general sort in
several or many other cases as well—sometimes cases that produced exonerations, sometimes
not—but formal discipline was limited to one or a few, or was not attached to any specific case.
For example, in 2011 Chicago Police Commander Jon Burge—who had already been fired from
the police department for a pattern of misconduct over several years—was convicted of perjury
for lying about his role as the leader of a group of officers who tortured numerous suspects.
200
We view that conviction as discipline that was imposed for every case in which he was
responsible for misconduct of that type. Overall, in 70% of exonerations with discipline for
official misconduct, that discipline was imposed for general patterns of behavior or for a case
other than the specific exoneration at hand (154/219).
Disciplinary proceedings may take months or years from the initiation to conclusion. We only
count those cases in which punishment was actually imposed, occasionally in reduced measure.
We do not count 1% of exonerations with misconduct in which discipline was imposed but later
reversed (15/1,295), or a handful in which, as of our last information, proceedings were still
pending final resolution (10/1,295).
Prosecutors who committed misconduct in criminal cases that led to exonerations were rarely
disciplined; it happened only 4% of the time. Police officers who committed misconduct were
disciplined almost five times as often as prosecutors, in 19% of the cases; and forensic analysts
who committed misconduct were more than twice as likely to be disciplined as police officers, in
200
See infra , Section XII.1.b.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 116 • National Registry of Exonerations • September 1, 2020
47% of the cases. We know of no exoneration in a case in which a child welfare worker was
disciplined. See Table 22.
Table 22: Rate of Discipline among Exonerations with Misconduct, by Type of Official
who Committed Misconduct*
Misconduct By:
Prosecutor
Police
Officer
Forensic
Analyst
ANY LAW
ENFORCEMENT
OFFICIAL
4%
(26/729)
19%
(160/843)
47%
(35/75)
17%
(219/1,295)
* Some cases include misconduct by more than one type of official.
Prosecutors and police officers are essential players in all criminal cases; forensic analysts
participate in a minor fraction. As a result, misconduct by forensic analysts occurred in only
about 3% of all exonerations—and, despite the comparatively high rate of discipline for that
misconduct, it only happened in about 1% of all cases (35/2,400). Prosecutors were also
disciplined in about 1% of all exonerations (26/2,400), and police officers in about 7%
(160/2,400).
Among the major crimes that account for about five-sixths of known exonerations, the rate of
discipline varies greatly. It’s highest by far in drug-crime exonerations, where some form of
discipline was imposed in 74% of cases with official misconduct, followed at some distance by
sexual assault at 17%, and murder and robbery at 9% each. White-collar and child sex abuse
exonerations bring up the rear, with discipline in 3% of cases with misconduct each. See Table
23.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 117 • National Registry of Exonerations • September 1, 2020
Table 23: Rate of Discipline among Exonerations with Misconduct, by Type of Official
who Committed Misconduct, for Major Crime Categories*
Misconduct By:
Prosecutor
Police
Officer
Forensic
Analyst
ANY LAW
ENFORCEMENT
OFFICIAL
Murder
4%
(15/404)
7%
(32/435)
38%
(13/34)
9%
(59/652)
Child Sex Abuse
3%
(2/67)
0%
(0/76)
25%
(1/4)
3%
(3/120) **
Sexual Assault
3%
(2/64)
3%
(2/68)
59%
(17/29)
17%
(21/126)
Robbery
9%
(3/33)
5%
(1/19)
-
9%
(4/46)
Drug Crimes
3%
(1/29)
84%
(90/107)
100%
(1/1)
74%
(91/123)
White-collar
Crimes
3%
(1/36)
0%
(0/7)
0%
(0/1)
3%
(1/39)
ALL CASES WITH
MISCONDUCT (1,295)
4%
(26/729)
19%
(160/843)
47%
(35/75)
17%
(219/1,295)
* Some cases include misconduct by more than one type of official.
** The total number of child sex abuse cases with misconduct includes 48 cases with misconduct by child welfare
workers, most of which also include misconduct by other officials.
The rates of discipline for particular types of law enforcement officials vary a good deal for
prosecutors and forensic analysts—from 3% to 9% for prosecutors, and from 25% to 59% for
forensic analysts (not counting two crime categories with only one case of misconduct each).
For police, however, the range is huge. Discipline was imposed in 84% of 107 drug crime
exonerations in which a police officer committed misconduct, but in none of 76 child sex abuse
cases in which they did so. See Table 23.
“Discipline” for official misconduct can come from three separate sources of authority:
Employment. The agency that employs misbehaving officials may reassign, suspend,
demote or fire them, or pressure them to resign.
Professional. If the official has a professional license or certification, the licensing or
certifying authority (for example, a state bar) may revoke or suspend that license or
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 118 • National Registry of Exonerations • September 1, 2020
certification, or reprimand the official who holds it.
201
We do not know of any
professional discipline of police officers who worked on the cases we list.
Criminal. If the misconduct is a crime, the official may be prosecuted and convicted.
202
In the normal course of events, law enforcement officers who are convicted of crimes related to
their work are also disciplined by their employers. In most cases, they are fired if they weren’t
already forced to resign or retire. Similarly, a convicted official who has a license or a certificate
is very likely to have it revoked or suspended, or to be reprimanded or otherwise disciplined by
the professional association involved. We only tabulate employment or professional discipline
for officials who were not convicted of crimes based on the same pattern of misconduct. For
officials who were convicted of crimes, those forms of discipline (if applicable) should be
assumed.
Prosecutors who committed misconduct were highly unlikely to be disciplined in any manner, as
we mentioned. In particular, in only two of 727 cases with prosecutorial misconduct was a
prosecutor convicted criminally. By contrast, in 127 exonerations, police officers who worked on
the cases were convicted of crimes because of misconduct that contributed to false convictions;
and forensic analysts were disciplined by their employers in more than a third of exonerations in
which they committed misconduct. See Table 24.
201
For prosecutors, we also count judicial discipline—reprimanding prosecutors, holding them in contempt, and
removing them from the case in question—as professional discipline. That occurred in three cases.
202
Successful civil lawsuits for monetary damages could, in theory, be a form of discipline for the offending parties,
but we don’t count them here for a variety of reasons. The most important are that civil damages are almost always
paid by the agencies or local governments that employ the offending officials, or by their insurance companies, rather
than the official actors themselves; and that most successful claims end in settlements, with no findings or
admissions of misconduct.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 119 • National Registry of Exonerations • September 1, 2020
Table 24: Rate of Discipline among Exonerations with Misconduct, by Type of Official
who Committed Misconduct and Form of Discipline*
Misconduct By:
Prosecutor
Police
Officer
Forensic
Analyst
ANY GOVERNMENT
OFFICIAL
Form of
Discipline
Employment**
2%
(11/729)
4%
(33/843)
37%
(28/75)
5%
(70/1,295)
Professional**
2%
(14/729)
-
9%
(7/75)
2%
(21/1,295)
Criminal
0.3%
(2/729)
15%
(127/843)
4%
(3/75)
10%
(133/1,295)
ANY TYPE OF
DISCIPLINE (1,295)
4%
(26/729)
19%
(160/843)
47%
(35/75)
17%
(219/1,295)
* Some cases include more than one type of discipline and some cases misconduct by more than one type of
official.
** Employment and Professional discipline are not listed for any official who was convicted of a crime in the same
case or a similar case.
As with misconduct itself, we have missed cases in which discipline was imposed. That is least
likely for criminal convictions of law enforcement officials—especially of prosecutors—which are
uncommon and conspicuous events. It is most likely for comparatively mild discipline by
employers, such as reassignment or demotion, which may go unnoticed if it is not deliberately
hidden.
In the sections that follow, we focus separately on each of these three categories of official
actors, and discuss differences in discipline by crime and by type of discipline that was imposed.
2. DISCIPLINE BY CATEGORY OF GOVERNMENT OFFICIAL
a. Prosecutors
Prosecutors are hardly ever disciplined for misconduct that contributes to false convictions.
We’ve probably missed some cases in which prosecutors received employment or professional
discipline that did not end their careers in prosecution—suspensions, private reprimands, and
so forth—but the general pattern is no surprise. Several earlier studies found the same pattern in
other contexts, as we have discussed.
203
203
Supra Section II.1. See also Ken Armstrong & Maurice Possley, Trial & Error: How Prosecutors Sacrifice Justice
to Win. Part 1: The Verdict: Dishonor, Chi. Trib. (Jan. 11, 1999); Ctr. for Pub. Integrity, Harmful Error:
Investigating America’s Local Prosecutors (2003); Kathleen M. Ridolfi & Maurice Possley, Preventable Error: A
Report On Prosecutorial Misconduct In California, 1997–2009, A Veritas Initiative Report (2010).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 120 • National Registry of Exonerations • September 1, 2020
We know of some discipline for prosecutors in 4% of exonerations with prosecutorial
misconduct: 15 murders, nine other violent felonies, a drug case and a white-collar crime. In
most of those cases, the discipline that was imposed was comparatively mild.
Eleven prosecutors were disciplined by the offices that employed them: two were fired, four
resigned or retired under pressure, and five were demoted, suspended or received additional
training. Fourteen prosecutors were disciplined in their professional capacity as lawyers, but
only three were disbarred, one of them for misconduct in the case of two exonerees who were
tried together. Two others were suspended, four were reprimanded, and three received
sanctions from judges that were limited to the cases at issue.
Only two prosecutors in exonerations were convicted of crimes for professional misconduct,
both in highly notorious cases, and both received nominal sentences:
204
In 2013, former Williamson County (Texas) District Attorney Ken Anderson served four
days in jail for contempt for concealing exculpatory evidence that would have prevented
the murder conviction of Michael Morton, who spent 24 years in prison.
In 2007, Michael Nifong, the former District Attorney of Durham County, North
Carolina, spent one day in jail, also for criminal contempt, for concealing exculpatory
DNA evidence in a nationally publicized prosecution of three white members of the Duke
University Lacrosse team who were falsely accused of raping a Black exotic dancer.
205
There was no exoneration in the Duke Lacrosse case; the defendants were never
convicted because charges were dropped before trial. In 2016, however, the attention
Nifong received in that case helped secure a murder exoneration for Darryl Howard,
whom Nifong had prosecuted and convicted of murder in 1991, and in whose case he had
also concealed crucial exculpatory evidence. We count the discipline in the Duke
Lacrosse case as applying to Howard’s exoneration because the prosecutor who
committed it was punished for similar behavior in a different case.
Nifong and Anderson were both also disbarred and lost their jobs.
206
204
See supra Section VIII.2.b. and infra Section X.2.a.
205
Shaila Dewan, Duke Prosecutor Jailed; Students Seek Settlement, N.Y. Times, Sept. 8, 2007.
206
One reason that few prosecutors are charged with crimes for misconduct in criminal cases may be that the
common form of state statutes of limitations makes that impossible in most cases. Ken Anderson, for example, could
have been charged with evidence tampering under Texas Penal Code §§ 37.09 and 37.10, but only if charges had
been filed within three years of the commission of the misconduct—which, as is typical for exonerations, would have
been more than 20 years before his misconduct was discovered. Instead he was charged with criminal contempt
under Texas Government Code § 21.002(a) because he violated a court order and lied to the trial judge; the statute
of limitations for that offense only begins to run when the contempt is discovered. Barry Scheck, Four Reforms for the
Twenty First Century, 96 Judicature 323 (2013) at 332 n.50. In 2016, California enacted Penal Code §141(c), which
makes it a felony for a prosecutor “to intentionally and in bad faith” conceal or destroy exculpatory evidence in a
criminal case. The applicable statute of limitations is California Penal Code § 803(c), which provides that “A limitation
of time prescribed in this chapter does not commence to run until the discovery of an offense….” That solves the
statute of limitations problem—if the authorities are interested in filing felony charges for concealing exculpatory
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 121 • National Registry of Exonerations • September 1, 2020
b. Police officers
Police officers were disciplined for misconduct in exonerations more than six times as often as
prosecutors, 160 times vs. 26. In 79% of those cases, officers were convicted of crimes (127/160);
in 20%, officers were disciplined by the police forces for which they worked (33/161); we know
of no cases in which police officers received professional discipline for misconduct that
contributed to false convictions.
These numbers, however, may be misleading in two respects: (i) On the one hand, the number
of officers who were disciplined for misconduct, especially the number of officers who were
convicted of crimes, is considerably smaller than the number of exonerations in which those
officers helped secure false convictions. (ii) On the other hand, we know that many instances of
discipline of police officers short of criminal conviction are concealed from public view; we don’t
know how often that occurred among exonerations with misconduct, or in which cases.
i. Criminal conviction in related cases
We know of 92 exonerations in Chicago in which police officers committed misconduct in
investigating the cases and were later convicted of crimes for that misconduct or for similar
misconduct in other cases, and another three such exonerations in a Chicago suburb within
Cook County, Illinois. However, that does not mean that these cases produced 95 separate
criminal convictions of police officers. Far from it.
Sixty-six of the 95 Cook County cases were in the Sergeant Watts group exoneration.
207
In all 66
cases, the defendants were framed for possession of drugs (or in one case, a gun) that officers
planted. The findings of discipline in these exonerations were based on two convictions for acts
that were part of the same pattern of criminal misconduct, but occurred in a single sting case in
federal court: the conviction of Sergeant Ronald Watts himself, who participated directly in
some of the 66 exonerations and supervised the officers in all of them, and that of his
subordinate, Officer Kallatt Mohammed, who played a direct role in several of the Watts
exonerations.
An additional 19 Cook County exonerations with criminal discipline are torture cases in which
Chicago Police Commander Jon Burge—who was convicted of perjury in a related federal case—
played a role as supervisor or direct participant.
In sum, 85 of the 95 cases of criminal discipline of police officers for misconduct in exonerations
in Cook County are based on three criminal convictions.
208
More telling yet: those three
evidence. So far, we are not aware of any charges that have been filed alleging violations of California Penal Code
§141(c).
207
See supra Section III.3.c.i.
208
The remaining 10 Cook County cases with criminal discipline of police officers include six exonerations based on
misconduct by the officers in the now disbanded “Special Operations Section” [SOS] of the Chicago Police
department; a total of eleven officers were convicted of crimes for their work in that unit. In addition, two individual
Cook County cases had one conviction of an officer each; and a set of three related cases led to two convictions. The
32 exonerations outside of Cook County with criminal convictions of police officers also include several groupings of a
few cases each. The most noteworthy are the Norfolk Four exonerations in Virginia whose confessions were
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 122 • National Registry of Exonerations • September 1, 2020
convictions of Watts, Mohammed and Burge account for two-thirds of all exonerations in the
country in which police officers who conducted or supervised the investigations were convicted
of crimes for the type of misconduct they committed in those cases (85/127). Indeed, the 66
Watts exonerations alone comprise more than half the national total.
The large number of criminal convictions for misconduct in the Watts cases is a direct
consequence of the structure of that group exoneration.
209
As we discussed, most group
exonerations are cases that only came to light after it becomes known that the officers involved
were guilty of concerted schemes of misconduct, usually planting drugs on innocent suspects.
That is also true for some of the Burge exonerations: they were identified as candidates for
exoneration because Burge was involved, after he was exposed for presiding over a program of
systematic torture, fired and eventually convicted.
For the Watts cases, the exonerations were also a direct response to the discipline itself—the
convictions of Watts and Mohammed in federal court. Since the Watts cases were identified for
exoneration by the participation of officers who’d been convicted of crime, all of them, by
definition, include criminal discipline for those officers.
Neither Watts nor any of his subordinates was convicted of any crimes directly related to the
many extreme miscarriages of justice they perpetrated. Internal police investigations of 15 other
officers were still pending two-and-a-half years after the first large group of Watts exonerations
and nearly seven years after Watts was convicted in federal court.
210
As far as we know, there
have been no state criminal investigations of Watts or any of the officers who worked under him.
Even though many fewer police officers were convicted of crimes than the number of cases with
such convictions suggests, police are far more likely than prosecutors to be criminally convicted
and imprisoned for misconduct that leads to false convictions. We don’t have a precise count of
the police officers who were convicted across these 127 exonerations—in some cases, there may
be more convictions than we have been able to identify—but we know of at least 30, and some
served substantial periods of time in prison.
211
Burge was sentenced to 4 1/2 years in prison,
Watts to 22 months, and Detective Robert Ford, who coerced the confessions of four innocent
sailors in a Virginia rape-murder case, was sentenced to 12 1/2 years in prison for extortion and
for lying to the FBI about related misconduct.
extracted by the misconduct of a single officer who was later convicted of related misconduct, and the convictions of
five murder codefendants in Los Angeles, who were framed by a single officer who went to prison for a pervasive
pattern of misconduct on the Rampart Group Exoneration.
209
See supra Section III.3.c.i.
210
Editorial Board, Take the lid off probes of officers tied to wrongful convictions, Chicago Sun Times, Jun 14, 2020.
211
We know of 18 officers who worked in Cook County, Illinois, who were among those convicted in crimes for
misconduct they committed in interrogations, or for similar misconduct in other cases—almost certainly more than
half the total for the country. On this issue, as with false confessions and the use of violence in interrogation, Cook
County is unique.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 123 • National Registry of Exonerations • September 1, 2020
The only two prosecutors who were convicted of any sort of criminal behavior in connection
with exonerations served a total of five days in jail between them.
ii. Unreported employment and professional discipline
As we mentioned, we don’t know of any police officer who was disciplined by a professional
organization for misconduct that contributed to the conviction of an exoneree. But it probably
happened.
Forty-four states require that police officers be certified or licensed after taking a training
program that satisfies state requirements. Those requirements are typically administered by a
state Police Officer Standards and Training (“POST”) Commission that also has the authority to
decertify an officer or revoke their license.
212
That disciplinary process, however, is generally
secret. There is a national database of decertifications, the National Decertification Index (NDI),
that is maintained by the International Association of Directors of Law Enforcement Standards
and Training, but access is limited to law enforcement and POST organizations.
213
Nonetheless,
USA Today and The Invisible Institute in Chicago have put together a database with information
on some 30,000 decertifications stretching back over dozens of years.
214
Most were based on use
of excessive force, drug or alcohol abuse, domestic violence, or other misconduct unrelated to
criminal investigations, but in about 15% of cases in which reasons for decertification are listed,
they include “dishonesty” or “official misconduct.”
215
Decertification means that the officers involved lost their jobs in the police forces at which they
worked when the misconduct occurred (although decertified officers are sometimes hired by
other police departments
216
). Terminating a police officer’s employment is a reasonably visible
event. Nonetheless, it’s likely that unbeknownst to us, some officers who helped obtain
convictions of exonerees were decertified for misconduct in those cases—and it’s very likely that
we have missed decertifications for similar misconduct in other criminal cases.
It is also highly likely that we missed cases in which police officers were disciplined by their
departments but not terminated. In some states—including Florida, Texas, Minnesota and
Arizona—employment information on police officers, including their disciplinary records, is
available to the public.
217
In others, including New Hampshire, Colorado and Vermont, that
212
Roger Goldman, Importance of State Law in Police Reform, 60 Saint Louis Univ. L. J. 363, 381 (2016); See Barry
Scheck, The Integrity Of Our Convictions: Holding Stakeholders Accountable In An Era Of Criminal Justice Reform,
48 Geo. L.J. Ann. Rev. Crim. Proc. iii, xviv-xxv (2019).
213
“About NDI,” International Association of Directors of Law Enforcement Standards and Training.
214
John Kelly and Mark Nichols, Search the list of more than 30,000 officers banned by 44 states, USA Today, Oct.
14, 2019.
215
John Kelly and Mark Nichols, We found 85,000 cops who’ve been investigated for misconduct. Now you can read
their records, USA Today, June 11, 2020.
216
See Goldman, supra note 212.
217
Jonathan Abel, Brady’s Blind Spot: Impeachment Evidence in Police Personnel Files and the Battle Splitting the
Prosecution Team, 67 Stan. L. Rev. 743, 770 (2015).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 124 • National Registry of Exonerations • September 1, 2020
information is secret;
218
California was in that group until late 2019.
219
And in some states,
including New York, Maryland, West Virginia and Oregon, prosecutors have the authority to
access police disciplinary records (and then, presumably, disclose them to the defense if they
contain exculpatory evidence)—but many prosecutors don’t know they have that access, or don’t
use it.
220
And if records are available, they may be incomplete: many departments have contracts
with police unions that require that disciplinary records be erased after a period that may be as
short as two years or less.
221
Disciplinary measures that are hidden from view may be better than no discipline at all, but they
have limited impact. They are less likely to deter officers from committing misconduct, and they
make it harder for the other participants in the process—prosecutors, defense attorneys, judges
and juries—to identify those officers who commit misconduct and lie about it repeatedly.
c. Forensic Analysts
All of the forensic analysts who were disciplined for their roles in convictions that led to
exonerations had committed forensic fraud. In four-fifths of those cases—28 out of 35—the
punishment was imposed by the agencies they worked for.
In three cases, forensic analysts were convicted of crimes: one for a lab scandal that included
writing false drug analysis reports; and two others for lying about their credentials in trial
testimony.
222
In seven cases, forensic analysts had professional licenses or memberships in
professional societies suspended or revoked (including three in which the analyst was also
disciplined by her employer). In 18 cases, the forensic analysts were fired from their jobs
(counting one who resigned as he was about to be fired), and in 10 others, the analyst was
reassigned to administrative duties.
The rate of discipline for forensic analysts is comparatively high: 47% of all known cases in
which they committed misconduct, and 56% of cases in which they committed forensic fraud,
that we know of (35/62). But—as with discipline for police officers—the number of analysts who
were disciplined is considerably smaller than the number of exonerations in which those
analysts participated.
In most cases, the analysts who were disciplined were serial offenders. Altogether, only 13
different analysts were disciplined for their misconduct across these 35 cases—and six of them
were responsible for 80% of the total (28/35). In fact, in nearly half of all cases with discipline
218
Id. at 762-770.
219
Marco della Cava, When police misconduct occurs, records often stay secret. One mom’s right to change that,
USA Today, Oct. 14, 2019.
220
Abel, supra note 217, at 775-779.
221
DeRay McKesson, et al., Police Union Contracts and Police Bill of Rights Analysis, June 29, 2016.
222
The sentences were comparatively mild: 30 days, 9 months home confinement, probation. The three are: Dollard,
Jermaine (linked); LeFever, Virginia; Kotler, Kerry.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 125 • National Registry of Exonerations • September 1, 2020
for a forensic analyst, the offending analyst was either Fred Zain (seven exonerations in West
Virginia and Texas) or Pamela Fish (10 cases in Chicago).
As we noted earlier, Zain, Fish, and other serial fraudsters are likely to be overrepresented
among known cases of forensic fraud.
223
After they are exposed, otherwise inconspicuous cases
they worked on are likely to be reexamined and misconduct that had been missed brought to
light. At the same time, their notoriety also makes them likely targets for discipline.
3. SUMMARY
The rate of known discipline for misconduct that led to the convictions of innocent exonerees is
low, 17% all told. That’s an undercount, inevitably—especially for police discipline by employers,
which may fly under the radar, or, in some departments, be concealed.
Very few prosecutors were disciplined in any manner. It occurred in 4% of exonerations with
prosecutorial misconduct, 26 cases altogether. In each case, the prosecutor was disciplined for
misconduct in the prosecution of an individual defendant or a pair of codefendants. Only two
prosecutors were convicted of crimes for their misconduct, both for contempt of court, and both
received minimal sentences.
Police officers and forensic analysts were much more likely to be disciplined; it happened in 19%
and 47% of the cases in which we know that they committed misconduct, respectively. In almost
80% of exonerations in which police officers were disciplined, they were convicted of crimes; by
contrast, in 80% of cases in which forensic analysts were disciplined, they were fired or
reassigned by their agencies to non-investigative duties.
There are many more exonerations in cases in which officers or analysts were disciplined than
officers or analysts who were the subjects of that discipline. In 30% of all known exonerations
with discipline, that discipline consisted of two criminal convictions, those of Chicago Police
Sergeant Ronald Watts and Officer Kallatt Mohammed, who supervised or participated in
framing 66 innocent drug defendants who were later exonerated (66/219). More than half of all
exonerations with discipline are based on punishments meted out to seven individuals: Watts
and Mohammed plus three other police officers
224
and two forensic analysts
225
(111/219).
As we have pointed out, the notoriety of some analysts who committed forensic fraud on a
regular basis drew attention to cases that would otherwise be overlooked, and produced
exonerations of innocent defendants who would not otherwise have been cleared.
226
The same
applied to Jon Burge and the torture cases he presided over
227
—and in both contexts, the same
223
See supra Section VI.1.
224
Jon Burge in Chicago (19 cases), supra Section V.3; Rafael Perez in Los Angeles (5 cases); and Robert Ford in
Norfolk, Virginia (4 cases).
225
Pamela Fish (10 cases): and Fred Zain (7 cases).
226
Supra Section X.2.c.
227
Infra Section XII.1.b.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 126 • National Registry of Exonerations • September 1, 2020
process also produced disciplinary actions that applied across all the cases involved, in both
groups.
The connection is even more direct for the 66 Watts group exonerations in Chicago. None of
these comparatively low-level drug convictions would have resulted in exoneration, regardless of
the defendants’ innocence, if they had not been part of a notorious group. And they all include
discipline because it was the discipline itself—the federal convictions of two officers involved,
and the publicity that followed—that led to the reinvestigations that produced these
exonerations.
While the publicity surrounding notorious malefactors such as Sergeant Watts and Commander
Burge has led to many exonerations—85 in Cook County alone—it has not led to a corresponding
imposition of discipline against other police officers who participated in the same misconduct.
No state criminal charges have been filed against any officer connected with these exonerations,
and, as far as we know, the only officers disciplined in any manner are the three who were
convicted of federal crimes.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 127 • National Registry of Exonerations • September 1, 2020
XI. Changes in Official
Misconduct over Time
1. IN GENERAL
Has official misconduct in criminal investigations and prosecutions in the United States
changed over time? Is it more common than in the past? Less common? Different? The answers
to these questions would be highly useful for any attempts to reduce misconduct in the future.
We can say something about this issue, but less than we’d like.
There are two major difficulties:
We don’t know about most misconduct in criminal cases, as we’ve discussed.
228
As best we can
tell, the majority of victims of torture in Chicago were guilty of the crimes they were convicted
of. Others were never convicted at all. Neither group appears in our data because the
misconduct did not contribute to false convictions. We also don’t know about misconduct that
did contribute to false convictions in cases that did not produce exonerations. And even among
the exonerations we report, we don’t know how much official misconduct has remained
successfully hidden from view. The best we can do is report on changes in known misconduct
that contributed to the false convictions in known exonerations.
Measuring change in this domain is very difficult. In part, that follows from the previous
problem: since we don’t know how often this behavior actually occurs, it’s hard to tell whether
its frequency has changed. But even if we limit ourselves to what we can observe—the numbers
of exonerations with known misconduct, or with particular types of known misconduct—
changes over time are extremely difficult to interpret. The main problem is the time lag from the
commission of misconduct to its discovery.
Usually, when we discuss a date in this report we are interested in the time of exoneration, the
final act in the legal story of a wrongful conviction. Here, however, we are interested in official
misconduct that contributed to the false conviction, the erroneous legal judgment that an
exoneration eventually corrected. We peg that behavior to the date of conviction, the last
possible time when misconduct can influence a criminal conviction.
228
Supra Section III.2.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 128 • National Registry of Exonerations • September 1, 2020
After that, of course, it may be years before the defendant is exonerated and becomes eligible for
inclusion in our database. The average time from conviction to exoneration is about 11 years for
all exonerations; for murder—the most common category of exoneration and the one with the
highest rate of official misconduct—the average lag is almost 15 years, and more than 60 murder
exonerations happened more than 30 years after conviction.
These time lags mean that we may have to wait decades before we can use data on exonerations
to accurately identify changes in official misconduct that might already have occurred in murder
convictions in the past few years.
To illustrate, there were 414 murder exonerations with official misconduct among convictions in
the 16-year period from 1987 through 2002, but only 68 such exonerations from convictions in
the next 16 years, from 2003 through 2018—one-sixth as many.
At first glance, this suggests a sharp decrease in false murder convictions caused by misconduct,
starting in the early years of the twenty-first century—but that impression is misleading. For
cases with convictions before 2003, the average time from conviction to exoneration for a
murder with official misconduct was more than 17 years. Assuming that time lag is similar for
more recent cases, we still don’t know about most exonerations that will eventually accumulate
for murder convictions with official misconduct that occurred in 2003, let alone those in 2017 or
2018. Twenty years from now, when we know about virtually all murder exonerations that will
ever occur for convictions from 2003 through 2018, we might see a reduction in misconduct
compared to the period before 2002, or an increase, or no change.
In short, on many issues and for most types of exonerations, it’s simply too early to tell whether
there has been change in the occurrence of misconduct over time.
229
The problem is particularly
severe for detecting reductions in misconduct: as we saw, an apparent decline may just reflect a
time lag.
But there are exceptions. In a few contexts—improper questioning of children, misconduct in
interrogations and forensic fraud—we see strong evidence of decreases in official misconduct
even with these limited data.
In addition, in one setting—federal white-collar crime prosecutions—we see clear evidence of a
recent increase in official misconduct. As we will see, detecting an increase is easier in this
context than detecting a decrease; if anything, the time lag means we may underestimate it.
229
The same issue can occur in medicine. Imagine that workers in canneries have been contracting a form of
stomach cancer at a high rate because of workplace exposure to lead, but that starting in 2010 canaries have
provided protective clothing that may reduce that exposure and prevent many of these cancers. If the average time
from initial exposure to diagnosis of the disease is 25 years, we may have to wait until 2030 or 2040 to begin to learn
whether the protective clothing is effective.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 129 • National Registry of Exonerations • September 1, 2020
2. IMPROPER QUESTIONING OF A CHILD VICTIM
230
Improper questioning of children has decreased sharply since the mid-1990s. It occurred in 24%
of child sex abuse exonerations (64/270) overall, but 77% of those cases were part of the child
sex abuse hysteria epidemic that lasted from the mid-1980s into the late 1990s (49/64). The last
child sex abuse hysteria exonerations were for convictions in 1998; since then, only a handful of
ordinary child sex abuse exonerations have included this improper and often abusive form of
interrogation. Overall, this improper questioning of children occurred in 28% of child sex abuse
exonerations from convictions through 2002 (58/207)—about 4% of all exonerations in that
period—but only 10% of such exonerations from convictions since 2003 (6/63), or less than 1%
of all exonerations.
3. VIOLENCE AND OTHER MISCONDUCT IN INTERROGATIONS
231
The frequency of misconduct in interrogations—and especially, violence in interrogations—has
dropped dramatically in the past 16 years. As we’ve mentioned, violence and other misconduct
in interrogations is primarily an issue in murder investigations. More than three-quarters of
exonerations with misconduct in interrogations are murder cases (126/165), as are 84% of
exonerations with interrogations that included violence or the threat of violence (88/105). The
change in the raw numbers for murder interrogations is dramatic:
Police officers committed misconduct in 119 interrogations in exonerations from murder
convictions through 2002; they used or threatened physical violence in 86 of those
interrogations.
Police committed misconduct in only seven murder interrogations of exonerees who
were convicted since the beginning of 2003—a decrease by a factor of 17 (119/7). They
used or threatened violence only twice in interrogations since the beginning of 2003, a
decrease by a factor of more than 40 (86/2).
This is not, of course, the final word on murder exonerations from convictions between 2003
and 2018. Exonerations yet to come may include additional instances of misconduct and
violence in interrogations, but those future cases are unlikely to reverse a trend as strong as
what we see here.
The proportion of murder exonerations with violent or otherwise abusive interrogations is a
better measure of change than the raw numbers of cases. That proportion is likely to be less
sensitive than raw number of cases to the effects of the time lag from conviction to exoneration,
since both the numerator and the denominator are reduced by as yet incomplete reporting:
15% percent of exonerations of murder convictions before 2003 included misconduct in
interrogations (119/782), but only 6% of exonerations for murder convictions since 2003
(7/126)—a decrease of 73%;
230
See supra Section IV.4.
231
See Section V.3.a.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 130 • National Registry of Exonerations • September 1, 2020
11% of murder exonerations from the earlier period included violent interrogations
(86/782), but only 2% of those for convictions from 2003 on (2/126)—a decrease of
85%.
232
It is hard to imagine how declines that steep could be erased by data on future exonerations.
233
More than 40% of pre-2003 murder interrogations with misconduct (50/119)—and more than
half of those with physical violence (46/85)—took place in Chicago.
234
The change since 2003
occurred both in Chicago and elsewhere. The proportion of murder exonerations with
misconduct in interrogations went from 46% (50/108) to 20% (2/10) in Chicago, and from 10%
(69/674) to 4% (5/116) in other places; use of violence dropped from 46% (50/108) to 20%
(2/10) in Chicago, and from 6% (39/674) to less than 1% (1/116) in the rest of the country.
4. FORENSIC FRAUD
235
Forensic fraud appears to have declined sharply among exonerations for convictions after 2002:
Overall, 5% of exonerations through 2002 included forensic fraud (72/1598), compared
to 0.4% of all exonerations of defendants convicted in 2003 or later (3/802).
Almost half of exonerations with forensic fraud were murder cases (36/75). Among
murder cases, 4% of exonerations though 2002 included forensic fraud (35/782), but
only 1% of those since 2003 (1/126).
Forensic fraud is one of several types of false or misleading forensic evidence (FMFE). The
overall rate of FMFE in exonerations is nearly the same through 2002 (25%, 400/1598), and
since 2003 (24%, 190/802). The steep decline in forensic fraud persists when we look at FMFE
cases separately:
18% of exonerations with FMFE from convictions through 2002 included forensic fraud
(72/400), but only 2% of those with FMFE based on convictions since 2003 (3/190);
232
We see a similar pattern if we limit our comparison to cases of false confessions. Before 2003, 64% of murder
exonerations with false confessions included misconduct in interrogations (119/186), and 46% included violence
(85/185); since 2003, the comparable rates are 39% (7/18) and 11% (2/18).
233
By comparison, exonerations after 2003 were only about 40% less likely than those since 2003 to include witness
tampering, that we now know of, and about 40% less likely to have had exculpatory evidence that was concealed.
The value of a change in the proportion of murder exonerations with misconduct or violence in interrogations as a
measure of change in the occurrence of those types of behavior would be undercut if murder exonerations with
abusive interrogation have longer time lags from conviction to exoneration than other murder exonerations. That is
the case, but the differences are too modest to account for the large disparities in rates of misconduct. For
exonerations of murder convictions through 2002, the average time from conviction to exoneration was 16 years for
all murder cases, 17.2 years for those with any type or misconduct, 18.9 years for those with misconduct in
interrogations, and 19.7 years for those with violent interrogations.
234
See supra Section V.3.a.i.
235
Supra Section VI.1.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 131 • National Registry of Exonerations • September 1, 2020
18% of murder exonerations with FMFE and convictions through 2002 had forensic
fraud (35/195), compared to 3% of such cases with convictions since 2003 (1/30).
As with misconduct in interrogations, we are likely to see additional exonerations with forensic
fraud for convictions since 2003, but the decrease from earlier years is so great that we are
confident the observed decline is real.
236
5. FEDERAL WHITE-COLLAR CRIME CASES
237
The rate of all federal white-collar crime exonerations, and the rate of such exonerations with
official misconduct, have both doubled or more in the past 17 years.
Thirty-two of the 46 federal white-collar exonerees were convicted in the 15 years from 2003
through 2017, more than twice as many as the 14 exonerees who were convicted of federal white-
collar crimes in the 15 years from 1988 through 2002. The number of federal white-collar
exonerations with official misconduct doubled from the earlier to the later period, 20 compared
to 10.
238
The time lag from conviction to exoneration for federal white-collar crimes is comparatively
short, 3.6 years on average for convictions before 2003, compared to 16 years for murders
exonerations. That means that changes in the numbers of exonerations are a better measure of
underlying behavior for federal white-collar cases, since we probably already know about the
great majority of exonerations for such crimes that will ever occur.
More important, the impact of a time lag from conviction to exoneration depends on the
direction of the observed change. It means that an observed decrease in the number of cases
may be misleading because there are more exonerations to come. But when we see an increase—
as we do for the number of federal white-collar exonerations, with and without misconduct—a
time lag to exoneration can only mean that the true increase in federal white-collar exonerations
may be larger than what we see so far because future exonerations may further increase the rate
for recent cases.
In short, judging from exonerations, the number of convictions of innocent federal white-collar
crime defendants has increased sharply since 2003; most of those cases (before and after 2003)
236
Our conclusion is limited to forensic fraud in prosecutions for violent felonies, which account for 95% of
exonerations with forensic fraud. In Section XII.2.b.1, we discuss two forensic analysts who falsified tests in
thousands of low level drug possession cases in Massachusetts between 2004 and 2013. None of those cases have
produced exonerations to date, and it is likely that the great majority of the defendants were guilty. As we discuss
below, that sort of misconduct is easier to conceal in cases that routinely produce quick guilty pleas rather than the
trials that occur in most exonerations.
237
Supra Section IX.
238
This pattern does not show up for other federal exonerations. We know of 32 exonerations from non-white-collar
federal convictions before 2003, and 34 since 2003; 21 of those exonerations from the earlier period included
misconduct, and 17 of those from the later period.
.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 132 • National Registry of Exonerations • September 1, 2020
involved official misconduct; and all federal white-collar exonerations with misconduct,
regardless of time, include misconduct by prosecutors.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 133 • National Registry of Exonerations • September 1, 2020
XII. Discussion and Conclusions
We have covered a lot of ground in some detail. We now pull back and address two fundamental
questions:
Why do law enforcement officials commit misconduct that leads to convictions of
innocent criminal defendants?
What can we do to reduce this sort of misconduct?
On the first question, we conclude that the most important causes of official misconduct in
criminal cases are systemic: pervasive practices that permit if not encourage bad behavior; lack
of the resources needed to train, supervise and conduct high quality investigations and
prosecutions; and ineffective leadership by police commanders, crime lab directors and chief
prosecutors. If these systemic problems are corrected, misconduct is less likely to occur—and
when it does happen, more likely to be counteracted before innocent people are condemned.
On the second question, we are confident that misconduct in criminal cases can be reduced,
perhaps dramatically. It has already happened for a few types of misconduct, as we’ve seen in
the preceding section—but that is no guarantee of further progress, especially since misconduct
has also increased in some contexts in the last two decades.
We discuss several categories of reforms that address the varieties of misconduct we have
examined. All have been tried, at least in part, and all will improve the operation of the criminal
justice system beyond reducing false convictions. Misconduct that leads to convictions of
innocent defendants also does harm in other cases. Guilty defendants are deprived of their
rights, and innocent defendants who are not ultimately convicted are arrested and charged—
they must defend themselves, and may held in custody for long periods. In addition, changes
that prevent this sort of misconduct will also reduce other poor practices that can lead to
errors. Reforms that prevent misconduct that sends innocent people to prison will benefit
criminal justice across the board.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 134 • National Registry of Exonerations • September 1, 2020
1. WHY DO LAW ENFORCEMENT OFFICIALS COMMIT MISCONDUCT?
Let’s start with a few striking examples.
a. Ken Anderson – Routine Practice
A quick review:
239
In 1987, when Ken Anderson was District Attorney of Williamson County,
Texas, he successfully prosecuted Michael Morton for murdering his wife, Christine. To do so,
Anderson concealed the following evidence: neighbors had seen a suspicious stranger hanging
around the Mortons’ house; after the murder, while Morton was in custody, someone else
attempted to use a credit card belonging to his wife and cashed a $20 check that was in her
missing purse; and the Mortons’ three-year old son, who witnessed the killing, told his
grandmother that “a monster” killed his mother when “Daddy was not there.”
What followed was a parade of horribles. In 2011, DNA testing of a bandana found near the
crime scene identified the real killer. The District Attorney’s Office had successfully resisted
testing that bandana for many years. Morton spent 24 years in prison for a crime that he did not
commit—a crime that was itself an unspeakable tragedy for him and his family. The real killer
went on to bludgeon another woman to death in 1988. Anderson himself was disgraced. He pled
guilty to contempt of court, spent four days in jail, was disbarred, and was forced to resign from
the position he then held as a judge.
But why did Ken Anderson conceal all that evidence of Michael Morton’s innocence?
We don’t know. We could ask, but we wouldn’t trust the answer—if any was given—and
Anderson himself may no longer know, if he ever did. Still, it’s worth considering the
possibilities.
Did Anderson deliberately frame a man he believed to be innocent? That seems highly unlikely.
It does happen, but as best we can tell it’s far more common for prosecutors and police officers
to lie, cheat and conceal in order to convict defendants they believe are guilty.
In this case, that may seem like a stretch: the concealed evidence of innocence would have been
devastating to the prosecution’s case. Anderson was an experienced prosecutor; he knew about
the credit card, the check and the statement by the victim’s child. Wouldn’t he at least have had
doubts?
Maybe, but we think probably not. There is extensive evidence that all of us, including
prosecutors, have a hard time paying attention to evidence that contradicts a theory we have
already adopted.
240
Anderson may have known about the discordant evidence but failed to grasp
its importance, or he may have found some way to dismiss it. Sloppiness, laziness and self-
239
See supra Section VII.2.b.
240
Keith A. Findley & Barbara O’Brien, Psychological Perspectives: Cognition and Decision Making in Examining
Wrongful Convictions: Stepping Back, Moving Forward. Bruce R. Acker & Allison D. Redlich, eds., Carolina Academic
Press, 2014; Barbara O'Brien, A Recipe for Bias: An Empirical Look at the Interplay between Institutional Incentives
and Bounded Rationality in Prosecutorial Decision Making, 74 Mo. L. Rev. 999 (2009); Keith A. Findley & Michael
Scott, The Multiple Dimensions of Tunnel Vision in Criminal Cases, 2006 Wisconsin Law Review 291.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 135 • National Registry of Exonerations • September 1, 2020
deception are ubiquitous and powerful. Plus, in most cases prosecutors are right: suspects
arrested by the police are usually guilty. Like anybody else, they can make the disastrous
mistake of confusing usually right with always right.
That account, however—if true—only explains why Anderson probably believed, in the face of
strong evidence to the contrary, that Morton bludgeoned his wife to death. It does not explain
why he broke the rules so flagrantly in order to convict him.
It may be that Anderson concealed critical evidence of innocence for reasons peculiar to the
murder of Christine Morton. Perhaps he was deeply concerned to get a conviction because the
case had attracted a great deal of attention. Our best bet, however, is the opposite. We think it’s
more likely that he concealed inconvenient evidence in Morton’s case because he did it all the
time, because it was his routine practice.
That explanation, if true, just pushes the question back a level: Why did he do it all the time? It’s
tempting to wonder about Anderson’s personality and motivation: Was he amoral? Cynical?
Ambitious? Maybe. Most prosecutors don’t routinely violate the basic rules of fair prosecution
and trial, and most, we are confident, have a more reliable moral compass than Ken Anderson.
Still, the most powerful explanation for Anderson’s behavior is structural. If Anderson routinely
concealed exculpatory evidence, he did it because he could. Nobody and nothing stopped him,
and it made his work easier. In most cases, the defendants were guilty—in many, there would
have been no dispute about guilt—and the overwhelming majority pled guilty. So who would
ever know? And since (he assumed) they’re all guilty, what did it matter?
But Michael Morton was innocent. And how many others? We have no idea. Very few criminal
defendants insist on their innocence, go to trial, and are sentenced to life imprisonment—and
few innocent defendants of any sort can prove their innocence by DNA testing on available
physical evidence. It took that perfect storm to bring this outrage to light; many lesser tragedies
may remain hidden.
Anderson, of course, was the elected head of his office. He set policy; he set an example for his
subordinates. In at least one case, his First Assistant District Attorney, Paul Womack, followed
that example. In 1993, he persuaded Troy Mansfield to plead guilty to second-degree indecency
with a child by threatening Mansfield with a life sentence if he went to trial. In the process,
Womack concealed—among other items—this memo in the prosecution’s file: “Victim will be
difficult to sponsor in Court. She told me she does not remember what happened! I suggest this
case be disposed of w/out trial, since victim cannot testify…Spent 2 hours w/this victim—will
be nigh impossible to sponsor her in court. At one point, told me nothing happened, then says
little boy might have [done] it (D’s son).”
Mansfield was exonerated in 2016 because, in the wake of Michael Morton’s exoneration, the
Williamson County District Attorney’s Office made its files available to defense attorneys.
Mansfield served three months in jail under his plea bargain, and 10 years on probation—and he
was required to register as a sex offender for life. Because of that last provision, his case was still
active in 2016. Otherwise, he probably would never have been exonerated.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 136 • National Registry of Exonerations • September 1, 2020
After the exoneration, Paul Womack agreed with the state bar that his license to practice law be
suspended indefinitely, because of an unspecified disability.
241
b. Richard M. Daley – Encouraging Brutality
On January 21, 2011, former Chicago Police Commander Jon Burge was sentenced to four and a
half years in federal prison for perjury and obstruction of justice because he lied under oath
about his use of torture to extract confessions from numerous criminal suspects,
overwhelmingly Black men.
It’s hard to summarize the enormity of the harm Burge and his underlings inflicted on those
men, their families and friends, and the City of Chicago. But they didn’t do it on their own. What
follows, from an article by Natalie Y. Moore published by The Marshall Project,
242
barely
scratches the surface.
“[Burge] tortured his first documented victim in 1973. … His officers had arrested
a man named Anthony Holmes on suspicion of murder and wanted him to
identify an accomplice. When Holmes refused, the officers left him handcuffed in
an Area 2 investigation room and went to find Burge. A few minutes later, Burge
strolled into the interrogation room with a mysterious box in a brown paper bag.
The box had a hand crank on one end and two wires with alligator clamps coming
out the other end. … Burge then picked up the alligator clamps and barked,
‘Nigger, you’re going to tell me what I want to know.’ He fastened the alligator
clamps and pulled a plastic bag down over Holmes’s head, warning him not to
bite through it when the pain hit. Then he started turning the crank.
“When the first blast of electricity rolled through him, Holmes bit through the
plastic bag—a reflex reaction that opened a vent for his screams of agony. Then
everything went black, and when he woke up, Burge was putting a fresh plastic
bag over his head—the terrible panic of suffocation compounded the pain of the
electric shocks. …Burge released another blast of electricity. Again Holmes felt a
thousand needles piercing every nerve in his body, and they kept on piercing and
piercing until his brain couldn’t take it and everything went black again.
“When he woke up, Burge was laughing and preparing his infernal box—he called
it ‘the nigger box’—for another round. Holmes broke. He confessed to a murder
he didn’t commit. …
“The torture ring was great for Burge's career. Because of his high clearance rate,
he was promoted to sergeant and then lieutenant—even though he was so open
241
See Tony Plohetski, A Life of Shame: How Williamson County Prosecutors’ 23-Year-Old Notes Exposed a
Father’s Wrongful Conviction, Austin American-Statesman, November 15, 2019.
242
Natalie Y. Moore, Payback, The Marshall Project (Oct. 30, 2018).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 137 • National Registry of Exonerations • September 1, 2020
about his methods that he sometimes displayed his ‘nigger box’ on a table at the
police station.”
243
In 1983, Darrell Cannon was questioned by Burge’s crew, who wanted him to lead them to one
“A.D.,” a gang member suspected of murder. By then, Burge and the officers working under him
had expanded their range of techniques. Years later, as reported by Moore, Cannon testified that
they drove him to a remote spot on railroad tracks, and there:
“ ‘The officer with the pump shotgun played Russian roulette on me by showing
me a shotgun shell, then turning his back to me, says, “Listen, nigger” and all I
could see was his back and not the shotgun nor the shell, and I heard two clicks
of the shotgun being clicked. Then he turn to face me forcing the barrel into my
mouth saying, “Nigger! Are you going to tell us where A.D. is?” ’
“Cannon said he didn't know, which he didn't, but the officer with the shotgun—
Lt. Peter Dignan, Burge's second-in-command—shoved the shotgun into his
mouth again. The other officers shouted, ‘Blow that nigger’s head off!’ like a
cheering section. But Cannon still didn’t know where A.D. was, so Dignan played
another round of turning his back to load the shotgun and coming back around to
shove it into Cannon's mouth. …”
244
When that didn’t work,
“Dignan and the officer with the cattle prod—a sergeant named John Byrne,
threw [Cannon] in the unofficial vehicle they’d so ominously chosen to use …and
once in the car, they pulled Cannon's pants down around his ankles, armed the
cattle prod and pushed it into Cannon's crotch. When that didn’t produce the
answer they wanted… Byrne hit his crotch with another blast of electricity, and
this time he kept it up for a solid 30 seconds.
“Cannon couldn’t take any more. The torture team had broken him. He'd say
anything they wanted, he begged. He’d sign any document, admit to killing the
president of the United States, just please God make it stop.
“Three years later, Byrne was promoted to commander.”
245
243
Id.
244
Id.
245
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 138 • National Registry of Exonerations • September 1, 2020
In 1991, after these horrifying stories finally started to came out, Burge was suspended.
246
In
1993, he was fired by the police department, but kept his pension. Two of his subordinates were
suspended, but ultimately were fully reinstated with back pay.
247
In 2009, two years before Burge was convicted of perjury, the Illinois State Legislature created a
Torture Inquiry and Relief Commission,
248
which eventually concluded that 32 defendants who
had been convicted of felonies in Cook County presented sufficient evidence that their
confessions were obtained by torture to require court hearings
249
—not counting many who were
no longer alive, or did not file claims for other reasons, or had been tortured but did not confess,
or were not convicted, or whose confessions were not used in court. In 2014 and 2015—after the
United Nations Committee Against Torture condemned the City of Chicago for failing to provide
sufficient redress to torture victims
250
—the City Council issued a formal apology to Burge’s
victims; provided up to $100,000 in reparations for each victim; gave free city college tuition,
job training, and psychological services to victims and their families; and required Chicago
schools to teach the history of police torture in 8th and 10th grade history classes.
251
By 2018,
city, county, and state taxpayers had paid at least $132 million in settlements and legal fees for
claims of torture by Burge and his men.
252
Many, probably most of the men tortured under Burge were guilty of at least some of the crimes
to which they confessed. But 19 exonerations in the Registry were from convictions based on
false confessions obtained by Burge and his officers, one rape and 18 murder cases, including
four death sentences.
Burge and his men committed countless premeditated violent crimes against helpless people in
their custody. They were vicious, sadistic and racist. There is no point in exploring their motives.
Such people exist, some become police officers; some corrupt their colleagues. The glaring
question is different:
How did they get away with it for so long, with impunity?
Dozens, probably hundreds of people must have known about what they were up to, at least in
one or a few particular cases: other police officers, sheriff’s deputies at the county jail, medical
246
David Jackson, Questions About Police Torture Persist, Chicago Tribune, Nov. 11, 1991.
247
Demoted Detectives Win Reinstatement, Chicago Tribune, Jan. 28, 1994. For a detailed history of Burge’s legacy
of police torture and its impact on Chicago, see Flint Taylor, The torture Machine: Racism and Police Violence in
Chicago (2019).
248
Illinois Torture Inquiry and Relief Commission Act, 775 Ill. Comp. Stat. 40 (2009).
249
State of Illinois Torture Inquiry and Relief Commission, TIRC Decisions. The TIRC decisions by outcome are listed
here.
250
People’s Law Office, “UN Committee Against Torture Calls Out the US Government.”
251
Chicago City Council, Anti-Torture Resolution (2015).
252
Elvia Malagon, 4 Things: The Cost of Jon Burge’s Police Torture Legacy, Chicago Tribune, Sept. 21, 2018.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 139 • National Registry of Exonerations • September 1, 2020
personnel, prosecutors and their investigators, and probably some judges. Didn’t anybody alert
someone with the authority to stop this reign of torture and bring the torturers to account?
Actually, that did happen—probably many times, but at least once that is well documented.
On February 17, 1982, Dr. John Raba, Medical Director of the Cook County Jail hospital, sent a
letter to the Superintendent of the Chicago Police Department concerning Andrew Wilson, who
had been arrested by Burge and his officers for killing two police officers.
“I examined Mr. Andrew Wilson on February 15 & 16, 1982. He had multiple bruises,
swelling, and abrasions on his face and head. His right eye was battered and has a
superficial laceration. [He] had several linear blisters on his right thigh, right cheek
and anterior chest which were consistent with radiator burns. … He also stated that
electrical shocks had been administered to his gums, lips, and genitals.
“All these injuries occurred before his arrival at the Jail.
“There must be a thorough investigation of this alleged brutality.”
253
A week later, the Superintendent wrote to Richard M. Daley, State’s Attorney for Cook County
from 1980 to 1989 when he became Mayor of Chicago, transmitting Dr. Raba’s letter and adding
that “because the person in question” was a defendant in a criminal prosecution by Daley’s
office, “I will forebear from taking any steps … until I hear from you or one of your assistants.”
254
And that was where it ended.
In June 2006, then-Mayor Daley gave a sworn statement about that letter to a special
prosecutor investigating the Burge torture regime. It’s a rambling account in which Daley
basically says several times that he was not responsible for the inaction—although he admitted
that he must have seen the letter—because it should have been handled by his subordinates.
In 2008, asked if he would apologize for the torture by Burge and those working for him, Daley
said, sarcastically:
“The best way is to say, ‘Okay. I apologize to everybody [for] whatever happened to
anybody in the city of Chicago.’... So, I apologize to everybody. Whatever happened to
them in the city of Chicago in the past, I apologize. I didn’t do it, but somebody else did
it. …” Daley said, laughing. … “But I was not the mayor [at that time]. I was not the
police chief. I did not promote [Burge]….”
255
True. Daley was just the prosecutor who used confessions Burge extracted by torture instead of
prosecuting the torturers. It was an open secret in the Chicago Police Department that Daley
253
Letter from Dr. John M. Raba, Medical Director, Cermak Prison Health Services, to Richard J. Brzeczek,
Superintendent, Chicago Police Department (Feb. 17, 1982).
254
Letter from Richard J. Brzeczek, Superintendent, Chicago Police Department, to Richard M. Daley, State’s
Attorney, Cook County (Feb. 25, 1982).
255
Daley Sarcastic About Burge Torture, Chicago Sun-Times, Nov. 23, 2008.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 140 • National Registry of Exonerations • September 1, 2020
and his deputies knew of the torture and tacitly condoned it.
256
And, of course, so did the
command structure of the Chicago Police Department.
Why did Daley and Burge’s superiors in the police department tolerate if not encourage this
reign of terror? The simplest answer is probably the best: they thought the defendants were
guilty, they wanted murder convictions, and they didn’t worry about the means. Plus they
probably didn’t mind the infliction of torture on men they believed were murderers—at least not
when those men were Black.
So torture became routine. Not in all cases, or most—not even in most murder prosecutions—
but if a confession was needed to close a case and the suspect wouldn’t confess, Burge and his
men went to work. And it wasn’t just them. They may have been the most notorious and
systematic torturers; they may have led the way; but they were not alone. Once it became clear
that torture was permitted, why wouldn’t other officers join in? Not all, probably not most, but
more than a few. As we saw, there were 52 exonerations with false confessions produced by
violent interrogations in Chicago.
257
Fewer than half were conducted by officers in Burge’s unit;
13 occurred after he was fired.
c. Joyce Gilchrist – Rewarding Fraud
On May 8, 1985, a woman was raped in her apartment complex in Oklahoma City. Jeffrey
Pierce, who worked at the complex as a landscaper, was taken by police to the victim; she said he
was not the rapist. In 1986, police created a photographic lineup with a picture of Pierce wearing
a tan shirt—an element of the victim's initial description of her attacker—and the victim
identified him.
258
Pierce was arrested, convicted of rape and related crimes, and sentenced to 65
years in prison. He was exonerated by DNA testing 15 years later, in 2001. Pierce’s conviction
depended heavily on the testimony of Joyce Gilchrist, a forensic chemist at the Oklahoma City
Police Department.
Gilchrist testified that 33 scalp and pubic hair samples from the crime scene were
“microscopically consistent” with hairs taken from Mr. Pierce's body. In other words, she
testified that the hairs could have come from Mr. Pierce. Such testimony is scientifically
meaningless because there are no systematic data on the frequency of various microscopic
characteristics in human hair— even if the two hair samples do in fact share similar
characteristics. In this case, that basic factual premise was false. A re-examination by the FBI in
2001 “concluded that none of the hairs taken from [Pierce’s] body exhibited the same
microscopic characteristics as those found at the crime scene.”
259
In other words, Gilchrist
256
See generally Flint Taylor, The Torture Machine: Racism and Police Violence in Chicago (2019).
257
See supra, Section V(3)(a)(i); Klara Stephens, Misconduct and Bad Practices in False Confessions: Interrogations
in the Context of Exonerations, 11 Northeastern Univ. L. Rev. 593, 607 (2019).
258
Pierce v. Gilchrist, 359 F.3d 1279 (10th Cir. 2004).
259
Id. at 1283.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 141 • National Registry of Exonerations • September 1, 2020
concealed the fact that the hairs found at the crime scene could not have come from Pierce’s
body.
In addition, Gilchrist violated a court order to deliver the hair samples in a timely manner for
review by Pierce's own expert. And, most important, Gilchrist concealed her own findings that
Pierce could not have been the rapist because his blood contained an enzyme that was absent
from the semen found in the victim.
260
This was hardly the only case in which Gilchrist committed forensic fraud. The 2001 FBI review
“found that at least five of the cases involved contrived and erroneous statements by Ms.
Gilchrist.”
261
And in the case of Alfred Brian Mitchell, a federal judge found that Gilchrist knew
that “testing [by the FBI] revealed that [Mitchell’s] DNA was not present on the samples” of
semen taken from the victim—but lied and testified that “the DNA analysis performed by the
FBI was ‘inconclusive.’”
262
A Federal Court of Appeals concluded on review that “Ms. Gilchrist
thus provided the jury with evidence implicating [Mitchell] in the sexual assault of the victim
which she knew was rendered false and misleading by evidence [that she] withheld from the
defense.”
263
Ultimately, hundreds of Gilchrist’s cases were reviewed. Six defendants she testified against
have been exonerated, including two who were sentenced to death, and others are still disputed.
In one of those death row exonerations, Gilchrist erroneously concluded that the real killer
could not have been involved because he had the wrong blood type. And in a rape-murder in
Beatrice, Nebraska, in 1985, an initial suspect—the real killer—moved to Oklahoma City shortly
after the crime, where Gilchrest obtained a blood sample and reported the wrong blood type. As
a result, he was falsely cleared. That led to a four-year investigation that produced false
convictions of six innocent defendants, who were ultimately exonerated in 2008 and 2009 after
DNA evidence proved the crime was committed by the suspect Gilchrist had cleared in 1985.
Why did Joyce Gilchrist embark on this career of systematic fraud and at least occasional
incompetence? We can only guess at her motivation, but the effect was unambiguous. She
became a star. After Pierce was convicted in 1986, Gilchrist received an honorary citation from
the Oklahoma City police and a commendation from the district attorney, for her “skillful work
in the careful analysis of the forensic evidence.”
264
In 1990, she was promoted to supervisor,
260
Id. at 1282.
261
Id. at 1283.
262
Mitchell v. Ward, 150 F. Supp. 2d 1194, 1226 (W.D. Okla.1999).
263
Mitchell v. Gibson, 262 F.3d 1036, 1064 (10th Cir. 2001). Mitchell was sentenced to death for murder; the alleged
sexual assault on the victim was presented to the jury as an aggravating factor to consider in deciding whether to
sentence him to death. Gilchrist’s misrepresentations on that issue led the court to vacate Mitchell’s death sentence,
but not his murder conviction.
264
David Kohn, Under the Microscope: Forensic Scientist Accused of Mishandling Cases, CBS: 60 Minutes (May 8,
2001).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 142 • National Registry of Exonerations • September 1, 2020
years earlier than usual. In 1995, she was named Oklahoma City Police Department Civilian
Police Employee of the Year.
265
Gilchrist’s work made her popular with police officers. She became known to them as Black
Magic “because she was able to get results that no other chemist could. When … homicide
detectives gave Gilchrist hair samples from a suspect, they would often let her know that this
was the person that they wanted to arrest.”
266
Some later claimed that they “didn’t believe
Gilchrist was doing proper lab work, because her results were ‘too good’”–but that didn’t stop
them from using those results in court.
267
Gilchrist was also a forceful and effective witness, and prosecutors came to rely on her to win
difficult jury trials. In some cases, they committed misconduct themselves in presenting her
evidence. Alfred Brian Mitchell’s trial was a striking case, as a federal court described in 2001:
“Compounding [Gilchrist’s] improper conduct was that of the prosecutor, whom the district
court found had ‘labored extensively at trial to obscure the true DNA test results [which cleared
Mitchell of rape] and to highlight Gilchrist's [misleading] test results….”
268
Gilchrist became a particular favorite of Oklahoma County District Attorney Bob Macy. In
February 2001, Gilchrist, who was being investigated by the police department, was placed on
administrative leave. In June 2001 Macy resigned unexpectedly, saying he wanted to spend
more time with his family. Three months later, Gilchrist was fired.
269
There had been warnings, for years. Defense attorneys had complained about Gilchrist. John
Wilson, chief chemist at the Kansas City police crime laboratory, testified for the defense in
several cases that Gilchrist worked on; in 1987, he filed a complaint against her with the
Southwestern Association of Forensic Scientists.
270
In 2001, after his claims were confirmed by
the FBI, Wilson said:
“I think you have to look at the prosecutor's office … they have to understand what's been
going on. They have to have seen all the flags that have been waved. The judges are no
different. … It's not just the police, it's not just the prosecutors, it's everyone in the entire
system.”
271
But as long as the engine was humming along, nobody wanted to look under the hood.
265
Belinda Luscombe, When the Evidence Lies, Time (May 13, 2001).
266
Mark Fuhrman, Death and Justice: An Expose of Oklahoma’s Death Row Machine 91 (2003).
267
Id. at 71.
268
Mitchell, 262 F.3d at 1064.
269
Lois Romano, Police Chemist’s Missteps Cause Okla. Scandal, Wash. Post (Nov. 26, 2001). See generally Daniel
LaChance, Executing Freedom: The Cultural Life of Capital Punishment in the United States (2016), ch. 6.
270
Paul C. Giannelli, Scientific Fraud, 46 Criminal Law Bulletin, 1313, 1325 (2010).
271
See David Kohn, Under the Microscope: Forensic Scientist Accused of Mishandling Cases, CBS: 60 Minutes (May
8, 2001). See also Scott Cooper, Former DA Bob Macy, ex-forensic chemist Joyce Gilchrist settle case, The
Oklahoma Gazette, June 18, 2009.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 143 • National Registry of Exonerations • September 1, 2020
d. Officers Iannotto, Palmer, Pecorale, Martin, Visconti and Bishop, and Detective Massanova
– Closing Cases
Around 2 a.m. on the morning of November 18, 1990, a young man in a group of several bumped
into another youth in a different group on a street in Greenwich Village in New York City. One of
the two who collided was wearing a reversible “orange aviator-style jacket.” He produced a .32
caliber pistol and fired repeatedly at the walls of nearby buildings. He then handed the gun to a
companion, who fired at the opposing group, killing one man and seriously injuring another.
The shooters ran away, followed by one of those they had shot at—Jose Fontanez—who flagged
down a police car.
Officers Glenn Iannotto and Drew Palmer drove to a nearby train station to intercept the
suspects if they tried to escape by train. They were joined there by officers Dean Pecorale and
Jill Martin, who spotted Luis Rojas on a train about to leave the station, wearing a jacket similar
to the one worn by the man who produced the gun, but with the orange shell on the inside and
the reversible maroon side was on the outside. Rojas was taken off the train, instructed to turn
his jacket inside out so the orange shell was on the outside, and handcuffed. In the meantime,
Officers Visconti and Bishop brought Fontanez to the station, where he identified Rojas as the
man who fired the first shots.
Shortly after that, officers Pecorale and Martin took Rojas to the scene of the shooting, and
officers Bishop and Visconti drove Fontanez there as well. An appellate court, reviewing the case
several years later, described what followed:
“All but three of the youths who were shot at … testified at trial that they were at the
crime scene when defendant arrived, handcuffed, in the back of the patrol car. Defendant
testified… that Mr. Fontanez, in front of other witnesses, pointed at him through the
cruiser’s window saying that he was the one because ‘[h]e has the colors on.’ At some
point, [one of the men who had been shot at—all of whom had been drinking] began
punching the car and was restrained by the officers present at the scene.”
272
The following morning, Detective Daniel Massanova conducted a lineup that was seriously
compromised. It consisted of Rojas himself and several “police cadet fillers who all had short
hair … were neat, fresh and clean-shaven with crew cuts, as they were on their way to the
[Police] Academy, while the suspect had been up all night before the 10:20 A.M. lineup, so the
cadets were easily identified.”
273
Needless to say, Rojas was once again identified repeatedly as the man who started the shooting.
On February 26, 1992, he was convicted of second degree murder and related offenses, and was
sentenced to 15 years to life in prison. Six years later, Rojas was exonerated by overwhelming
evidence that he had been misidentified.
272
People v. Rojas, 213 A.D.2d 56, 630 N.Y.S.2d 28 (N.Y. App. Div. 1995); the Rojas opinion included a detailed
description of this case.
273
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 144 • National Registry of Exonerations • September 1, 2020
In a 1995 opinion reversing Rojas’s conviction, the Appellate Division of the New York Supreme
Court wrote:
“Once the identification and arrest had been made, there was no need for an on-the-
scene investigatory showup and defendant should have been kept segregated from the
potential lineup witnesses. Instead, as demonstrated by the officers’ radio transmissions,
defendant was intentionally… returned for additional showups.
“This on-the-scene showup was further tainted by the fact that Fontanez was also
returned to the crime scene and …[said] in front of other witnesses, while pointing to
defendant that he was the one because ‘[h]e has the colors on’….
“The lineup itself approaches the bizarre. [T]he fillers were police cadets, clean shaven,
crew-cut and neat on the way to the Academy after a good night’s sleep [while the
suspect] had ‘normal’ length hair, had been up all night and [was], no doubt, disheveled,
probably looking the worse for wear.”
274
In short, as the court concluded:
“The entire identification procedure, or rather the series of identifications were both
improper and prejudicial.”
275
This too was routine misconduct, if not in the same range as Ken Anderson’s or Richard Daley’s.
The police officers responded quickly to the scene of a murder, located a plausible suspect—and
then went off the rails. From the look of it, they were convinced that they had the shooter and
knew they needed eyewitness identifications, so they went about getting those identifications in
the easiest manner possible without considering the possibility that they were orchestrating the
conviction of an innocent man.
2. CAN WE REDUCE OFFICIAL MISCONDUCT IN CRIMINAL CASES?
Yes, we can reduce official misconduct in criminal cases. It has happened. There have been
dramatic reductions in some types of misconduct among exonerations involving convictions
since 2003, and even greater improvements in the middle decades of the twentieth century. But
misconduct that leads to false convictions could also get worse; we’ve seen that too, in some
settings.
There is no panacea for official misconduct in criminal prosecutions because it’s not one thing
but several different types of behavior, by different actors in different settings. A single remedy
is no more likely than a general cure for “disease.”
In this section, we discuss several categories of reforms, some of which may have already
produced favorable changes. We start with the most specific—rules that dictate the conduct of
274
Id.
275
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 145 • National Registry of Exonerations • September 1, 2020
law enforcement officials—and proceed to the most general, changes in our legal and political
culture.
a. Categories of Reforms
i. Rules
(a) Procedural rules
Michael Morton was convicted of a murder he did not commit because his prosecutor, Ken
Anderson, deliberately concealed powerful exculpatory evidence. In 2013, in response to his
exoneration, the Texas Legislature passed and the governor signed the Michael Morton Act,
which created “open file” discovery in criminal cases in Texas.
276
It requires prosecutors to
provide defendants with copies of, or access to:
“[A]ny offense reports, any designated documents, papers, written or recorded
statements of the defendant or a witness, including witness statements of law
enforcement … or any designated books, accounts, letters, photographs, or objects or
other tangible things … that constitute or contain evidence material to any matter
involved in the action and that are in the possession, custody, or control of the state….”
The Act also addresses the specific duty that Anderson violated. It provides that “the state shall
disclose to the defendant any exculpatory, impeachment, or mitigating document, item, or
information in the possession, custody, or control of the state that tends to negate the guilt of
the defendant or would tend to reduce the punishment for the offense charged.”
277
That
provision was not new; it reiterated existing constitutional and ethical requirements.
Anderson, of course, broke existing rules by concealing exculpatory evidence. Not only did he
violate Morton’s right to due process of law, he also lied to the judge who presided over the case
when he falsely claimed to have provided him with all potentially exculpatory evidence.
Would this sort of general procedural rule have made a difference to Morton’s prosecution in
1987? Maybe. If you’re required to disclose everything, it may be harder to hide the evidence
you’d hate to face in court. Where open discovery becomes the rule, defense attorneys and
judges may become better at spotting gaps and omissions, and prosecutors may become more
wary of holding back critical information.
The key, of course, is enforcement. We think it’s likely that Ken Anderson ignored fundamental
rules in Michael Morton’s case because they were unenforced and ignored by prosecutors in
Williamson County generally. That can happen to any rule, from those that require open
discovery in criminal cases to those that prohibit speeding, tax evasion, or the murder of
civilians in war zones. Every type of misconduct we discuss violates existing legal and ethical
rules. Why not just enforce those that we already have?
276
Texas Code of Criminal Procedure, Art. 39.14.
277
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 146 • National Registry of Exonerations • September 1, 2020
Strict enforcement of existing rules may work as well as anything—if it happens. New rules can
make enforcement more likely by mobilizing support and drawing attention to the problem,
especially if they specify what should be done in concrete terms. In November 2017, for
example, the Chief Judge of the State of New York “announced the adoption of new rules that
will require judges presiding over criminal trials to issue an order notifying and reminding
prosecutors” in detail of their obligation to disclose exculpatory evidence.
278
The announcement
specifically notes that “the new order… does not in any way change existing law, [but] provides a
mechanism by which to educate inexperienced prosecutors and defense attorneys—and remind
experienced ones—about their constitutional and ethical duties.”
279
This new rule could make a
major difference, if implemented and taken seriously, or it could become an empty formality.
Prosecutors and police can only conceal evidence they have obtained in the process of
investigation. The “procedural” rules of discovery that bear on this sort of misconduct govern
the disclosure and use of information that is already in the hands of state officers. The same is
true for prosecutorial misconduct at trial, which is governed by elaborate procedural rules of
evidence and trial practice.
On the other hand, there are several contexts in which specific rules on how to obtain and
preserve evidence could improve the quality of information in criminal trials generally, and
reduce misconduct along the way.
(b) Evidence gathering rules
Eyewitness identification. In People v. Rojas,
280
the Appellate Division of the New York
Supreme Court had many complaints about the lineup that Detective Massanova used to obtain
identifications of the defendant, including this: “Since no lineup photos were taken … appellate
review, to a large extent, [is] speculation.”
281
It would have been better if Massanova had
photographed the lineup, and better yet if he had recorded the entire process on video.
Failing to photograph or record a lineup is common, and it’s not misconduct. But if there had
been a rule that required the police to photograph the lineup (there was no such requirement), it
might have prevented the misconduct that did occur: subjecting Rojas to a patently suggestive
procedure that was guaranteed to produce identifications whether he was guilty or not.
Photographing lineups reduces the incentive to deliberately cook an identification because it
makes (at least some forms of) suggestiveness visible for all to see later on. It may also reduce
the likelihood of unintentional suggestiveness by focusing attention on the procedure and
278
Press Release, N.Y. State Unified Court System, Chief Judge DiFiore Announces Implementation of New
Measure Aimed at Enhancing the Delivery of Justice in Criminal Cases (Nov. 8, 2017). The new rules specifically say
that the “order [from the trial court] shall not contain any reference to materiality,” and include provisions that require
trial courts to remind defense attorneys in detail of their duty to provide criminal defendants effective legal
representation.
279
Id.
280
People v. Rojas, 213 A.D.2d 56, 630 N.Y.S.2d 28 (N.Y. App. Div. 1995) (N.Y. App. Div. 1995).
281
Id., 213 A.D.2d at 70.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 147 • National Registry of Exonerations • September 1, 2020
reminding the officer who administers a lineup that others will review it. And, of course, it
makes it easier for judges and juries to assess the value of an identification after the fact.
Recording the identification process is one of several rules that have been proposed, and
sometimes adopted, for the purpose of increasing the accuracy of eyewitness identifications.
Others include: having the lineup administered by a person who does not know the identity of
the suspect, and therefore cannot communicate it to the witnesses, purposefully or
inadvertently; selecting “fillers” for the lineup who resemble the witnesses description of the
criminal; telling each witness that the criminal may or may not be in the lineup; recording the
level of confidence of any witness makes an identification; and never permitting multiple
witnesses to view a lineup together.
282
These procedures are designed to make all misidentifications less likely—including those that
might be generated by “tainted identification procedures” in which “police intentionally
structure the identification to induce the witness to identify the exonerees.”
283
They keep police
from deliberately stacking lineups by requiring them to run lineups that can’t be stacked. At
least that’s the theory.
Similar rules have been proposed, and sometimes enacted, to govern other aspects of criminal
investigations. The most important are those that require that interrogations of criminal
suspects be recorded, preferably on video.
Interrogations. As we have noted, much of the coercive behavior that produces false confessions
is not considered misconduct.
284
Officers are allowed to question suspects in isolation and at
great length, to tell them that their guilt has been determined beyond doubt, and to lie to them
about physical evidence (“we have your fingerprints”) or other witnesses (“your buddy already
told us you did it”). Sometimes, this behavior spills over into misconduct; often, it does not.
If a defendant’s guilt is in dispute despite his confession, it’s essential to know what actually
happened in the interrogation—in some cases, to decide whether a confession by the defendant
must be excluded from evidence because it was coerced; in others, to evaluate the credibility of a
confession that has been admitted.
But in many cases—until recently, almost all—we don’t know what happened in the
interrogations because they were not recorded in any form. The Supreme Court described the
issue 75 years ago, when E. E. Ashcroft claimed that he was forced to confess during a long
interrogation by the Memphis police:
“As to what happened … during this thirty-six hour secret examination, the testimony
follows the usual pattern, and is in hopeless conflict. Ashcraft swears that the first thing
said to him when he was taken into custody was, ‘Why in hell did you kill your wife?’;
that, during the course of the examination, he was threatened and abused in various
282
See, e.g., CA Penal Code § 859.7 (2018). See generally infra Section XII.2.b.ii(b).
283
See supra Section VI.3.
284
Supra Section V.4.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 148 • National Registry of Exonerations • September 1, 2020
ways, and that, as the hours passed, his eyes became blinded by a powerful electric light,
his body became weary, and the strain on his nerves became unbearable. The officers, on
the other hand, swear that, throughout the questioning, they were kind and
considerate.”
285
In the last 30 years, exonerations by DNA and other means have proven beyond doubt that false
confessions occur on a regular basis, especially in murder cases. In response, there has been a
growing movement to require that interrogations be recorded, in all felony cases or at least in
homicides. In 2000, recording interrogations was required in two states, and explicitly frowned
upon by federal law enforcement agencies. It is now required, in some form, by 26 states, the
District of Columbia and all agencies of the United States Department of Justice.
286
Needless to say, recorded interrogations are far less likely than unrecorded ones to include
misconduct of any sort—and will probably never include torture. That connection is not
incidental. One of the turning points in the spread of this reform occurred in 2003, when Illinois
passed a law that requires police to record interrogations in murder cases—a law that was an
explicit response to the horrific history of torture by Jon Burge and his lackeys.
287
Witness interviews. Eyewitness identification and interrogation are two major contexts in which
police deal with potential witnesses. A true confession, like an accurate identification, is factual
evidence from a “witness” to the crime. Rules that require recording these procedures make it
easier to tell if a suspect confessed to a crime he did not commit—or a witness identified a
suspect she did not see—and also discourage misconduct that may produce false confessions and
erroneous identifications. Conceptually, however, these are specific applications of a more
general rule.
We could require police to record all conversations with people who are interviewed during the
investigation to a crime, and to make video recordings whenever the information they seek
includes visual content. If such a rule were enacted and observed, it would produce accurate
information about witness statements, and along the way, prevent the worst sorts of witness
tampering: the use of threats, lies, bribes or promises of favors to induce witnesses to give false
evidence.
No such rule exists anywhere in the United States, but it’s technologically feasible. Officers, for
example, could be required to wear bodycams and to turn them on when talking to witnesses.
This is, in fact, one aspect of several general police bodycam rules that have been proposed to
285
Ashcraft v. Tennessee, 322 U.S. 143 (1944) (footnotes omitted).
286
See infra XII.2.b.ii(a). The 26 states include 24 that added statutes or court rules that required recording all or
some types of interrogations, and two where all police agencies in the state did so on their own.
287
3 705 ILCS §405/5-401.5 and 725 ILCS §5/103-2.1, relating to investigations of first degree murder suspects (the
statute took statewide effect in July 2005). See Steven Thomma, Police interrogation law showed Obama's skill in
Illinois senate, McClatchy Newspapers (Impact 2020), March 27, 2008, for a description of the crucial role Barak
Obama—then an Illinois state senator—played in getting that legislation passed.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 149 • National Registry of Exonerations • September 1, 2020
address police violence, especially as a part of the Black Lives Matter movement.
288
Of course,
officers who are about to commit a crime by threatening witnesses in order to procure perjury
will not do it on camera, rule or no rule. But if there were such a rule, they might do it less often.
Physical evidence. Most police forces have no general rules that govern the collection and
preservation of physical evidence, from fingerprints and biological trace evidence to weapons,
clothing, vehicles, stolen objects, and so forth. Such rules might both improve the quality of
evidence that’s available at trial and make it more difficult for officials who are so inclined to
conceal or destroy exculpatory physical evidence.
The rules we describe in this section are not specifically aimed at misconduct. Their main
purpose is to improve criminal investigations by treating all evidence—objects and the memories
of witnesses alike—as fragile material that must be carefully collected and preserved: Record
what witnesses and suspects say, and the context in which they speak; avoid biasing their
memories or leading them to say things they don’t believe; find all available physical evidence
and store it intact. Along the way, following these rules is likely to reduce misconduct, perhaps
dramatically.
What possible objection could there be to these reforms? Some officers, of course, would like to
continue to influence lineups, coerce confessions, or ignore physical evidence that suggests that
a suspect might be innocent. That’s not an argument anybody actually makes, but there are
slightly more respectable relatives in the vicinity. Some police officers, for example, object to
recording interrogations on the ground that jurors will not “understand” why ugly interrogation
techniques are “necessary”—that it’s better for them to remain ignorant of how their evidentiary
sausages are made. It’s a hard position to defend. If you can’t explain to a jury why your method
of interrogating suspects is OK, you probably shouldn’t be doing it—and you certainly shouldn’t
be allowed to hide the process from those who decide those suspects’ fates.
There is, however, a different objection to detailed rules on obtaining and preserving evidence.
Implementing them takes time, money, training and supervision—in short, resources. Without
adequate resources, rules like these—if adopted—will not be followed or enforced.
ii. Resources
When we described the 1992 murder conviction of Luis Rojas, we said that he was exonerated
six years later “by overwhelming evidence that he had been misidentified.”
289
But we didn’t say
what that evidence was (other than the extreme suggestiveness of the eyewitness identifications
the police obtained). The new evidence of Rojas’s innocence that was presented in 1998
included:
288
See, e.g., ACLU, A Model Act For Regulating the Use of Wearable Body Cameras by Law Enforcement, Version
2.1 (June 2018); Jay Stanley, Police Body-Mounted Cameras: With Right Policies in Place, A Win for All, ACLU
(March 2015). See generally Shelley S. Hyland, Body-Worn Cameras in Law Enforcement Agencies, U.S.
Department of Justice, Bureau of Justice Statistics (2016).
289
Supra Section XII.1.d.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 150 • National Registry of Exonerations • September 1, 2020
(i) An alibi from a police sergeant for the New York Port Authority, who saw Rojas at the
train station 10 minutes before the city police officers arrived—which meant that he
could not have participated in the shooting.
(ii) Testimony from an uninvolved eyewitness to the shooting—a security guard at a
restaurant across the street—who gave the police a description of the man they were
looking for, in an aviator jacket, that did not resemble Rojas. Later, when that witness
saw Rojas back at the scene of the crime, handcuffed in a squad car, he told police that
Rojas “was definitely the wrong man.”
290
That security guard was not asked to view the
lineup the next morning at which Rojas was identified by several companions of the
victims who had been shot.
In other words, if the police had done a decent investigation on the night of the shooting, or in
the days that followed, Rojas would never have been charged with any crime, let alone convicted
of murder and imprisoned for six years. Why didn’t they do a better job?
It’s tempting to blame laziness or incompetence, and they may have played a role. But the work
load of the New York City police department was probably a bigger factor. There were more than
2,000 homicides in New York in 1992, compared to 287 in 2018. In that unruly context, the
detective in charge of the investigation probably thought the Rojas case was easy: the police
caught the gunman as he was trying to escape (or so they thought), and he was immediately
identified by several witnesses to the shooting. A lineup was needed to nail the lid shut—and
then on to the endless line of other murder cases.
When police officers fail to conduct a careful investigation in a criminal case, the consequences
may be catastrophic. In this case, they didn’t interview a readily available alibi witness, and
ignored a credible (and sober) eyewitness who gave them a description of the criminal that
didn’t fit Rojas and then told them that Rojas “definitely” didn’t do it. As a result, an innocent
man was convicted and imprisoned, and the actual criminals were never apprehended. In other
cases, they might ignore fingerprints that are visible at the scene of a killing, or blood, or a gun—
with equally disastrous results.
All the same, failure to conduct a good, or even a minimally competent, criminal investigation is
not in itself “official misconduct” as we, and as courts, use the term. Misbehavior that harms
another person comes in two basic varieties: abuse and neglect. “Misconduct” by law
enforcement officials in a criminal case is always abuse—some form of affirmative action that
breaks the rules and deprives the defendant of a legal right.
Criminal suspects and defendants have legal rights to be free from physical abuse, obstruction of
justice, illegal arrests and searches, official perjury, and so on. But they have no affirmative right
to high quality investigations of their cases, or indeed, to any investigation at all. Concealing
exculpatory evidence violates their rights; in rare cases, so does a failure to collect or preserve
evidence, but only if it is done in bad faith, which usually means that an officer deliberately
290
People v. Rojas, 213 A.D.2d 56 at 68, 630 N.Y.S.2d 28 (N.Y. App. Div. 1995).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 151 • National Registry of Exonerations • September 1, 2020
decided to ignore evidence that might undermine the case against a criminal defendant.
291
Neglecting to do an adequate job may be just as harmful, or worse, but it is not “misconduct.”
Lack of resources, however, does intersect with misconduct, in two ways:
First, the absence of a competent investigation, whatever its cause, leaves a suspect vulnerable
to misconduct. That’s what happened to Luis Rojas. If Detective Massanova or other officers had
interviewed the Port Authority police sergeant who would have provided Rojas with an alibi,
their misconduct in obtaining obviously suggestive identifications probably would not have
mattered. The alibi would likely have led to further investigation, after which Rojas would have
been released and the tainted identifications would not have contributed to a false conviction.
That happens regularly. As best we can tell from limited data, most misidentifications of
innocent suspects
292
—and most false confessions
293
—do not produce wrongful convictions,
usually because other evidence that proves the suspects’ innocence emerges before the case gets
to trial. In Rojas’s case, the other evidence was never sought, seen or heard.
Second, lack of resources may tempt officers to break the rules because they don’t have the time
to conduct proper investigations. That may have happened to Luis Rojas. The officers who
arrested him might have intentionally set up a hopelessly tainted identification process in order
to get identifications quickly and easily, and wind up the investigation within hours of the crime.
Sometimes, police officers who are overwhelmed by murder cases take a different impermissible
short cut. They detain several young men or boys they find in the vicinity of a killing, interrogate
them until one of them confesses or names another as the killer—and then close the case.
That seems to be what happened to Kendrick Scott and Justly Johnson, who were arrested and
charged with murder within hours after Lisa Kindred was shot and killed in Detroit on May 9,
1999. Antonio Burnette and Raymond Jackson had been arrested hours earlier as part of the
same investigation, and they told the police—and later testified in court—that Scott and Johnson
told them that they had committed the murder.
Scott and Johnson were convicted of murder in 2000. They were exonerated in 2018 after the
victim’s son—who witnessed the killing at close range when he was eight years old, but was
never questioned by the police, the prosecution or the defendants’ lawyers—testified that he was
certain that neither Scott nor Johnson was the gunman. In addition, a cousin of Jackson’s
testified that before he died in 2008, Jackson told her that he had lied because he was afraid of
the police and the prosecution; and Burnette testified that the police “whooped” him during his
291
See Arizona v. Youngblood, 488 U.S. 51, 58 (1988) (“We therefore hold that unless a criminal defendant can show
bad faith on the part of the police, failure to preserve potentially useful evidence does not constitute a denial of due
process of law.”).
292
Samuel R. Gross, "Loss of Innocence: Eyewitness Identification and Proof of Guilt." J. Legal Stud. 16 (1987): 395-
453.
293
Steven A. Drizin & Richard A. Leo, The Problem of False Confessions in the Post-DNA World, 82 N.C. L. Rev.
891, 951 (2004).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 152 • National Registry of Exonerations • September 1, 2020
interrogation, that he believed they considered Johnson a suspect, and that he was afraid he
would be charged with the murder himself if he did not say what the police wanted to hear.
The resource we’ve focused on so far is time, which is inversely proportional to the number of
cases each officer must handle. Resources may also be structural. It might have helped in the
Rojas case if the police department had explicit protocols that required detectives, at least in
homicide cases, to interview all possible alibi witnesses, construct lineups with foils who
resemble the description of the criminal, and present lineups to all known eyewitnesses, not just
those who have already identified their suspect. Of course, it takes time and money to devise and
enforce such protocols, train the officers involved, and provide the support they need. But
funding alone is not enough. It also takes expertise, commitment, and leadership.
One reason for the comparatively low rate of police misconduct among federal exonerations may
be that most federal law enforcement agencies have considerably greater financial and
organizational resources than most state and local police forces: fewer cases per officer, better
training and equipment, more detailed record keeping and closer supervision. If true, this
means that federal officers have fewer incentives to commit misconduct because they have the
wherewithal to work cases properly, more opportunities to correct initial errors and avoid
erroneous convictions, and more reason to fear that misconduct will be detected.
One last point: Defense resources are as important as those that are available to police and
prosecutors.
Luis Rojas’s conviction was not reversed on appeal because of misconduct of the police, but
because the appellate court found misconduct by his defense lawyer:
“Defendant's trial counsel, Mr. Fronefield, not only made no effort to investigate the
foregoing evidence [of Rojas’s innocence] or defendant's alibi, but further committed
numerous errors during the course of the trial which tended to implicate defendant
rather than to prove his innocence. …”
294
“Mr. Fronefield's performance… deprived defendant of the effective assistance of counsel
requiring reversal of the conviction.”
295
It’s hard to fathom the magnitude of Fronefield’s incompetence in representing Rojas. Not only
did he fail to interview witnesses who would have cleared his client, he went out of his way to
discredit Rojas’s own defense. At trial, Rojas testified, truthfully, that he was not at the scene of
the shooting. That did not stop his defense attorney from arguing to the jury that Rojas actually
was there, and that he had a gun. As the appellate court noted, “[it] almost appears, at times,
that Mr. Fronefield acted as a second prosecutor.”
296
294
People v. Rojas, 213 A.D.2d 56 at 66, 630 N.Y.S.2d 28 (N.Y. App. Div. 1995).
295
Id. at 69-70.
296
Id. at 69.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 153 • National Registry of Exonerations • September 1, 2020
Ineffective assistance of defense counsel is misconduct. In most cases, the defendant is deprived
of his constitutional right to effective representation by neglect rather than abuse, usually by the
lawyer’s failure to investigate and prepare for trial. Whatever its form, however, ineffective
assistance of counsel is not official misconduct. It’s not covered by this Report—although, as
we’ve said, it may do as much to generate false convictions as official misconduct, if not more.
297
There’s no justification for the abysmal representation Mr. Rojas received, regardless of
circumstances. Across cases, however, the level of funding for indigent criminal defense has a
huge impact on the quality of work defense attorneys do.
Attorney Fronefield tried to persuade Luis Rojas to plead guilty to whatever charges and
sentence the prosecution offered, in a murder case that Fronefield had not bothered to
investigate. Rojas refused, probably because he would have had to spend years if not decades in
prison if he had pled guilty. But overworked and underpaid public defenders and private defense
attorneys do persuade defendants to plead guilty every business day of the year, by the hundreds
and thousands, in cases the attorneys know very little about. Usually the defendants are not
charged with killing anybody, and the penalties are much lower than what Rojas faced. Some of
those defendants are victims of official misconduct, and some are innocent. We rarely hear
about either group.
On the other hand, effective defense attorneys reduce both the frequency and the impact of
misconduct by government officials. They may deter misconduct. Police officers, prosecutors
and forensic analysts may be less likely to break rules if they believe that hard-working defense
lawyers are likely catch them. And good defense work can certainly prevent misconduct from
producing convictions of innocent defendants. That’s what should have happened to Luis Rojas,
if he had not had the misfortune to be assigned a defense lawyer who “acted as a second
prosecutor,” and a lazy one at that.
In general, a well-staffed and well-funded criminal justice system—from police investigators and
forensic analysists, to prosecutors and defense attorneys, to trial judges—is likely to see less
misconduct, and the misconduct that does occur is more likely to be caught somewhere along
the way before the worst happens. It can take the failure of a village to convict an innocent
person.
iii. Accountability
What about punishing those who commit misconduct? Will that keep others from doing it in the
future?
We know that some sort of discipline was imposed on at least one government official in about a
sixth of exonerations with known official misconduct (219/1,295). Did holding officials
accountable for misconduct in those cases have any impact on misconduct by other officials?
It seems unlikely the disciplinary actions we know about had any major impact on misconduct
in later cases. Discipline is uncommon—in 83% of cases with misconduct no discipline was
297
See supra Section II.2.b.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 154 • National Registry of Exonerations • September 1, 2020
imposed—and very slow: the average time from conviction to exoneration in a case with official
misconduct is about 11 years. And there are conspicuous counter examples.
In 1993, when Commander Jon Burge was finally fired by the Chicago Police Department, the
most prolific known torturer on the Chicago police force who was not associated with Burge was
just getting started. Detective Reynaldo Guevara used beating, torture and threats to coerce false
confessions and to force witnesses to lie and implicate innocent defendants. He was responsible
for the false convictions in 14 murder exonerations; all but two of them occurred after Burge was
fired. And Chicago Police Detective Kriston Kato was responsible for abusing numerous
suspects, including two defendants who were beaten into false confessions, one for murder and
one for sexual assault, and later exonerated. Both of those exonerees were wrongly convicted
after Burge was fired.
On the other hand, disciplining brazen, active offenders is an essential component of any
attempt to stop misconduct. The impunity that Burge and his cronies enjoyed for decades
helped generate and preserve the culture of violence in interrogations in Chicago. Punishing
officers who are widely known to torture suspects may not be enough to stop it from happening,
but failing to do so will doom other efforts, if there are any. In fact, as we’ve seen, eventually—
years after Burge was fired—the rates of abusive interrogations and false confessions in Chicago
did fall to a fraction of what they were in his heyday.
Discipline comes in many flavors. The most severe is criminal conviction, which will prevent the
official involved from committing any additional misconduct because he will be fired and
probably imprisoned. But will it deter others from doing the same? By the time an officer is
convicted of criminal misconduct, other likeminded officials will have been getting away with
similar misconduct for years. Perhaps they will change their ways and hope they don’t get
prosecuted for past misdeeds; perhaps others will be deterred from taking up abuse and
deception; but at best this is a slow and inefficient way to promote law abiding behavior.
Immediate and mild sanctions for low level infractions by the agencies for which the officials
work are likely to do more to shape behavior than swinging an axe when an egregious act is
discovered years later. Reprimands from supervisors for failing to follow departmental protocols
on eyewitness lineups—or for failing to provide exculpatory impeachment evidence to the
defense—are cheap and quick and may change the behavior of the officials involved
immediately. Any action in response to an exoneration means, on average, a 10-year delay or
longer, by which time the officials involved may have done the same thing many more times, or
retired, or both. If the sanctions are severe, the process of imposing them may take years more.
In other words, the most effective sanctions are aspects of ongoing supervision of the officials
who investigate and prosecute criminal cases. That’s the stage at which misconduct is easiest to
detect and prevent. That type of supervision is best done when these investigations and
prosecutions are subject to explicit rules, and when the agencies involved have sufficient
resources to conduct careful supervision—two categories of reform we have already discussed.
Of course, even with adequate resources, many government agencies are reluctant to discipline
or even investigate their employees. It can damage morale, disrupt relations with other agencies
and make the agency and its leaders look bad. The officials under scrutiny may be well-liked and
well-connected. Their superiors may actually support their misconduct, and they may have other
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 155 • National Registry of Exonerations • September 1, 2020
powerful allies—especially if they are members of police unions. Effective oversight depends on
the priorities and the power of those who lead the agencies involved—an issue we address next.
Discipline can also be imposed by professional organizations. As we’ve described, there is a
system in most states for decertifying police officers; that means that they will lose their jobs,
but might get hired by other police forces because decertification records are kept secret.
298
Prosecutors, of course, are lawyers and subject to regulation by state bar associations and state
supreme courts. Courts and bar associations don’t have the same specific incentives to ignore
misbehavior as the agencies in which it takes place, but they may be deterred by a general
reluctance to publicly condemn and punish practitioners in their own profession, or they may
lack the resources to investigate and sanction more than a token number of offenders, or they
may just not care. For whatever reasons, they rarely discipline prosecutors, even when trial or
appellate judges cite them for misconduct in cases in their courts.
299
b. Local leadership and local culture
The most significant features of the American system of criminal justice—if you can call it that—
are fragmentation and local control.
We have 50 states, plus the District of Columbia, the Commonwealth of Puerto Rico, several
territories and the federal government, each with its own set of penal and procedural laws and
its own court system. More than 90% of criminal cases are filed in state courts, but in almost all
states those courts are county rather than state institutions, and there are 3,142 counties (or
equivalent geographical units) in the 50 states. As of 2007, there were 2,330 independent
prosecutorial offices that handled state law felony cases, mostly one per county.
300
Policing is
even more fragmented. There are about 18,000 separate police forces in the United States,
301
including more than 12,000 local police departments.
302
With rare exceptions, these thousands of agencies are independent of each other and subject to
local political control. Most state court judges and prosecutors are directly elected in county
elections. Most sheriffs and some local police chiefs are also directly elected; the rest are
appointed by elected county and municipal officials.
The institutions of national government—Congress, the Supreme Court, the President, the
Department of Justice—play important roles in the administration of criminal justice across the
298
See supra Section X.2.b.ii.
299
See Kathleen M. Ridolfi & Maurice Possley, Preventable Error: A Report On Prosecutorial Misconduct In
California, 1997–2009, A Veritas Initiative Report (2010). See generally supra Section II.1.
300
Steven W. Perry and Duren Banks, “Prosecutors in State Courts, 2007 – Statistical Tables,” Bureau of Justice
Statistics (Dec. 2011).
301
Duren Banks, et al., “National Sources of Law Enforcement Employment Data,” Bureau of Justice Statistics (Rev.
Oct. 4, 2016).
302
Brian A. Reaves, “Local Police Departments, 2013: Personnel, Policies, and Practices,” Bureau of Justice
Statistics (May 2015).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 156 • National Registry of Exonerations • September 1, 2020
country. State governments—legislatures, governors, attorneys general, state appellate courts—
are more important within their jurisdictions. But local crime labs, police departments and
prosecutorial offices investigate and prosecute the overwhelming majority of crimes, and for the
most part set the working policies for doing so. Any significant reforms depend on the
leadership of those who administer these local agencies.
Those leaders, in turn, are constrained by the institutions they run. A practice that’s been
followed for years may be embedded in the institutional culture, and, as management
professionals like to say, “Culture eats strategy for breakfast.”
303
That doesn’t mean reform is
impossible—we’ve seen it happen—but it can be difficult. The organizational culture itself may
have to be changed, which is not a simple process.
i. Crime Labs
Independent crime labs in general. Most crime labs in the United States are run by police
departments.
304
In many, the forensic analysts are sworn police officers, sheriffs’ deputies or
state troopers; in other labs, they are civilian employees of the police agencies. This creates an
obvious conflict: forensic scientists should prize accuracy and clarity above all else. However, as
a British court explained in 1993, “forensic scientists employed by the government may come to
see their function as helping the police,”
305
rather than providing accurate and complete
scientific information.
The temptation to lie and cheat in order to convict a defendant is hardly limited to forensic
analysts. It’s what leads some law enforcement officers to manufacture evidence of guilt or hide
evidence of innocence. But it’s particularly troublesome for forensic evidence for two reasons:
First, there are strong reasons why forensic analysts should not work in law enforcement
agencies in any capacity. The investigation of a crime is a basic police function, but testing
objects and trace evidence can be and often is done just as well—or better—by independent
scientists who have no other role in the case.
Second, precisely because they do not perform a core law enforcement function, forensic
analysts may be viewed within police departments as playing a secondary, supportive role in
investigations: police and prosecutors determine who’s guilty, and take them to court; forensic
experts provide ammunition to get them convicted.
When that happens, forensic officers may be rewarded for their efficiency in obtaining
convictions rather than their scientific skill. That’s why Joyce Gilchrist was promoted after she
303
This aphorism is routinely attributed to management guru Peter Drucker, but there is a debate about whether he
ever actually said it. Andrew Cave, Culture Eats Strategy For Breakfast. So What's For Lunch? Forbes, November 9,
2017.
304
National Research Council of The National Academies, Strengthening Forensic Science In The United States: A
Path Forward (2009) at 183.
305
R v. Ward, [1993] 96 Crim. App. 1, 68 (U.K.).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 157 • National Registry of Exonerations • September 1, 2020
became known as the “go to” expert in difficult homicide cases in Oklahoma City.
306
That’s why
Fred Zain, a serologist who was ultimately implicated in hundreds of cases of forensic fraud in
West Virginia and Texas, was promoted to head of the West Virginia State Police crime lab a
year after two officers who worked under him complained that he wrote reports based on blank
laboratory slides.
307
“[P]rosecutors relied on Zain because the remaining West Virginia
serologists were incapable, in their view, of reaching the ‘right’ results.”
308
After he moved on to
Texas, “several prosecutors expressed dissatisfaction with the reports they were receiving … and
specifically requested that the evidence be analyzed by Zain.”
309
The solution to the dangers of undue influence by law enforcement is straightforward. In a 2009
report on forensic science in the United States, a panel of the National Research Council of the
National Academies of Sciences, Engineering and Medicine recommended that:
“Scientific and medical assessment conducted in forensic investigations should
be independent of law enforcement efforts either to prosecute criminal suspects
or even to determine whether a criminal act has indeed been committed.
Administratively, this means that forensic scientists should function
independently of law enforcement administrators.”
310
In other words, crime labs should be run by forensic scientists rather than police officers. This
position enjoys broad support among legal and forensic scholars
311
for reasons that are more
basic and extend beyond preventing deliberate misconduct.
Forensic analysts who work for police forces are also in danger of unconsciously biasing their
findings to coincide with the conclusions already reached by the officers who arrested the
defendants and the prosecutors who charged them. Independent crime labs would reduce if not
eliminate that danger.
In addition, independent laboratories are likely to be more effective at preventing incompetence.
Joyce Gilchrist not only falsified and hid evidence in cases of charged defendants, she also
erroneously cleared two capital murderers whose blood was found at the scene of their crimes.
312
We know about those errors because other, innocent suspects were charged and convicted
306
See Section XII.1.c.
307
Court Invalidates a Decade of Blood Test Results in Criminal Cases, The New York Times, Nov. 12, 1993, p. A20.
308
Paul C. Giannelli, Independent Crime Laboratories: The Problem of Motivational and Cognitive Bias (2010) at 251.
309
In re Investigation of the W. Va. State Police Crime Lab., Serology Div., 438 S.E.2d 501 at 513 n.16 (W. Va.
1993) (referring to deposition of T.S. Smith).
310
National Research Council Of The National Academies, Strengthening Forensic Science In The United States: A
Path Forward (2009) at 24.
311
Paul C. Giannelli, Independent Crime Laboratories: The Problem of Motivational and Cognitive Bias (2010) at 251;
Paul C. Giannelli, The Abuse of Scientific Evidence in Criminal Cases: The Need for Independent Crime
Laboratories, 4 Va. J. Soc. Pol’y & L. 439, 441 (1997); Sandra Guerra Thompson, Cops in Lab Coats, Carolina
Academic Press (2015) at 231-32.
312
Supra Section XII.1.c.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 158 • National Registry of Exonerations • September 1, 2020
instead, and ultimately exonerated. We have no idea how many guilty suspects were never
charged because of Gilchrist’s errors in cases that did not lead to false convictions and
exonerations. If Gilchrist’s supervisors had been forensic scientists themselves, they might have
been less dazzled by her ability to convince jurors to convict, and more aware of her professional
ineptitude.
The main problem is not that police chiefs deliberately or even negligently permit bad practices
in crime labs. As Professor Sandra Guerra Thompson points out, even with the best of
intentions, “[o]ne cannot supervise what one does not understand.”
313
The Houston crime lab. How much of a difference can the structure of a forensic lab make?
Consider Houston:
In November 2002, the Houston Police Department Crime Laboratory ran into a major crisis.
Two outside experts, working with a local television news program, identified serious problems
in the lab’s work in seven DNA and serology cases: deficient documentation of procedures and
results; mistakes in analyses; errors in calculating the meaning of their results; and
mischaracterizations of those results in testimony.
314
Less than a month later, an audit by the
Texas Department of Public Safety Crime Laboratory System concluded that the serology and
DNA section of the Houston Police Crime Lab “was in shambles.”
315
The section was closed
immediately.
In January 2003, one of those outside experts, Professor William Thompson of the University of
California, Irvine, reviewed the lab’s DNA work in the rape and kidnapping conviction of Josiah
Sutton and concluded that Sutton was innocent. The crime lab asked an outside lab to review its
work, and that lab reported that the DNA evidence used in court was “probably incorrect.” On
March 10, 2003, retesting by an independent lab found that Sutton did not contribute to the
semen recovered at the crime scene. He was released from prison two days later and exonerated
that May. In the wave of publicity that followed Sutton’s exoneration, the lab was repeatedly
described as the worst in the country.
316
This crisis was hardly the first sign of trouble at the Houston Police lab, but it was the most
serious.
317
It generated 25 investigations by the Houston Police Department, two grand jury
investigations, and finally, in 2007, a major report by a former Inspector General of the United
States Department of Justice, a team of lawyers and a scientific advisory board.
318
That report
313
Sandra Guerra Thompson, Cops in Lab Coats, Carolina Academic Press (2015) at 182.
314
Michael R. Bromwich, “Final Report of the Independent Investigator for the Houston Police Department Crime
Laboratory and Property Room” at 54-57 (June 13, 2007), [hereinafter Bromwich Report].
315
Id. at 8.
316
See, e.g., Adam Liptak, Houston DNA Review Clears Convicted Rapists, and Ripples in Texas Could Be Vast,
N.Y. Times, March 11, 2003; William C. Thompson and Cliff Spiegelman, Who Should Control Houston’s Crime Labs,
Houston Chron., May 31, 2016.
317
Bromwich Report at 24-49; Thompson, Cops in Lab Coats, supra note 313, Ch.7, at 205-230.
318
Bromwich Report at i-ii, 58-59.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 159 • National Registry of Exonerations • September 1, 2020
describes the “Sutton case [as] a microcosm of the range of problems” they observed in the
handling and testing of DNA evidence by the Houston Police lab: poor technical skills and
inadequate training of the forensic analyst led to ambiguous results; and the analyst
misrepresented those results and testified that they showed a much higher probability that
Sutton participated in the crime than would have been warranted even if they had been
accurate.
319
In April 2014, the Houston Police Department Crime Laboratory closed its doors and
transferred its responsibilities to the Houston Forensic Science Center, an independent
laboratory, administered by forensic scientists, with adequate funding. The new crime
laboratory is now often described as one of the best in the United States.
320
Most of the problems at the old Houston Police lab involved poor performance rather than
misconduct: incompetence, inadequate training, useless supervision, and insufficient resources.
But misconduct played a role. In the Sutton case, a forensic analyst incorrectly analyzed the
results of the DNA test she conducted, and then misreported her own findings at trial to lend
support to the prosecution’s theory that Sutton was guilty. The 2007 report identified several
other cases of clear misconduct in DNA testing
321
and in drug identification,
322
and of course,
many other instances will have gone undetected, especially given the lab’s deeply flawed record
keeping.
The new Houston Forensic Science Center has had problems, but it has dealt with them openly
and effectively.
323
The change in culture from the old Houston Police lab is unmistakable, even
from a distance.
The process that brought about that change was drastic. The now-defunct police lab was
effectively put into receivership for 10 years, major parts of its work were halted entirely and
outsourced to other labs, and it was ultimately replaced entirely by a new, better funded and
better run independent crime laboratory that was built from scratch, physically and
organizationally. Maybe that’s what it takes for a crime lab to go from worst to best in eleven
years. It’s a big deal, but we’ve seen it done. Healthier police labs could no doubt restructure and
become independent over time, without ever closing down.
The Houston Forensic Science Center is not unique. In a few states, independent crime labs
have existed for decades, typically under the auspices of medical examiner offices, and a couple
319
Id. at 14, 217-219.
320
Nicole B. Cásarez & Sandra G. Thompson, Three Transformative Ideals to Build a Better Crime Lab, 34 Ga. St. U.
L. Rev. 1007, 1012 (2018) ("In the short time since the laboratory assumed its corporate structure, the HFSC has
already produced remarkable outcomes that set the city’s laboratory apart from any other forensic laboratory in the
country, earning it local, national, and international recognition."); Jed Rakoff, Jailed by Bad Science, The New York
Review of Books, Vol. LXVI, No. 20 (Dec. 19, 2019) at 79-80, 85.
321
Bromwich Report at 104.
322
Id. at 153-157.
323
See Samantha Ketterer, Houston Crime Lab Fires Investigator After Alleged Testing Policy Violation, Houston
Chron., October 26, 2018.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 160 • National Registry of Exonerations • September 1, 2020
of other jurisdictions changed from police to independent labs since the National Research
Council report in 2009.
324
For now, however, police crime labs are still the rule.
Caveat. Independent crime labs are not cure-alls. No formal administrative structure will
overcome inadequate resources or sloppy supervision. The most notorious cases of forensic
fraud in recent years involved two analysts at the Massachusetts state crime lab who falsified
results in tens of thousands of drug tests at a time when the lab was run by the Massachusetts
Department of Health's Office of Human Services. One of them faked tests in thousands of cases
to improve her performance evaluations (Annie Dookhan);
325
the other did so because she was
an addict who used the drugs herself (Sonja Farak).
326
Annie Dookhan and Sonja Farak managed to fly under the radar, faking results in mostly low
level drug cases that overwhelmingly ended in guilty pleas.
327
It’s an outrage that they were able
to do it in so many cases, for so long. But even a dysfunctional independent crime lab might
have prevented Joyce Gilchrist and Fred Zain from achieving star status by systematic fraud in
highly visible rape and murder trials.
ii. Police
In 2019, the Houston Forensic Science Center had 200 employees
328
and a budget of about $27
million.
329
That year the Houston Police Department had about 6,500 employees, including
5,300 sworn police officers,
330
and a budget of over $900 million.
331
The crime lab performs
functions that can, if necessary, be handled by other public or private labs at other locations. The
police department is responsible for a host of essential functions—emergency services, physical
safety, traffic control, order maintenance, crime prevention, criminal investigation,
apprehension and arrest of suspects—most of which can only be done by police officers, at the
scene of the relevant events.
The radical cure that worked for the Houston crime lab—closing it down and building a new one
from the ground up—is not an option for a major police force.
324
Thompson, Cops in Lab Coats, supra at 182, notes 7-9.
325
Shawn Musgrave, The Chemists and the Cover-Up, Reason, Feb. 9, 2019.
326
Id.
327
The Dookhan and Farak scandals ultimately resulted in the dismissal of thousands of drug convictions, mostly
guilty pleas to misdemeanors. Id. These cases are not included in the Registry for the reason we discussed in
III.3.c.i, on Group Exonerations: the dismissals did not alter the result of individual post-conviction re-examinations of
the evidence in each case.
328
Hannah Dellinger, COVID-19 in staff puts Houston Forensic Science Center ‘precariously close’ to limiting crime
scene responses, Houston Chronicle, June 29, 2020.
329
Houston Forensic Science Center, Inc., FY19 Budget.
330
Houston Police Department, Wikipedia, (last visited June 28, 2020).
331
City of Houston, “Proposed Operating Budget for the Period July 1, 2018 to June 30, 2019.”
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 161 • National Registry of Exonerations • September 1, 2020
Police forces in America are quasi-military organizations. Officers wear uniforms and have
ranks, from Officer or Deputy, through Corporal, Sergeant, Lieutenant and Captain, to Chief,
Commissioner or Sheriff. This suggests a command structure in which superior officers can
institute change by issuing orders that subordinates are required to obey. That is not the case.
A patrol officer may have the lowest rank in the police hierarchy, but her job is nothing like that
of a private in an infantry platoon. Most police work is done away from police precincts, by
officers acting alone or in pairs, responding to calls, patrolling neighborhoods and investigating
crimes. Ongoing supervision is impossible, and yet every officer has a great deal of authority
over the civilians she interacts with, and discretion to decide whether and how to exercise that
authority.
It doesn’t help that police officers work in a notoriously insular subculture. They see themselves
(with some justice) as performing indispensable, difficult and sometimes dangerous work that
civilians don’t understand but are often quick to criticize. That produces a sense of mission
combined with cynicism, suspicion of outsiders and top commanders—and solidarity with other
officers. One manifestation is the Blue Wall of Silence that protects many officers who commit
crimes from exposure by fellow officers; another is resistance to change.
In addition, police chiefs are hemmed in by external forces. Local elected officials have their own
agendas; influential police unions exercise substantial power in local politics, and oversee
employment contracts that limit and sometimes cripple the commanders’ authority to discipline
subordinates.
332
In some departments, the disciplinary process has become such a dysfunctional
tangle of hearings and appeals that officers who ignore orders or break rules face few
consequences, if any.
333
A lot has been written about reforming American police departments. A common theme is that
reforms are difficult to implement, and more difficult to sustain.
334
But the reforms that have
received most attention all concern basic aspects of police interactions with civilians:
community relations; use of force; responsiveness to requests for assistance and to complaints;
patterns of stops and arrests; race relations. The type of reforms that we’re concerned with are
more specific and limited: procedures for collecting and preserving testimonial and physical
332
See DeRay McKesson, et al., Police Union Contracts and Police Bill of Rights Analysis, Campaign Zero (June 29,
2016); Editorial Board, Take the lid off probes of officers tied to wrongful convictions, Chicago Sun Times, Jun. 14,
2020.
333
For example, in August 2019, two Chicago police officers were fired “a decade after they were first accused of
hitting a child in their care and failing to seek medical attention for the 8-year-old boy, who suffered fractures to the
face and arm that the child said was inflicted by one of the officers.” The Chicago Police Board said it was “deeply
troubled” that the case took so long to resolve. Jeremy Gorner, Two Chicago Police Officers Were Accused of Child
Abuse. A Decade Later, They’ve Been Fired, Chicago Tribune, Aug. 27, 2019.
334
See Samuel Walker, Institutionalizing Police Accountability Reforms: The Problem of Making Police Reforms
Endure, 32 Saint Louis University Public L. Rev. 57 (2012); Trent Ikerd & Samuel Walker, Making Police Reforms
Endure, The Keys for Success, U.S. Department of Justice Office of Community Oriented Policing Services (2010);
Wesley G. Skogan (2008) Why reforms fail, Policing & Society, 18:1, 23-34. See generally Samuel Walker & Charles
Katz The Police in America: An Introduction, 9th Edition (2017).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 162 • National Registry of Exonerations • September 1, 2020
evidence in criminal investigations. It may be that changes in that limited and largely reactive
domain are easier to achieve.
In the worst abuses, “reform” does not require any change in operating rules. In Chicago in the
1980s, Jon Burge’s regime of systematic torture was tolerated by his superiors—in fact,
encouraged—and other police officers followed his sadistic lead.
335
Torture became routine. It
took decades to untangle that knot, and it was a major accomplishment; institutional culture
was a serious obstacle, even a culture so pathological that torture was normalized. But the end
product is simply that officers have (mostly) stopped violating basic, longstanding rules against
extreme physical abuse.
No one would cheer if a hospital with an unacceptable rate of post-operative mortality was
finally able, in 2020, to get all surgeons to always wear sterile surgical gloves. It’s not a “reform”
to crack down on bad behavior that should never have been permitted. Ending an unacceptable
practice might be hard—violent interrogations, or deadly surgical hygiene—but the surgeons or
officers involved all knew the rules, and probably followed them most of the time.
The more novel reforms in police investigations are not prohibitions on misconduct but
affirmative rules on how to gather evidence. Their goal is to obtain and preserve unbiased
evidence—and, as we’ve mentioned, they may also reduce misconduct along the way.
(a) Recorded interrogations
There is an easy and effective remedy for misconduct in police interrogation, as we have
discussed: record the whole process, preferably on video.
336
Most violence and other misconduct
just won’t happen on camera—and if it does, it’ll be there for the world to see. In addition, and
perhaps more important, recording may reduce or end ugly permitted practices that also cause
false confessions, and it will give judges and juries the information they need to evaluate the
voluntariness and truthfulness of a confession.
All of this was known 20 years ago. Nonetheless, as of 2002, only Alaska and Minnesota
required recording interrogations, both under decisions by their state supreme courts,
337
and the
federal Department of Justice had a written policy against recording.
338
By 2019, 24 additional
states and the District of Columbia had added statutes or court rules that required recording all
335
Supra Section XII.1.b.
336
See supra Section V.4.e.
337
Brandon L. Bang, et al., Police Recording of Custodial Interrogations: A State-by-State Legal Inquiry, 20 Int’l J. of
Police Science & Management 4, 10 (2018).
338
Thomas P. Sullivan, Recording Federal Custodial Interviews, 45 Am. Crim. L. Rev. 1297, 1297 (2008).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 163 • National Registry of Exonerations • September 1, 2020
or some types of interrogations (usually homicides, especially capital murders),
339
and the
Department of Justice had reversed course 180 degrees.
340
Clearly a major change is underway. At first blush, it appears to be driven by courts and
legislatures. In fact, police departments have played a central role in the process.
In two states, police have implemented state-wide recording requirements independent of court
mandates or legislative action—which brings the total number of states that require some
recording by all police agencies to 26. In Hawaii, all four police departments in the state require
recording interrogations for “serious crimes.”
341
And the Rhode Island Police Accreditation
Commission requires all 43 Rhode Island police departments to implement mandatory
recording procedures for interrogations in murder cases.
342
More generally, the police have played an important role in the spread of recorded
interrogations in all states, both those that have statutes or court rules and those that don’t. By
2004, when no state other than Alaska and Minnesota had yet implemented a recording
requirement, at least 240 police departments in 37 states had done so—most of them 10 years
before or earlier.
343
Those initiatives, and the positive experiences the departments reported, paved the way for
statewide changes. Every state that has adopted a statewide recording rule had police
departments that did so earlier on their own. The same is true now in the substantial but
dwindling minority of states without statewide rules: Numerous local police forces in those
states do record some or all interrogations—and that may lead the way to more reforms at the
state level.
344
(b) Improved eyewitness identification procedures
The danger of eyewitness misidentification has been well known for many years.
In 1932, Edwin Borchard’s classic book, Convicting the Innocent,
345
described eyewitness errors
as the most common cause of the 65 documented false convictions Borchard collected. And in
339
See Bang et al., supra note 337, at Table 1, pp. 11-12, for the status of these rules in 2017. Since then, two
additional states have added recording requirements: Nevada, see Nev. Rev. Stat. Ann. § 171.1239 (West), and
Oklahoma, see Okla. Stat. Ann. tit. 22, § 22 (West).
340
Office of Pub. Affairs, “Attorney General Holder Announces Significant Policy Shift Concerning Electronic
Recording of Statements,” U.S. Dep’t of Justice, May 22, 2014; Michael S. Schmidt, “In Policy Change, Justice Dept.
to Require Recording of Interrogations,” N.Y. Times, May 22, 2014.
341
Bang et al., supra note 337, at p.13.
342
Id.
343
Thomas P. Sullivan, Police Experiences with Recording Custodial Interrogations, Northwestern University School
of Law: Center on Wrongful Convictions, Appendix A (2004).
344
For example, as of February 25, 2019, there were 31 such police departments in Arizona, 33 in Ohio, 22 in
Tennessee, and 58 in Florida. “Electronic Recording Project” National Association of Criminal Defense Lawyers,
Feb. 25, 2019.
345
Edwin M. Borchard, Convicting the Innocent, Errors of Criminal Justice (1932).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 164 • National Registry of Exonerations • September 1, 2020
1967—decades before any serious concerns about false confessions that were not obtained by
violence—the Supreme Court observed that “The vagaries of eyewitness identification are well
known; the annals of criminal law are rife with instances of mistaken identification,”
346
and
quoted an author who wrote: “The influence of improper suggestion upon identifying witnesses
probably accounts for more miscarriages of justice than any other single factor—
perhaps…[more] than all other factors combined.”
347
As we’ve noted,
348
eyewitness identification has also been the subject of extensive psychological
research—and of proposals for reform.
349
For the most part, they focus on four areas of
procedural reform: using a lineup administrator who does not know the identity of the suspect
and therefore cannot intentionally or unintentionally suggest that a witness identify that person;
proper instructions to witnesses; proper selection of fillers to avoid a suggestive lineup; and
recording the entire process, including the witness’s initial level of confidence in any
identification that is made.
350
In 1999, the Department of Justice released a set of recommendations entitled Eyewitness
Evidence, A Guide for Law Enforcement.
351
The release of that “guide” is often described as a
turning point for public policy on eyewitness identification, but for the most part, change on the
ground didn’t begin in earnest until about 2010.
In 2013, the Police Executive Research Forum released a study of identification policies and
practices in 619 departments that responded to a survey that was sent in 2011 to 1,377 of the
approximately 18,000 American police departments.
352
They found that about “56 percent of all
responding agencies reported one or more changes in policy or procedure since 1999,”
353
but
half of these changes were made in 2010 or 2011.
354
We can’t generalize from this sample to all police departments in the country. It’s likely, for
example, that departments that undertook reforms were more likely to respond to the survey
than those that didn’t. But even if the rate of reform is overstated by a factor of five, the total
346
United States v. Wade, 388 U.S. 218, 228 (1967).
347
Patrick Wall, Eye-Witness Identification in Criminal Cases (1965).
348
Supra Section IV.3.
349
See Gary L. Wells et al., Policy and Procedure Recommendations for the Collection and Preservation of
Eyewitness Identification Evidence, 44 L. & Hum. Behav. (2020).
350
See supra IV.2.
351
National Institute of Justice, Eyewitness Identification: A Guide for Law Enforcement (Oct. 1999); see also
National Institute of Justice, Eyewitness Evidence: A Trainer‘s Manual for Law Enforcement (Sept. 2003).
352
Police Executive Research Forum, A National Survey of Eyewitness Identification Procedures in Law Enforcement
Agencies (2013).
353
Id. at xi.
354
Id. at 69-70, Table 24.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 165 • National Registry of Exonerations • September 1, 2020
number of departments that addressed the issue in some respect between 1999 and 2011 might
be in the thousands.
There have also been changes at the state level in eyewitness identification practices. By early
2020, 31 states had some set of statewide rules, standards, or model policies—from minimal to
detailed and exacting.
355
Two-thirds of these statewide reforms took place after 2011—that is,
after hundreds, if not thousands, of local police departments had taken similar steps on their
own.
In other words, as with recording interrogations, local police departments—the most numerous
and dispersed agencies in the criminal justice system—led the way.
Local police departments are also essential for any statewide reform to succeed in practice.
Virginia, for example, adopted a comprehensive model eyewitness identification policy in 2011.
Professor Brandon Garrett has conducted two studies of how this policy was implemented. In
2013, he found that “the vast majority of Virginia law enforcement agencies still followed earlier
and outdated” eyewitness identification practices.
356
Five years later, Garrett found that “[b]y
2018, the vast majority of Virginia residents live in jurisdictions in which best practices
regarding eyewitness identifications have been adopted.”
357
Garrett concludes that over time,
“self-policing [by local police departments]…improved police policy.”
358
Professor Keith Findley
studied the implementation of an eyewitness identification policy in Wisconsin, and concluded
that a state mandate that provides guidance but requires police departments to adopt their own
policies is more likely to succeed than one that attempts to impose a uniform policy from on
high.
359
These are affirmative reforms. They set the terms for conducting interrogations and lineups
correctly, and reduce misconduct along the way. A crackdown on deliberate misconduct—
witness tampering, for example, or concealing exculpatory evidence—might be harder to
implement.
We can’t say how common these police-initiated reforms have been, or how well the new rules
are followed in the departments that have adopted them. Our point, rather, is that local police
departments can initiate investigative reforms that are designed to reduce both errors and
misconduct; we’ve seen it done. In fact, they seem to be more effective at that task than other
law enforcement agencies.
355
National Registry of Exonerations, State Eyewitness Identification Procedures, compiled the from Innocence
Project website, State Policy pages (March, 2020).
356
Brandon Garrett, Eyewitness Identifications and Police Practices: A Virginia Case Study, 2 Virginia J. of Crim. L. 1,
17 (2014).
357
Brandon Garrett, Self-Policing: Dissemination and Adoption of Police Eyewitness Policies in Virginia, 105 Va. L.
Rev. Online 96 (2019).
358
Id.
359
Keith Findley, Implementing the Lessons from Wrongful Convictions: An Empirical Analysis of Eyewitness
Identification Reform Strategies, 81 Mo. L. Rev, 377 (2016).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 166 • National Registry of Exonerations • September 1, 2020
iii. Prosecutors
In all but a few states,
360
county and district prosecutors are the most powerful actors in the
American system of bringing criminals to justice. They have the exclusive power to decide
whether and who to charge for most known or suspected crimes, and what charges to file. When
they decide not to file charges, their decisions are unreviewable. After a defendant has been
charged, they have virtually unlimited power to dismiss the case entirely, or to reduce the
charges or limit the punishment in return for a guilty plea without trial, which happens in about
95% of felony cases in which defendants are convicted. They prosecute the few criminal cases
that do go to trial, and they have a major voice in the sentences of those defendants who are
convicted after trial.
This enormous power is entrusted to a chief prosecutor known as District Attorney, Prosecuting
Attorney, County Attorney, State’s Attorney, or some other less common name. The great
majority are elected directly by the voters in the geographic units they serve, usually counties.
Typically, they obtain most of their funding from county boards of supervisors or other elected
local politicians, but otherwise they do not report on their administrative or legal decisions to
any superior officials.
A newly elected prosecutor, like a new police chief, may face a local culture that tolerates or
encourages misconduct. New Orleans is a good example.
361
John Thompson, as we discussed,
362
was sentenced to death in 1985 because a New Orleans
prosecutor hid (and eventually lost or destroyed) a critical item of physical evidence: a piece of
cloth with a stain of the criminal’s blood that tests had proven could not have come from
Thompson. In 2012, nine years after he was exonerated, the Supreme Court considered
Thompson’s claim that he was the victim of a systematic practice of concealing exculpatory
evidence by the Orleans Parish District Attorney’s office under Harry Connick, Sr. who ran the
office from 1974 through 2003.
363
A year after they decided Thompson’s case, the Supreme
360
In Alaska, Delaware and Rhode Island, state-wide attorneys general have primary responsibility for criminal
prosecutions throughout the states. George Coppolo, States That Elect Their Chief Prosecutor, Connecticut General
Assembly (Feb. 24, 2003).
361
But not unique. In August 2019, for example, Chief Judge Julie A. Robinson of the United States District Court for
the District of Kansas issued an opinion concluding that the “culture of the Office of the United States Attorney” for
Kansas had produced a situation in which “the fairness of the adversary system is called into question by systemic
prosecutorial misconduct.” Specifically, the office had been systematically recording and listening to confidential
conversations between defense attorneys and their clients. U.S. v. Karl Carter, Case No. 2:16-cr-20032-JAR,
Document 758, Filed 08/13/19, at pp.6, 184. A former senior attorney in that office later described how he had told
the United States Attorney “about systemic prosecutorial misconduct on a daily basis … in person and in documented
form,” and that similar reports were made by “defense attorneys, probation officers, citizens, law enforcement and the
courts,” all to no avail. Mike Warner, Letter to the editor: Why I resigned from the Kansas U.S. Attorney’s Office.
362
Supra Section III.3.a.
363
Connick v. Thompson, 563 US 51 (2011).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 167 • National Registry of Exonerations • September 1, 2020
Court reversed a 1995 murder conviction from New Orleans because exculpatory evidence was
concealed.
364
Justice Ginsberg, in her separate opinion in Thompson’s case, summed up her view of the
problem: “Connick created a tinderbox in Orleans Parish in which Brady violations [concealing
exculpatory evidence] were nigh inevitable. And when they did occur, Connick insisted there
was no need to change anything….”
365
The numbers support Ginsberg’s description. Orleans Parish has the highest per capita
exoneration rate of any county in the country. In 78% of exonerations from New Orleans,
exculpatory evidence was concealed (18/23); prosecutors committed misconduct in nearly 90%
of those cases (16/18).Three of those exonerations (including Thompson’s) reached the Supreme
Court; in each, the defendant was sentenced to death because Connick’s deputies concealed
exculpatory evidence.
366
Looking beyond exonerations, we know of 51 criminal cases from New Orleans in which courts
of have found—or the prosecution has acknowledged—that the New Orleans District Attorney’s
office violated Brady, plus several others in which strong claims of Brady violations were
bypassed when the convictions were reversed for other reasons.
367
All of these cases were
decided after Harry Connick, Sr. became District Attorney. The actual total of cases with
concealed exculpatory evidence is much higher because the great majority of criminal
defendants don’t go to trial, don’t appeal, and never learn about evidence that was deliberately
concealed. Like torture in Chicago, concealing exculpatory evidence in New Orleans became
routine.
368
Can this sort of prosecutorial culture be changed? We think so, but we can’t yet say whether it’s
happened in New Orleans. Leon Cannizzaro, who has served as District Attorney for the last
twelve years, decided in late July 2020 not to seek reelection.
369
His tenure was marred by his
defense of the office against lawsuits and complaints stemming from the Connick era, and by a
364
Smith v. Cain, 565 U.S. 73 (2012).
365
Connick, supra, 563 U.S. at 108.
366
Kyles v. Whitley, 514 U.S. 419 (1995); Connick v. Thompson, 563 US 51 (2011); Smith v. Cain, 565 U.S. 73
(2012).
367
Robert Jones v. Leon Cannizzaro, Jr., E.D. LA., No. 18-cv-503, Expert Report of Professor Laurie Levenson,
Document 230-2 Filed 01/28/20, pp. 8-9.
368
See Ellen Yaroshefsky, New Orleans Prosecutorial Disclosure in Practice After Connick v. Thompson, 25 Geo. J.
Legal Ethics 913, 934 (2012); Janet Moore, Democracy and Criminal Discovery Reform After Connick and Garcetti,
77 Brook. L. Rev. 1329, 1342 (2012).
369
Matt Sledge & John Simerman, New Orleans DA Leon Cannizzaro says he won't run again, ending era and
controversial career, The Times-Picayunne, July 24, 2020.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 168 • National Registry of Exonerations • September 1, 2020
separate scandal involving the office’s use of phony “subpoenas” to coerce interviews with
witnesses.
370
To date, the Registry lists five exonerations of defendants who were convicted in New Orleans
since Connick left office in 2003. In one, an exoneration from a 2011 conviction, we know of
exculpatory evidence that was concealed. Perhaps the Orleans Parish District Attorney’s Office
has turned the corner on concealing exculpatory evidence, but it’s too early to say, since we will
see additional New Orleans exonerations from that period in the years to come, possibly more
than we have already seen.
An elected prosecutor who wants to change an established pattern of misconduct can exercise
much tighter control over her subordinates than most police chiefs. Unlike police officers,
deputy prosecutors work primarily in their offices and in court, within a stone’s throw of
supervisors. Their most important tasks are conducted (or completed) in public, on the record,
frequently in writing. Every formal action they take is done in the name of the head of the office,
who can always override them—and who, unlike most police chiefs, usually has the power to fire
or demote them.
If a county prosecutor announces a policy on handling criminal cases in her jurisdiction—
marijuana possession will no longer be charged as a crime, for example, or a defendant who is
charged with spouse abuse will not be allowed to plead guilty to a non-violent crime—it is likely
to be carried out.
A prosecutor has the power to attack official misconduct in criminal cases by several means. She
can order her deputies not to commit specific types of misconduct—concealing exculpatory
evidence, for example—or direct them to follow protocols that make that misconduct impossible,
such as open file discovery. She can discipline or discharge deputies who violate those orders or
commit other types of misconduct. She can dismiss charges in cases that are tainted by
misconduct by her own deputies or by other law enforcement officials, refuse to file charges
when the arresting officer has a serious history of misconduct, or decline to call such officers as
prosecution witnesses. She can prosecute police officers who intimidate or abuse defendants or
witnesses, or frame innocent suspects; and she can prosecute any official who commits or
procures perjury, or obstructs justice. She can reinvestigate past cases to see if misconduct was
committed or miscarriages of justice occurred, and exonerate any innocent defendants she
identifies.
In practice, local prosecutors’ power is limited by the resources at their disposal, by political
constraints, and by the independent authority of other law enforcement agencies, especially the
police. All the same, they have more power to prevent misconduct in criminal cases than any
other public officials.
Over the last dozen years, a growing number of prosecutors around the country have taken some
of these steps. The most common is the creation of a Conviction Integrity Unit (CIU) with a
370
Id.; Nicholas Chrastil, Three qualify to run for DA on first day, Cannizzaro not among them, The Lense, July 22,
2020.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 169 • National Registry of Exonerations • September 1, 2020
mandate to investigate possible wrongful convictions in the jurisdiction of the prosecutor’s
office, to secure exonerations when they conclude that the convicted defendants are innocent,
and—in some offices—to train prosecutors and other officials to avoid future false convictions.
371
Of the possible tools for preventing future misconduct, exonerating innocent defendants is the
least direct—but it may be influential, especially in conjunction with other, less visible policies.
The Brooklyn District Attorney’s Conviction Review Unit is one of the most vigorous CIUs in the
country. It was created by District Attorney Charles Hines in 2011, and became far more active
under his successor, Ken Thompson, who took office in 2014.
372
All told, the Brooklyn CIU has
been responsible for 27 exonerations, 21 of them murder cases. Eight of these exonerations, all
murder cases, involve extreme misconduct by now-disgraced retired police detective Louis
Scarcella, who lied, fabricated evidence, procured perjured testimony and coerced confessions in
many cases from the late 1980s through the 1990s, including five additional murder
exonerations that the CIU did not sponsor. Nine other Brooklyn CIU exonerations involved
misconduct by police officers other than Scarcella, and 10 included misconduct by prosecutors.
For example, in 2010, Wayne Martin was convicted of first-degree murder for shooting and
killing the owner and an employee of a tire store in Brooklyn. In 2016, lawyers in the Brooklyn
District Attorney’s Office discovered that their files contained two different versions of the
prosecution’s summary of the evidence. The complete summary said that the only eyewitness to
testify against Martin had identified another man as the gunman, but that paragraph was
missing from the version that was turned over to Martin’s defense attorney. A further review of
the prosecution file by the CIU turned up a police report of an interview with a witness who said
that he saw another suspect commit the shootings. That document had also been concealed from
Martin’s defense lawyer at trial. In September 2016, the District Attorney moved to dismiss the
charges against Martin.
Detective Scarcella retired in 1999, more than a decade before the first of his cases unraveled.
He has not been prosecuted for any crime, although he has been sued for money damages.
373
Marc Fliedner, the prosecutor who obtained Wayne Martin’s conviction, left the District
Attorney’s office before that case was reopened, and has never been disciplined for his conduct.
Still, these exonerations (and others) send a message that current prosecutors and police officers
might heed: Falsifying evidence of guilt and concealing evidence of innocence are not tolerated.
Since 2014, several prosecutors who campaigned on promises to reform the criminal justice
system have been elected in urban counties across the country—Boston, Chicago, Houston,
Philadelphia, Saint Louis, and San Francisco, among others. All inherited or created conviction
371
National Registry of Exonerations: Conviction Integrity Units.
372
In July 2020, the Kings County (Brooklyn) District Attorney’s Office released a report of the 25 exonerations by
that CIU that were initiated after Ken Thompson began his term, 426 Years: An Examination of 25 Wrongful
Convictions in Brooklyn, New York, District Attorney, Kings County, July 9, 2020. The report contains useful
descriptions of the cases and the patterns they fall into, but it’s value for our purposes is limited because the office
was obliged to anonymize the information reported.
373
Andrew Keshner, Family of Wrongly Convicted Brooklyn Man Suing Ex-Detective Louis Scarcella Who Coerced
Him Into Confession, N.Y. Daily News, Nov. 2, 2017.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 170 • National Registry of Exonerations • September 1, 2020
integrity units. Otherwise, most of the reforms they have proposed or enacted address issues
other than false convictions: reducing or eliminating money bail to avoid unnecessary pretrial
detention, declining to file charges for drug possession or other minor non-violent crimes,
reducing the sentences they demand in return for guilty pleas or recommend after conviction at
trial, and so forth.
Several of these reform-minded prosecutors have also adopted two types of proactive policies
designed to prevent official misconduct and the false convictions it may cause: providing
defense attorneys with open file discovery; and compiling lists of police officers whom they will
not call as witnesses because of their past misconduct, or will not rely on as a basis for filing
criminal charges.
374
A big issue in the background is the extent to which reform minded prosecutors will be able to
reduce misconduct by police officers. Chesa Boudin, the recently elected District Attorney of San
Francisco, announced an ambitious approach: “Prosecutors must do now what we should have
done long ago: Prosecute offending officers and adopt measures that focus on increasing
transparency, accountability and equal justice in a system plagued with impunity for the police
and racial injustice.”
375
Easier said than done—but it will never be done unless it’s said.
This spate of reforms is a promising development, but it’s too early to assess their impact. We
know of 62 CIUs—a small number in a country with over 2,300 prosecutorial offices, but almost
two-thirds of them are in counties with more than a million people. All told, those prosecutors
serve a population of nearly 120 million. Between them, these CIUs have participated in 434
exonerations—but more than half of these units have worked on only a single exoneration, or
none at all, and quite a few appear to be little more than window dressing.
376
If these programs remain in place—and if they spread widely across the country—they might
become highly influential. Both of these conditions, however, are uncertain.
There’s been considerable resistance to all these measures, and to the prosecutors who initiated
them. At the national level, Attorney General William Barr has excoriated state prosecutors who
“style themselves as ‘social justice’ reformers” but “spend their time undercutting the police,
letting criminals off the hook, and refusing to enforce the law.”
377
The results, Barr said, “will be
predictable. More crime; more victims.”
378
Strong words—with no evidence to back them up—
but the important opposition (like most important facts about American criminal justice) is
local.
374
See, e.g., Office of the State Attorney for the Ninth Judicial Circuit, Orange & Osceola County, State Attorney
Ayala Institutes an Enhanced Brady Policy for Recurring State Witnesses; Matt Kawahara, Boudin bans prosecutions
based on sole testimony of untrustworthy police officers, San Francisco Chronicle, June 15, 2020.
375
Chesa Boudin, The Police Answer to Us. What Will We Do About It?, The New York Times, July 27, 2020.
376
The National Registry of Exonerations, Exonerations in 2018 (April 9, 2019); “Conviction Integrity Units,” National
Registry of Exonerations, (last visited July 3, 2020).
377
Michael Balsamo, Barr Defends Police, Takes Swipe at Progressive Prosecutors, PBS News Hour, Aug. 12, 2019.
378
Id.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 171 • National Registry of Exonerations • September 1, 2020
Conviction integrity units—the most common innovation—are popular with voters. In some
elections, competing candidates have campaigned on the promise that they would do a better
job of running a CIU; we know of no candidate for a position as local prosecutor who has
opposed such a unit. Nonetheless, CIUs have run into opposition from other actors within the
criminal justice system.
In St. Louis, when the District Attorney (on the recommendation of her CIU) moved to dismiss a
24-year-old murder conviction on grounds of innocence, the judge took the extraordinary step
of asking the Missouri Attorney General to intervene. With the support of the Attorney General,
the judge ruled that the District Attorney had no power to act in the case; that ruling is now on
appeal.
379
In Maryland, after a few trial court judges denied motions by district attorneys to
vacate old convictions on the ground that the defendants were innocent, the issue was settled by
the legislature, which passed a law in 2019 explicitly granting them the power to do just that.
380
In 2020, Utah followed suit with the Conviction Integrity Units Act.
381
Open file discovery has not met with strong political opposition that we know of. But the
creation of lists of police officers who, because of prior misconduct, will not be used as witnesses
or trusted in charging decisions has been attacked by police unions, in public and in court.
382
The real issues, of course, are political. Larry Krasner, a former criminal defense and civil rights
lawyer, was elected as District Attorney of Philadelphia in 2017, and began his term in 2018. He
is probably the best known of the recent crop of progressive prosecutors,
383
and has attracted
more than his share of attacks. For example:
The President of the Philadelphia’s Fraternal Order of Police has not only called Krasner
“anti-law enforcement” and publicly referred to him and his supporters as “the parasites
of the city,”
384
but placed billboards on the main highway through Philadelphia that say
“HELP WANTED: NEW PHILADELPHIA DISTRICT ATTORNEY.”
385
379
Richard A. Oppel, These Prosecutors Promised Change. Their Power Is Being Stripped Away, N.Y. Times, Nov.
25, 2019.
380
Justin Fenton, Maryland Lawmakers Pass Bill to Make It Easier for Prosecutors to Overturn Conviction, Baltimore
Sun, April 10, 2019.
381
Utah Code 78B-9-501-503.
382
See Barry Scheck, The Integrity Of Our Convictions: Holding Stakeholders Accountable In An Era Of Criminal
Justice Reform, 48 Geo. L.J. Ann. Rev. Crim. Proc. iii, xiii-xxi (2019).
383
See, e.g., Jennifer Gonnerman, Larry Krasner’s Campaign to End Mass Incarceration, New Yorker, Oct. 22, 2018;
Ben Austen, In Philadelphia, A Progressive D.A. Tests the Power – and Learns the Limits – of His Office, N.Y. Times,
Oct. 30, 2018; Steve Volk, Philadelphia DA Larry Krasner On Radical Path to Remake Criminal Justice System,
Newsweek, Oct. 31, 2018; Steve Volk, Larry Krasner vs. Everybody: Inside the Philly DA’s Crusade to Revolutionize
Criminal Justice; Philadelphia Magazine, Nov. 23, 2019.
384
Joe Trinacria, Philly FOP President Blasts DA Krasner in Letter to Police Cadets, Philadelphia Magazine, March 2,
2018.
385
‘Help Wanted’: Philadelphia Police Union Calling For New District Attorney in I-95 Billboards, CBS Philly, June 27,
2019.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 172 • National Registry of Exonerations • September 1, 2020
In July 2019, the Pennsylvania state legislature passed a law that gives authority to the
state’s attorney general to prosecute some firearms violations in Philadelphia—normally
an exclusive power of the district attorney—but nowhere else in the state.
386
In August 2019, after a gunman injured several police officers in a shootout that lasted
seven hours, the United States Attorney for the federal district that includes Philadelphia
said the shooter was motivated by “the disrespect for law enforcement” that Krasner “is
fostering in this community.”
387
The critical question is whether Krasner will win reelection in 2021. If he’s defeated after one
term, his impact may be limited. In the meantime, however, in his first two years in office
Krasner hired 184 of the 318 attorneys who work in his office.
388
No big city police chief could
exercise that level of control over the composition of their professional workforce.
Like Larry Krasner, most of the current crop of progressive local prosecutors are in their first
terms in office. Historically, many local prosecutors are reelected repeatedly with support from
the local police and judiciary. These prosecutors have broken that mold. If they are able to stay
in office long enough, they may affect basic changes; if prosecutors in other jurisdictions follow
suit, those changes may become widespread.
That remains to be seen.
389
c. National Patterns
Whatever local crime lab directors, municipal police chiefs and county prosecutors do, their
direct influence, by definition, is local. If they are successful in their own territories, America
will remain a patchwork of counties and districts with widely divergent practices—some effective
at combating misconduct and preventing wrongful convictions, some not—unless fundamental
changes take place at a national level.
The United States Department of Justice (DOJ) has some capacity to monitor and address police
misconduct by local agencies. It can also lead by example, as it did in 2014, when it reversed
386
Akela Lacy & Ryan Grim, Pennsylvania Lawmakers Move to Strip Reformist Prosecutor Larry Krasner of Authority,
The Intercept, July 8, 2019.
387
Staff Reports, Philly Police Shooting Recap: Gunman Had AR-15 and Handgun, Officers Identified, McSwain-
Krasner Sniping, Philadelphia Inquirer, Aug. 15, 2019. As it happens, the gunman was a former federal gun-crime
defendant who received a reduced sentence and was released after multiple arrests because he was an informant for
the federal prosecutorial office headed by the United States Attorney who criticized Krasner. The standoff ended after
Krasner went to the scene and convinced the gunman to surrender safely. Alex Yablon, Larry Krasner’s Lonely,
Radical Crusade to Solve America’s Gun Problem, New Republic, Jan. 28, 2020.
388
Email to Samuel Gross from the Philadelphia District Attorney’s Office, March 13, 2020.
389
On August 4, 2020, Kim Gardner, the progressive Circuit Attorney of Saint Louis City, easily won her first
reelection primary (see, Joel Currier, Gardner easily wins primary for St. Louis circuit attorney, St. Louis Post
Dispatch, August 5, 2020), despite a host of attacks (see, e.g., Tom Jackman, 67 current, former prosecutors defend
St. Louis prosecutor from attacks in McCloskey gun case, The Washington Post, July 22, 2020), and is likely to win
the general election in November.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 173 • National Registry of Exonerations • September 1, 2020
course and began to require electronic recording of interrogations,
390
—and as it could do
pervasively by setting and maintaining high standards of conduct for its own agents and
prosecutors. In the past several years, the department has performed poorly on both fronts. That
could change.
At best, however, the role of the federal government will be limited. Any major national change
in the prevalence of misconduct in criminal investigations and prosecutions will require a
change in the national culture of the professions and institutions that conduct those functions.
This has happened before. It may happen again.
i. The United States Department of Justice
The federal Department of Justice is a unitary national institution. It includes 94 Offices of
United States Attorneys that handle federal prosecutions (and non-criminal matters) within
their local jurisdictions. These offices, however, unlike local state prosecutorial offices, are
headed by United States Attorneys who are appointed by the President and operate under the
authority of the Attorney General of the United States. The best unit of analysis for federal
criminal policy is the entire department.
The Department of Justice has very different roles in addressing misconduct by forensic
analysts, by police officers, and by prosecutors.
(i) For decades, DOJ has been a national leader in the use of forensic science.
(ii) A portion of the department’s case load is devoted to misconduct by local police officers,
both criminal prosecutions of officers who commit crimes in the course of pursuing particular
cases, and civil complaints against police departments for systematic violations of constitutional
rights. DOJ can also lead police reform by example, as we mentioned.
(iii) DOJ has no authority to systematically review or monitor the conduct of state
prosecutors,
391
but it could lead by setting an example of good practice for them to follow.
(a) Leading in forensic science.
The Federal Bureau of Investigation (FBI), the Department of Justice’s principal law
enforcement agency, administers the largest, best known, and most influential crime laboratory
in the United States. Over the past several decades, the FBI has also been the home to some of
390
Office of Pub. Affairs, “Attorney General Holder Announces Significant Policy Shift Concerning Electronic
Recording of Statements,” U.S. Dep’t of Justice, May 22, 2014. See supra Section XII.2.b.ii.a.
391
In theory, DOJ could charge individual prosecutors for violating 18 U.S. Code § 242, which makes it a crime to
“willfully subject[] any person … to the deprivation of any rights, privileges, or immunities secured or protected by the
Constitution or laws of the United States.” See United States Department of Justice, Statutes Enforced by the
Criminal Section. That would be appropriate, for example, if a prosecutor deliberately deprived a defendant of a fair
trial by concealing critical exculpatory evidence. But doing so would be difficult in exoneration cases—if DOJ were
otherwise interested—because, under 18 U.S. Code § 3282, charges must be filed “within five years next after such
offense shall have been committed.” (Emphasis added.) As we saw in Section X.2.a., Note 206, that is usually
impossible because the time lag from commission to discovery of misconduct in cases that produce exonerations is
typically much longer than five years.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 174 • National Registry of Exonerations • September 1, 2020
the country’s most prominent forensic scandals, and both the FBI and the DOJ have been
accused of unreasonable defensiveness of forensic science.
392
In 2009, the National Research
Council in its path breaking report on forensic science in the United States, explicitly
recommended against allowing DOJ to regulate forensic science because of its prosecutorial and
law enforcement orientation.
393
For several years after 2009 there were signs that the DOJ and the FBI might assume a leading
role in improving the use of forensic science, including preventing forensic fraud and other
misconduct. The FBI cooperated with professional organizations in a comprehensive review of
its troubled history of microscopic hair comparisons,
394
sponsored studies of the accuracy of
forensic techniques,
395
and initiated a procedure to prevent contextual bias from coloring the
judgment of forensic examiners.
396
In 2013—despite the National Research Council’s skepticism about allowing DOJ to regulate
forensic science—the department was designated as co-administrator of the newly created
National Commission on Forensic Science (NCFS), along with the National Institute of
Standards and Technology. The NCFS made many useful recommendations for regulating
forensic science.
397
They only bear directly on federal laboratories, but could begin to function as
national guidelines for state and local laboratories.
Unfortunately, the current presidential administration disbanded the National Commission on
Forensic Science in 2017,
398
and has shown no interest in forensic science reform in any other
context. But a start has been made. If a future administration decides to pursue the issue, there
is a foundation to build on.
392
John F. Kelly & Phillip K. Wearne, Tainting Evidence: Inside the Scandals at the FBI Crime Lab (1998); NRC,
Weighing Bullet Lead Evidence, (2004); Office of the Inspector General, A Review of the FBI's Handling of the
Brandon Mayfield Case, (Jan., 2006); ABS Group, Root and Cultural Cause Analysis of Report and Testimony Errors
by FBI MHCA Examiners, (Aug., 2018).
393
National Research Council Of The National Academies, Strengthening Forensic Science In The United States: A
Path Forward at 17 (Feb., 2009).
394
Norman L. Reimer, The Hair Microscopy Review Project: An Historic Breakthrough for Law Enforcement and A
Daunting Challenge for the Defense Bar, The Champion 16, July, (2013).
395
E.g., The President’s Council of Advisors on Science and Technology, Report to the President, Forensic Science
in Criminal Courts: Ensuring Scientific Validity of Feature-Comparison Methods, 9 (Sept., 2016).
396
Office of the Inspector General, A Review of the FBI’s Progress in Responding to the Recommendations in the
Office of the Inspector General Report on the Fingerprint Misidentification in the Brandon Mayfield Case, 5 (June,
2011).
397
National Commission on Forensic Science, Reflecting Back—Looking Toward the Future (April 11, 2017).
398
Spencer S. Hsu, Sessions Orders Justice Dept. to End Forensic Science Commission, Suspend Review Policy,
Washington Post, Apr. 10, 2017; Suzanne Bell et al., A Call for More Science in Forensic Science, 115 Proceedings
of the National Academy of Sciences 4541 (2018).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 175 • National Registry of Exonerations • September 1, 2020
(b) Policing local police
Chicago Police Sergeant Ronald Watts was arrested in 2012 for stealing money from an FBI
informant playing the role of a drug courier. The following year, he pled guilty in federal court
and was sentenced to 22 months in prison.
399
After his arrest, two Chicago police officers filed a
whistleblower suit in which they claimed that they had told their supervisors that Watts had
been shaking down drug dealers and framing innocent people for years, but their superiors
responded by calling them “rats” and assigning them to meaningless desk jobs.
400
Sixty-six
defendants who had been framed for drug crimes by Watts and his subordinates have been
exonerated, and many more are likely to be in the months and years to come.
401
The Watts cases constitute one of seventeen “group exonerations” that we know of in which
multiple defendants were exonerated after it was shown that corrupt police officers framed them
for drug crimes that did not occur.
402
Between them, these groups account for more than 2,500
exonerations. (The Watts cases are listed in the Registry; other group exonerations are not so
far, but more will be added, as we have discussed.
403
)
In nine of those groups, including every one with exonerations that began since 2004, local
authorities only acted after local police officers were indicted in federal court.
404
The Watts cases
are a good example: High-ranking Chicago police officers had known for years what Sergeant
Watts and his men were doing, but took no action until after the FBI arrested him; they still
have taken no disciplinary action against any of more than a dozen other officers involved, other
than Watts himself and one subordinate who was also convicted in federal court.
405
In Benton
Harbor, Michigan, and in Camden, New Jersey, local police initiated investigations of systematic
misconduct in their own departments—in Camden, it was the chief—but then asked the DOJ
rather than local prosecutors to file charges. Presumably, they expected federal authorities to be
more effective in prosecuting local police than county prosecutors.
In addition, since 1994, the Civil Rights Division of the Justice Department has had the
authority to file civil complaints against police forces that exhibit “a pattern or practice of
399
Supra Section III.3.c.i.
400
Jason Meisner, Former Chicago Police Officer Sentenced For Stealing Money from Drug Courier, Chicago
Tribune, Oct. 9, 2013.
401
See supra Section III.3.c.i.
402
National Registry of Exonerations, Mass Exonerations and Group Exonerations Since 1989 (Apr. 9, 2018).
403
See supra Section III.3.c.i.
404
See National Registry of Exonerations, Mass Exonerations and Group Exonerations Since 1989 (Apr. 9, 2018).
The groups in which federal indictments preceded any local exonerations are: Philadelphia PA, 1995-2008; Tulsa OK,
2009-2012; Benton Harbor MI, 2010-2012; Camden NJ, 2010-2012; Mansfield OH, 2012; Philadelphia PA, 2013-
ongoing; East Cleveland OH, 2016-2017; Chicago IL, 2016-ongoing; Baltimore, MD, 2017-ongoing. These criminal
charges do not face the statute-of-limitations issues that prevent the conviction of most prosecutors for misconduct in
cases that produce exonerations, See supra Note 391, because, when charges are filed, the long-term patterns of
criminal conduct by police that give rise to the charges are recent or still on-going.
405
See supra Section X.1.b.i.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 176 • National Registry of Exonerations • September 1, 2020
conduct by law enforcement officers” that violates the Constitution.
406
As of the beginning of
2017, 40 such cases had produced agreements requiring state and local police departments to
change their practices.
407
These agreements focus primarily on the issues that are central to most police reform efforts:
community relations, the use of force, racial discrimination in police stops and arrests, race
relations generally, and so on. But the DOJ could use this process to require local police to
improve evidence gathering practices. It did so in its 2013 consent decree with the City of New
Orleans, which includes more than a dozen specific rules for conducting interrogations (e.g.,
recording is required) and photographic lineups (the officers conducting a lineup may not know
who the suspect is, and so forth).
408
The United States Department of Justice has tools to address systematic violations of due
process by local police. Federal prosecutions can act as a backstop to local inaction in the face of
scandalous criminal behavior by police, and civil complaints can generate systemic reforms. The
department has used these tools over the past few decades with substantial success in dozens of
cities and counties, but it would take more resources and decisive action to drastically reduce
police misconduct on a national scale.
409
The administration of the current president has moved
in the opposite direction,
410
but that may change.
(c) Setting an example on prosecutorial conduct
Official misconduct in federal exonerations is severely lopsided.
411
In state cases, misconduct by
police officers is moderately more common than misconduct by prosecutors, 36% to 29%. For all
federal exonerations, it’s the reverse and by a large margin: misconduct by prosecutors is more
than two-and-a-half times as frequent as misconduct by police, 52% to 20%. Among federal
white-collar crimes—the most common type of federal exoneration—prosecutors committed
misconduct more than seven times as often as police, in 65% of the cases compared to 9%.
There are two sides to this striking pattern: a low rate of misconduct by police officers, and a
high rate by prosecutors. The rate of misconduct by prosecutors among exonerations for federal
406
42 U.S.C. § 14141.
407
Civil Rights Division, U.S. Department of Justice, The Civil Rights Division’s Pattern and Practice Police Reform
Work: 1994-Present (2017). Twenty of the agreements were court-enforced consent decrees, and 20 were settlement
agreements, typically known as memoranda of agreement, between the United States and the local jurisdiction.
408
United States v. City of New Orleans, Case 2:12-cv-01924-SM-JCW, United States District Court for the Eastern
District of Louisiana, Consent Decree, January 11, 2013, pp, 45-48.
409
See Joanna C. Schwartz, Who Can Police the Police?, 2016 U. Chi. Legal F. 437, 448; Rachel A. Harmon,
Promoting Civil Rights Through Proactive Policing Reform, 62 Stan. L. Rev. 1, 3-4 (2009).
410
G.R., Without Consent Decrees, Who Will Police the Police, Economist, Jan. 28, 2019, J. Brian Charles, Justice
Department Ends Era of Pushing Police Reform, Governing, Sept. 28, 2017.
411
See supra Section IX at Table 21.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 177 • National Registry of Exonerations • September 1, 2020
white-collar crimes is higher than for any other crime category, and the rate of misconduct by
police is lower than for any other crime group.
412
Federal prosecutors may be partly responsible for the low rate of police misconduct in federal
white-collar exonerations.
413
White-collar prosecutions are usually proactive cases that
prosecutors initiate and control. The investigators are overwhelmingly federal agents, most of
whom work for agencies of the Department of Justice. With that level of involvement,
prosecutors can discourage or prevent a lot of misconduct by police officers.
These numbers suggest a low rate of police misconduct for all federal white-collar cases, not just
the small number that produce exonerations. If the intense prosecutorial oversight that the DOJ
can exercise in big-ticket federal prosecutions prevents police misconduct, greater prosecutorial
involvement in police investigations might substantially reduce police misconduct in state law
criminal cases as well.
Prosecutorial misconduct, however, is an entirely different story. Federal prosecutors
themselves commit misconduct in nearly two-thirds of white-collar exonerations. Every one of
the handful of federal white-collar exonerations with police misconduct also included
prosecutorial misconduct, which suggests that the prosecutors could have prevented police
misconduct in many of those cases if they were not themselves part of the problem.
As always, we can’t use these data to estimate the overall rates of misconduct in federal white-
collar cases.
414
Federal prosecutors may prevent a great deal of police misconduct in white-collar
prosecutions—at least the types of misconduct that lead to false convictions—but misconduct by
federal prosecutors themselves is the rule in those cases that end in exoneration. The resources
at their disposal, and their ability to choose the targets for investigation, make the danger of
misconduct by prosecutors in federal white-collar crime cases particularly disturbing—especially
since the frequency of white-collar crime exonerations with misconduct by federal prosecutors
has more than doubled since 2002.
ii. National culture
(a) Questioning children
So far we’ve discussed the culture of individual offices and departments, or particular counties.
But culture also exists—and can change—at a national level. We saw that recently with improper
and frequently abusive questioning of the supposed victims of child sex abuse.
From the mid-1980s through the mid-1990s, this type of misconduct fueled a national epidemic
of prosecutions of child care workers and others based on false, bizarre and often impossible
claims of massive patterns of abuse that police and child welfare workers extracted from the
412
This comparison includes only crimes that have 10 or more exonerations. (We do not include Fraud on this list
because 32 of the 40 fraud exonerations we know about are federal white-collar cases; they constitute more than
70% of all federal white-collar exonerations.)
413
Supra Section III.3.d.
414
Supra Section III.3.d.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 178 • National Registry of Exonerations • September 1, 2020
children.
415
By 2003, the Chief of Police of Minnetonka, Minnesota, explained that child sex
abuse charges from 20 years earlier were being dismissed because questioning of children was
“less sophisticated in 1984” and produced unreliable evidence. There are stragglers—in a few
individual child sex abuse exonerations the children were subjected to improper questioning as
recently as 2014—but they are now rare exceptions.
(b) Forensic fraud
A similar change seems to be taking place with forensic fraud, at least in investigations of violent
crimes.
416
The number of known cases has decreased steeply, from 5% of exonerations for
convictions before 2002 to 0.4% for those since 2003. That might also be true for other types of
forensic misconduct; our data do not address that issue directly.
Forensic evidence is much more controversial than it was 30 years ago. One of the singular
effects of the DNA exonerations since 1989 has been to identify many errors and misstatements
by forensic analysts that contributed to convictions of innocent defendants. In some cases, the
problem was forensic fraud; in others there was misinformation but no misconduct. Either way,
they brought unprecedented scrutiny to conclusions by forensic experts that had been routinely
accepted without dispute.
One result of this new scrutiny was a surge in forensic lab scandals, uncovering systematic
patterns of incompetence, misconduct or both. By one estimate, there were at least 70 such
scandals from 1993 to 2013.
417
In the process, several serial forensic fraudsters were exposed,
including Fred Zain in West Virginia and Texas, Joyce Gilchrist in Oklahoma and Pamela Fish in
Chicago, who between them committed forensic fraud in 20 exonerations.
418
All three clusters
were based on fraud committed before 1998, and nearly three-quarters of the resulting
exonerations were complete before 2003 (17/25).
In 2009, the National Research Council’s report on forensic science in the United States set off
shock waves by concluding that many practices, tests and comparisons that were presented in
courts as “forensic science” had no scientific basis.
419
The message to practitioners on all sides was clear: Forensic evidence is powerful but
dangerous; it requires care and scrutiny. That means the work of forensic analysts is more likely
to be reviewed by supervisors in forensic labs, by police in their criminal investigations, by
prosecutors preparing for trial, and by defense attorneys and defense experts in court. Scrutiny
and care mean that fraud by forensic analysts (and other types of misconduct, and mistakes) is
415
See supra Section IV.4.
416
See supra Section VI.1.
417
Sandra Guerra Thompson, Cops in Lab Coats, Carolina Academic Press (2015) at 52-61.
418
See supra Section VI.1.
419
National Research Council of the National Academies, Strengthening Forensic Science In The United States: A
Path Forward (2009).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 179 • National Registry of Exonerations • September 1, 2020
less likely to happen, and more likely to be identified or weeded out by prosecutors, defense
attorneys or opposing experts.
The NRC report also recommended that forensic laboratories be run by forensic scientists rather
than police officers. So far, we’ve only seen baby steps in that direction, but if independent,
scientist-run labs become the rule, we expect the quality and integrity of forensic evidence to
continue to improve.
(c) Violence in interrogations
Perhaps the most important cultural transformations in criminal investigation in the United
States in the last hundred years is the huge reduction in the use of violence to extract
confessions that occurred in the middle decades of the twentieth century.
In 1931, the National Commission on Law Enforcement (the “Wickersham Commission”)
published its Report on Lawlessness in Law Enforcement,” which documented countless cases
of violence in interrogations and concluded that:
“The third degree—the inflicting of pain, physical or mental to extract confessions or
statements—is widespread throughout the country. …Physical brutality is extensively
practiced. The methods…range from beating to harsher forms of torture.”
420
The revelations of the Wickersham Commission began a process that unfolded over more than
35 years and included several components.
In 1936, the Supreme Court decided the first of many cases in which it recognized that the use in
court of a confession obtained by torture violates due process of law.
421
In 1944, the Court
unanimously agreed that “violence [in interrogations] is, an outlaw.”
422
About the same time,
Fred Inbau—a law professor, former police officer, and the most prominent expert on
interrogations in the second third of the twentieth century—warned police about “known
instances of miscarriages of justice resulting from the use of force and threats in obtaining
confessions,”
423
and began to try to teach police more “professional” non-violent interrogation
techniques.
424
Inbau’s efforts evolved into what is now known as the Reid Technique of
interrogation, a method that is based on isolation, deception and manipulation. This method is
often coercive. It has produced its fair share of false confessions—in the last 20 years, it has
become a major target of attempts to reduce the number of confessions by innocent
420
National Commission on Law Enforcement, Report on Lawlessness in Law Enforcement (1931) at p.153.
421
Brown v. Mississippi, 297 U.S. 278 (1936).
422
Ashcraft v. Tennessee, 322 U.S. 143 (1944).
423
Fred Inbau, The Courts on Confessions, The Police Digest, Dec. 1943, at 13, 15.
424
See F. Inbau, Lie Detection and Criminal Interrogation 119-33 (1st ed. 1942).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 180 • National Registry of Exonerations • September 1, 2020
suspects
425
—but it does so without violence, or (if properly applied) any other type of legally
recognized misconduct.
426
In 1966, the Supreme Court decided Miranda v. Arizona,
427
which famously requires the police
to provide warnings and an opportunity to consult with a lawyer before interrogating a suspect
in custody. Miranda established a formal legal structure for interrogations. It focused attention
on the dangers of the process—confessions that are coerced or false or both—and created a
procedural ritual to address them. It both built on and advanced the professionalization of
interrogations, and of police investigations generally, that Inbau began to promote more than 20
years earlier.
428
Other forces were at play as well. In particular, in 1963, in Gideon v. Wainwright,
429
the
Supreme Court greatly expanded the constitutional right to defense counsel in criminal
prosecutions. That made it harder to hide abusive interrogations. And the disproportionate use
of beatings and torture against African Americans and other minority suspects made them a
specific target of the civil rights movement in the 1950s and ‘60s. In any event, by 10 years after
Miranda, the “third degree” was largely an historical relic. In most police departments, it simply
was not done.
But not everywhere. The cultural shift that ended violence in interrogations did not reach every
police department, as we have discussed in detail. It took a much later change in the culture of
Cook County law enforcement—around the turn of the twenty-first century—to eliminate
systematic brutality in interrogations by Chicago police officers. The result is reflected in our
data: a huge decrease in the rate of violence in interrogations conducted after 2002.
430
There may be a few other jurisdictions where violence is still common in interrogations, but we
haven’t heard of them. There are, of course, exceptions that we do know about, but they’re just
that, exceptions.
Are other changes of this sort already underway? Maybe. We’ve identified several horses; we
don’t know which, if any, will make it around the track.
425
425
See Saul M. Kassin et al., Police-Induced Confessions: Risk Factors and Recommendations, 34 Law & Hum.
Behav. 13 (2010); Richard A. Leo, False Confessions: Causes, Consequences, and Implications, 37 J. Am. Acad.
Psychiatry & L. 332, 332–34 (2009).
426
See supra Section V.1.a.
427
Miranda v. Arizona, 384 U.S. 436 (1966).
428
Richard A. Leo, The Impact of Miranda Revisited, 86 J. Crim. L & Criminology 621, 668-69 (1996).
429
Gideon v. Wainwright, 372 U.S. 335 (1963).
430
See supra Section XI.3.
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 181 • National Registry of Exonerations • September 1, 2020
3. CODA
Over the past 15 to 20 years, we’ve seen declines in a few categories of official misconduct that
lead to false convictions: the use of violence and other abusive tactics in interrogations,
improper questioning of children, and forensic fraud.
There may have been a general decrease over the last two decades in misconduct that leads to
false convictions. It’s too early to say given the available data. We’re optimistic about the future,
but only time will tell.
On the other hand, over the same time period, there’s been a clear increase in false convictions
for federal white-collar crime in cases with prosecutorial misconduct.
Rules of conduct can reduce official misconduct in handling criminal cases, especially
affirmative rules that prescribe what should be done. The most valuable are rules that primarily
serve central goals of criminal investigation and prosecution: obtaining and recording accurate
information from witnesses and suspects; collecting and preserving probative physical evidence;
providing all relevant evidence to all parties. If followed, such rules will also prevent a great deal
of misconduct.
Rules, of course, are often ignored. They will be ignored in any organization that does not have
the resources to perform adequately, but even with adequate resources, rules must be enforced.
The most effective type of enforcement is expensive: ongoing supervision of those who work the
cases, with modest contemporaneous sanctions for poor performance or misconduct.
Regardless of cost, any attempt to change operational rules, or to institute this type of
supervision in an organization that does not have it, will be resisted by the established working
culture. Other things equal, resistance is likely to be stronger for police—whose profession
operates entirely within the criminal justice system—than for prosecutors, forensic scientists
and child welfare workers, whose training and professional colleagues are more likely to overlap
with other disciplines and other lines of work.
Changing a destructive work culture in an organization or a county requires determined and
effective leadership—and, of course, the resources to implement the change. Success is never
assured; ultimately, it may depend on local political support.
The United States has 3,142 counties, more than 2,300 separate prosecutorial offices and about
18,000 police forces. Local reforms alone will produce a patchwork of rules and practices at
best, and leave many places untouched.
The only effective national remedy to entrenched forms of official misconduct is a change in the
national culture that governs this work. That has happened for some types of misconduct in
criminal cases, in this century and in the last. It might happen again—slowly perhaps, or
gradually and then suddenly.
431
431
Ernest Hemingway, The Sun Also Rises (1926).
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 182 • National Registry of Exonerations • September 1, 2020
Methodological Appendix
1. The National Registry of Exonerations
Our data are exonerations posted on the National Registry of Exonerations, an online archive of
exonerations in the United States. The Registry classifies a case as an “exoneration” if a person
who was convicted of a crime is officially and completely cleared based on new evidence of
innocence. A more detailed definition appears here.
As we described in the body of this Report (Section II.2.a), the Registry relies entirely on non-
confidential, publicly-available sources, and it changes constantly. We add cases steadily—about
200 a year—and we change and correct old cases as we get new information, explore new issues,
or modify our coding scheme.
This report is based on the first 2,400 exonerations in the Registry, those posted by February 27,
2019. Each of those cases is reported in the Registry, in a separate story under the exoneree’s
name and in the tabulations of cases that appear on our Detailed View, our Summary View, our
Interactive Data Display, and in a variety or reports we release. Within that set of 2,400 cases,
the details of available data vary depending on the issue, as we describe below.
This report contains many links to the stories of individual exonerees whom we discuss, and
some data on them. The links take you to the current versions, which may have been updated
since the report was written. We also include links to sets of cases, such as all exonerations with
official misconduct in a particular county. Those groups may have grown to include cases posted
since the first 2,400.
2. Coding official misconduct
As we have discussed (Section II.2 and Section III.1) we only code as “official misconduct” acts
by law enforcement officials that undermine the factual accuracy of the determination that the
exoneree committed a crime. We don’t code as misconduct many bad practices that increase the
chance of errors—even incompetence—if the official involved did not violate a duty to behave
otherwise. And, of course, we can’t code the many acts of misconduct that remain hidden from
us, even in cases that have ended in exoneration.
3. The datasets
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 183 • National Registry of Exonerations • September 1, 2020
We began this study in 2014 with the first 1,361 cases that had been posted in the Registry, those
listed by May 14, 2014.
432
Over the following year-and-a-half we developed a detailed coding
system for official misconduct and applied it to those 1,361 cases.
In 2015 and 2016, based on what we learned from that initial dataset, we developed a shorter
and simpler (but still extensive) coding scheme for official misconduct that we have applied to
all exonerations in the Registry. We now use that coding system for official misconduct for all
exonerations in the Registry—the first 1,361, all 2,4oo that are in the main dataset for this
report, and all exonerations we have added since and continue to add every week.
Most of the data we report is based on the official misconduct coding that we now do for all
exonerations, and applies to all 2,400 cases in our main dataset. But the data we collect is
constrained by our resources. We code information by the case; we do not generally code
information separately for each category of misconduct or type of official who committed
misconduct. To do that would multiply the time required to process cases beyond our capacity.
As a result, we might be able to say, for example, that a particular case included misconduct by a
prosecutor, and also in the same case a law enforcement officer concealed exculpatory evidence,
but—unless the case included no misconduct by any other type of law enforcement officer—we
cannot say whether a prosecutor was the official who concealed that evidence.
However, in our initial study of the first 1,361 exonerations posted we do have information on
the categories of officials who committed particular types of misconduct. We also have more
detailed information on the types of exculpatory evidence that were withheld. We report some of
these data where appropriate in the text. In other words, some of our more detailed findings are
based on data from a subset of the first 1,361 exonerations posted among the entire set of 2,400
exonerations we use.
Any table that is based on the 1,361 exoneration dataset includes the notation “N = 1,361.” In the
text, we often include the numerators and denominators of proportions we discuss as
percentages. For example, we might describe a hypothetical proportion as based on the main
dataset as “25% (600/2,400).” If we had the same proportion for the 1,361 data set the
description would be similar, “25% (340/1,361),” but note that the actual proportion—
340/1,361—is italicized.
The same applies to subcategories. For example, if we reported that some fact applies to half of
all cases with official misconduct in the main data set, the denominator would be 1,296, the
number of cases out of 2,400 that include misconduct, and the description might be “50%
(648/1,296).” If we reported the same ratio for the 1,316 dataset the denominator would be 768
(the number of cases with misconduct in that dataset), and we would write “50% (384/768).
Note the italics.
3. Estimating the rate of fabricating evidence
432
To be precise, we started in May 2014 with 1,365 cases, but over the succeeding years four of those cases were
removed because we determined that they did not in fact qualify under our definition of “exoneration.”
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 184 • National Registry of Exonerations • September 1, 2020
“Fabricating Evidence” is a set of four different types of misconduct by officials who, in one form
or another, made up evidence that they falsely claimed to have witnessed themselves: forensic
fraud, planted evidence, phony assaults, and fabricated confessions.
We have data on each of the components of fabricating evidence, but they come from both of the
two data sets we discussed—the 1,361 cases posted by May 14, 2014, and the larger group of all
2,400 cases posted by February 27, 2019.
To get an overall estimate of the rate of fabricated evidence, we (i) determine the partial rate for
each component, in whichever dataset we can; (ii) eliminate any duplicates, cases that appear to
include more than one sub-species of fabricating evidence; and (iii) combine the resulting rates
after duplications are removed.
(i) Determining partial rates
• The rate of forensic fraud is determined from the 2,400 data set: 75/2,400 = 3.1%
• The rate of planted evidence is a combination of the rates for non-group exonerations and
the rate for group exonerations. For non-group exonerations we have data from the 1,361
data set: 16/1,361 = 1.2%. For group exonerations, we have data from the 2,400 dataset,
66/2,400 = 2.8%. No group exonerations are in 1,361 dataset, so there are no duplications,
and we can combine these rates to estimate the overall rate of planting evidence among all
exoneration in the study: 1.2% + 2.8% = 4%.
• The rate of fake assaults is a combination of the rates for non-misdemeanor exonerations
and the rate for misdemeanor exonerations. For non-misdemeanor exonerations we have
data from the 1,361 data set: 7/1,361 = 0.51%. For misdemeanor exonerations, we have data
from the 2,400 dataset, 11/2,400 = 0.47%. All but one of the misdemeanor cases were
posted after 1,361 dataset, so for this estimate we will treat the two sets as having no
duplications and combine these rates to estimate the overall rate of fake assaults among all
exonerations in the study: 0.51% + 0.47% = 0.98% = ~1.0%.
• The rate of fabricated confessions is determined from the 2,400 data set: 36/2,400 = 1.5%
(ii) Eliminating duplicates
There are only 3 cases that have more than one of these four types of misconduct: two cases
with fabricated confessions also had planted evidence, and one case with a fabricated
confession included forensic fraud. For the purpose of this estimate, we removed those three
cases from the total of cases with fabricated confessions, which, as adjusted, becomes:
33/2,400 = 1.4%
Go to Executive Summary • Go to Table of Contents
Government Misconduct and Convicting the Innocent
The Role of Prosecutors, Police and Other Law Enforcement
Page 185 • National Registry of Exonerations • September 1, 2020
(iii) Combining partial rates:
Forensic fraud 3.1%
Planted evidence 4.0%
Fake assaults 1.o%
Fabricated confessions 1.4%
(adjusted) ______
9.5%
The procedure for determining the rate of fabricating evidence for particular crimes is similar,
with data on each crime or set of crimes.
