Cole: Forensic Science and Wrongful Convictions
Cole: Forensic Science and Wrongful Convictions
Cole: Forensic Science and Wrongful Convictions
SIMON A. COLE*
INTRODUCTION
B
randon Garretts book, Convicting the Innocent, makes a number of
important contributions to the scholarly and public discourse on
miscarriages of justice. In this essay, I will focus on the contribution
that is most related to my own research interests: its contribution to our
understanding of the relationship between forensic science and
miscarriages of justice. I will first endeavor to place Garretts contribution
in historical context by briefly tracing the history of discussions about
forensic science and wrongful convictions. I will then highlight in what
way Garretts work has furthered our understanding. I will then discuss
some of the criticisms of Garretts work by advocates of forensic science
and try to explain how data limitations contribute to the difference of
opinion between Garrett and his critics. I will conclude by suggesting a
different, more theoretically grounded way of conceptualizing
miscarriages of justice that might help us move beyond these differences of
opinion. Ultimately, however, my suggestions will be highly speculative:
data limitations, again, will make it difficult to make any strong empirical
inferences about the relationship between forensic science and wrongful
convictions.
Ph.D. (Science and Technology Studies), Cornell University; A.B., Princeton University. I am
grateful to Jolle Vuille and Norah Rudin for helpful comments on drafts of this Article. This
material is partially based upon work supported by the National Science Foundation under
grant No. SES-0115305. Any opinions, findings, and conclusions or recommendations
expressed in this material are those of the author and do not necessarily reflect the views of
the National Science Foundation.
711
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
1 See, e.g., Laurence H. Tribe, Trial by Mathematics: Precision and Ritual in the Legal Process, 84
HARV. L. REV. 1329, 1333-34 (1971). Tribes characterization of forensic science is perhaps
unjust. See, e.g., D. H. Kaye, Revisiting Dreyfus: A More Complete Account of a Trial by
Mathematics, 91 MINN. L. REV. 825, 827 (2007); Franco Taroni et al., Forerunners of Bayesianism in
Early Forensic Science, 38 JURIMETRICS J. 183, 200 (1998).
2 PAUL ROBERTS & CHRIS WILLMORE, THE ROLE OF FORENSIC SCIENCE EVIDENCE IN CRIMINAL
PROCEEDINGS 1 (1993).
3 EDWIN M. BORCHARD, CONVICTING THE INNOCENT: ERRORS OF CRIMINAL JUSTICE, at xix
(1932).
4
See, e.g., Simon Dinitz, Foreword to C. RONALD HUFF ET AL., CONVICTED BUT INNOCENT:
WRONGFUL CONVICTIONS AND PUBLIC POLICY, at xiv-xv (1996); William O. Douglas, Foreword
to JEROME FRANK & BARBARA FRANK, NOT GUILTY 11-12 (1957); EDWARD D. RADIN, THE
INNOCENTS 231-35 (1964); Talia Roitberg Harmon, Predictors of Miscarriages of Justice in Capital
Cases, 18 JUST. Q. 949, 950 (2001). But see MICHAEL L. RADELET ET AL., IN SPITE OF INNOCENCE:
ERRONEOUS CONVICTIONS IN CAPITAL CASES 148 (1992) (discussing a case where the stray hair
of an unknown person was bottled along with the specimen collected during an autopsy
unbeknownst to the forensic expertcontributing to a miscarriage of justice).
5 Kent Roach, Forensic Science and Miscarriages of Justice: Some Lessons from Comparative
Ethics, Economics, and Traditional Ecological Knowledge, 21 VT. L. REV. 225, 267 (1996).
9 See Understand the Causes: Eyewitness Misidentification, THE INNOCENCE PROJECT,
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
Debate Over Capital Punishment in the United States, 34 LAW & SOC. INQUIRY 603, 617 (2009).
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
16 See Barry C. Scheck, Barry Scheck Lectures on Wrongful Convictions, 54 DRAKE L. REV. 587,
604-06 (2006).
17
See EDWARD CONNORS ET AL., U.S. DEPT OF JUSTICE, CONVICTED BY JURIES, EXONERATED
BY SCIENCE: CASE STUDIES IN THE USE OF DNA EVIDENCE TO ESTABLISH INNOCENCE AFTER
TRIAL 2 (1996), available at https://www.ncjrs.gov/pdffiles/dnaevid.pdf.
18 See Clive Walker & Russell Stockdale, Forensic Evidence, in MISCARRIAGES OF JUSTICE: A
REVIEW OF JUSTICE IN ERROR 119, 126-27 (Clive Walker & Keir Starmer eds., 1999).
19 CONNORS ET AL., supra note 17, at 15.
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
20 See RICHARD NOBLES & DAVID SCHIFF, UNDERSTANDING MISCARRIAGES OF JUSTICE: LAW,
THE MEDIA, AND THE INEVITABILITY OF CRISIS 117-22 (2000).
21 Id. at 189. Clive Walker & Carol McCartney, Criminal Justice and Miscarriages of Justice in
High Profile Criminal Appeals, 22 OXFORD J. LEGAL STUD. 53, 61 & n.43, 62 (2002).
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
23 Id. at 64.
24 Roach, supra note 5, at 72.
25 Id.
26
See generally BARRY SCHECK ET AL., ACTUAL INNOCENCE: FIVE DAYS TO EXECUTION AND
OTHER DISPATCHES FROM THE WRONGLY CONVICTED (2000).
27 Id. at 107, 158.
28 See id. at 170-71. Fred Salem Zain was a West Virginia state trooper who was in charge
of serology for the states crime laboratory. Id. at 109.
29 Id. at 158-71.
30 Id.
31 See id. at 1-33.
32 SCHECK ET AL., supra note 26, at app. 2.
33 See id.
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
Michael Saks and Jonathan Koehler cited the role of forensic science as
a contributor to miscarriages of justice to support their claim that forensic
science was not as reliable as it was often claimed to be and reform of
forensic science was urgently needed.34 Saks and Koehler suspected that
Scheck et al.s splitting of forensic science into multiple categories might
obscure the significant role that forensic science plays as a contributor to
miscarriages of justice. As a result, they published a slight reanalysis of the
Innocence Project datawhich at that time represented the first eighty-six
post-conviction DNA exonerationsby aggregating all forensic
contributors into just two categories.35 Errors in forensic science testing
were found to be present in sixty-three percent of cases, second only to
eyewitnesses misidentifications. The giving of false or misleading
testimony by forensic scientists was found to be present in twenty-seven
percent of cases, the fifth most common contributor.36 The articles
placement in the prestigious journal Science helped ignite renewed efforts
to reform forensic science in the United States. Some forensic scientists,
upset by the articles portrayal of their field, questioned the data sampling
techniques, methods, [and] criteria that went into Saks and Koehlers
representation of the forensic sciences as a leading contributor to
miscarriages of justice.37 Saks and Koehler responded that [r]esearch on
DNA exonerations is obviously in its infancy, and we support calls for a
more complete and scientific review of these cases.38
Of course, post-conviction DNA testing is only one method of exposing
miscarriages of justice. In a comprehensive study of U.S. exonerations from
1989, the beginning of the post-conviction DNA exoneration era, to 2003,
Gross and colleagues found a total of 340 exonerations, slightly less than
half of which were exposed by post-conviction DNA testing.39 Gross and
his colleagues devised their own system for categorizing the causes of
these wrongful convictions; forensic science was lumped into their
perjury category. The study found that twenty-four of the wrongful
convictions involved perjury by a forensic scientist.40 This analytic
34 See Michael J. Saks & Jonathan J. Koehler, The Coming Paradigm Shift in Forensic
09 (2006).
39 Post-conviction DNA testing exposed a total of 144 exonerations. Samuel R. Gross et al.,
Exonerations in the United States, 1989 through 2003, 95 J. CRIM. L. & CRIMINOLOGY 523, 523-24
(2005).
40 Id. at 543.
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
accurate results.57 The best term for this is probably validitythe term
Garrett has reserved for the other category of forensic problemssince
validity is generally understood to include reliability but not vice versa.
Specifically, a process cannot be more valid than it is reliableit cannot
produce accurate results at a greater rate than it produces consistent
results. However, a process can be more reliable than it is validit can
produce consistent results at a greater rate than it produces accurate
results.
Meanwhile, what Garrett calls validity, appears to refer primarily to
making correct inferences from the evidence. As Garrett himself notes:
[T]he term validity refers to whether claims or inferences are supported by
the evidence.58 However, this is the meaning of the term validity as
used in logic rather than the meaning used in science. Thus, it seems to me,
the distinction that Garrett is trying to make would have been clearer had
he chosen different labels for the two recurring types of problems with . . .
forensics.59 For example, he might have labeled them: (1) problems of
validity and (2) problems of inference.
This terminological confusion is exacerbated by some problems with
the integrity of the categories themselves. Garrett views only two forensic
techniques as reliable: serology and DNA profiling.60 This appears to be
because the techniques are not based on a subjective opinion and
because expert witnesses are able to defensibly convey the probative value
of serology or DNA profiling evidence by estimating the rarity of the
biological features found consistent with the suspect.61 However, Garretts
argument is slightly misleading. The issue is not the subjectivity of the
opinion. After all, even opinions about DNA profiles are in some sense
subjective.62 Rather the issue is whether the opinion is based on an
instrumental measurement or a visual analysis. A conclusion that serology
and DNA profiling are reliable ought to be based on: (1) studies that
measure the rate at which users of the technique achieve correct results on
samples of known origin, and (2) as Garrett correctly notes, expert
witnesses ability to provide defensible estimates of the rarity of the
biological features in question, so as to convey the probative value of the
evidence to the fact-finder.63
64 Id.
65 Id.
66 See id.
67 Id. at 86.
68 See COMM. OF IDENTIFYING THE NEEDS OF THE FORENSIC SCI. CMTY., NATL RESEARCH
COUNCIL, STRENGTHENING FORENSIC SCIENCE IN THE UNITED STATES: A PATH FORWARD 22
(2009), available at https://www.ncjrs.gov/pdffiles1/nij/grants/228091.pdf.
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
69 Id. at 2.
70 Compare id. at 161, with GARRETT, supra note 7, at 86.
71 GARRETT, supra note 7, at 90.
72 See Dawn McQuiston-Surrett & Michael J. Saks, Communicating Opinion Evidence in the
Forensic Identification Sciences: Accuracy and Impact, 59 HASTINGS L.J. 1159, 1188-89 (2008)
(*T+he language employed by forensic experts affects the inferences fact finders draw,
sometimes producing conclusions in the minds of fact finders quite different from what the
expert witnesses purportedly intend.); Dawn McQuiston-Surrett & Michael J. Saks, The
Testimony of Forensic Identification Science: What Expert Witnesses Say and What Factfinders Hear,
33 LAW & HUM. BEHAV. 436, 436 (2009) (discussing how variations in the presentation of
forensic science information affect fact-finders' judgments in a trial).
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
sixty-one percent of these 153 trials the testimony was invalid.73 The
analysis of this invalid testimony forms Garretts principal contribution
to the discussion of the role forensic science plays in miscarriages of justice.
Elsewhere, Garrett analyzed the invalid testimony in greater detail. He
defined six different categories of invalid testimony: (1) interpreting the
nonprobative evidence as inculpatory; (2) discounting exculpatory
evidence; (3) presenting inaccurate statistics; (4) providing frequencies or
probabilities in the absence of empirical data; (5) providing nonnumerical
statements of probability or frequency despite a lack of any empirical data;
(6) concluding evidence did in fact come from the defendant despite no
empirical data permitting such conclusions.74 Garretts data, therefore,
presents a very complex picture. The forensic evidence he discusses varies
along at least four important dimensions. We have forensic evidence that is
inculpatory, exculpatory, and inconclusive, that reports results from at
least seven different forensic techniques, resulting in trial testimony that is
either valid or invalid in one or more of six different ways, over a twenty-
four-year period. Making generalizations from such data is challenging to
say the least.
Next, we might ask what this data tells us about the role of forensic
science in miscarriages of justice that were exposed by post-conviction DNA
exonerations. The answer, clearly, is that forensic science played an
important role. However, this derives less from simply counting the
proportion of cases in which forensic evidence was present, or even the
proportion of cases in which forensic evidence was inculpatory, or even the
proportion of cases in which the evidence derived from unreliable
techniques or the testimony was invalid. A closer look at Garretts data
clearly shows that forensic science primarily contributed to the
miscarriages of justice that were exposed by post-conviction DNA
exonerations in two ways.
First, serological evidence which ought to have been interpreted as
either excluding the defendant or as having nothing useful to contribute to
the fact-finders perception of the defendants guilt was instead presented
to the fact-finder as inculpatory. This occurred in sixty-seven cases.79
Second, microscopic hair comparison evidence that ought, if used at all, to
have been conveyed to the fact-finder only as failing to exclude the
76 See DANIEL S. MEDWED, PROSECUTION COMPLEX: AMERICA'S RACE TO CONVICT AND ITS
80 Itiel E. Dror et al., Contextual Information Renders Experts Vulnerable to Making Erroneous
Identifications, 156 FORENSIC SCI. INTL 74, 75 (2006); Dan E. Krane et al., Sequential Unmasking:
A Means of Minimizing Observer Effects in Forensic DNA Interpretation, 53 J. FORENSIC SCI. 1006,
1006 (2008); D. Michael Risinger et al., The Daubert/Kumho Implications of Observer Effects in
Forensic Science: Hidden Problems of Expectation and Suggestion, 90 CAL. L. REV. 3, 9, 11-12, 30, 41-
42 (2002).
81 GARRETT, supra note 7, at 90.
82 Id.
83 Samuel R. Gross, Convicting the Innocent, 4 ANN. REV. L. SOC. SCI. 173, 176 (2008) (And
yet almost everything we know about false convictions in the United States depends on this
small, assorted, messy data set.).
84 See Justin Brooks & Alexander Simpson, Blood Sugar Sex Magik: A Review of Post-
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
evidence probative of guilt must have been present at the crime scene; that
this evidence was collected and preserved; that it was either not subjected
to DNA testing, subjected to an older, less discriminating method of
analysis, or erroneously interpreted; that the evidence remained available
for testing through post-conviction litigation; and that the crime was
serious enough that it yielded a sentence severe enough that the convict
had an incentive to pursue post-conviction DNA testing.85 In practice, this
means that sexual assault cases are highly over-represented in this data set,
that homicides are over-represented as well, and so are the combination of
the two: rape-murders.86 Thus, Garretts work allows us to conclude that
biased serology and microscopic hair comparison evidence were major
contributors to those miscarriages of justice that occurred in this selective
set of cases.
However, it is clear that there are other miscarriages of justice less
amenable to exposure through post-conviction DNA testing.87 For example,
as I have shown elsewhere, wrongful convictions involving fingerprint
evidence are less amenable to exposure through post-conviction DNA
testing than wrongful convictions involving serology or hair.88
What can we conclude about these cases from Garretts study? The
answer to this question is less clear. One perspective would be to treat
Garretts set of cases as deviantto assume that Garrett has identified a
problematic set of cases which are susceptible to specific problems. For
example, sexual assault prosecutions are especially problematic because
they are crimes that provoke high emotions and with specific evidence
collection problems (e.g., eyewitnesses who are not the victim or
perpetrator are rare). Or, perhaps serology and microscopic hair
Conviction DNA Testing Statutes and Legislative Recommendations, 59 DRAKE L. REV. 799, 805
(2011).
85 See Nicole Dapcic, A Quest for Exculpatory DNA Evidence or a Wild-Goose Chase? Expansion
of Searches for Lost Evidence Under Horton v. Maryland, 37 NEW ENG. J. ON CRIM. & CIV.
CONFINEMENT 77, 83-84 (2011).
86 Gross, supra note 83, at 179 (Once we move beyond murder and rape cases, we know
very little about any aspect of false convictions.); Alexandra Natapoff, Misdemeanors, 85 SO.
CAL. L. REV. (forthcoming 2012) (manuscript at 46) (*T+he innocence movement is centrally
concerned with serious offenses, typically murder and rape, which together comprise a small
fraction of the criminal systemaround 2 percent.); D. Michael Risinger, Innocents
Convicted: An Empirically Justified Factual Wrongful Conviction Rate, 97 J. CRIM. L. &
CRIMINOLOGY 761, 772 (2007).
87 See Gross, supra note 83, at 174-75 (We cannot study an event if we cant tell when it
happens. This is a severe problem for false convictions because, by definition, we dont know
when they occur.).
88 Simon A. Cole, More Than Zero: Accounting for Error in Latent Fingerprint Identification, 95
The difference between these two perspectives has been brought into
relief by a critique of Garretts work by two crime laboratory advocates.
Collins and Jarvis contend that the usual methods of analyzing data
derived from post-conviction DNA exonerations, which involve counting
the proportion of cases in which forensic science was involved, overstates
the contribution of forensic science to miscarriages of justice.90 They argue
that the presence of forensic science in a case does not necessarily imply
that forensic science caused the miscarriage of justice.91
This critique certainly has some merit for early analyses which simply
counted whether forensic science was present in post-conviction DNA
exonerations. But Garrett has improved on this method by counting the
number of cases in which forensic science was not merely present but
unreliable or invalid. However, Collins and Jarvis still contend that
Garrett overstates the contribution of forensic science. In their own analysis
92 Id. at 25.
93 Id.
94 Their coining of this term is criticized in Norah Rudin & Keith Inman, Who Speaks for
Forensic Science?: The Conviction and Exoneration of a Straw Man, CAL. ASS'N OF CRIMINALISTS
NEWS, Fourth Quarter 2008, at 10, 11.
95 Collins & Jarvis, supra note 90, at 24.
96John Collins & Jay Jarvis, The Wrongful Conviction of Forensic Science, CAL. ASSN OF
CRIMINALISTS NEWS, Fourth Quarter 2008, at 15, 21.
97 On January 26, 2012, I wrote to the authors through their website, CRIME LAB REPORT,
the crime, would have a forty percent chance of not being excluded.
Suppose that the same defendant is subsequently convicted, based in
part on the serological evidence. Suppose further that DNA typing is
subsequently developed, the convict requests post-conviction DNA typing,
the test is performed, and it excludes the defendant. DNA profiling is far
more discriminating than blood type. The temptation is to say that the
serological evidence was wrong or false or flawed, but, of course,
this is somewhat unfair. The serologist never said that the suspect was the
perpetrator. True, the serological evidence was used against a defendant
who turned out to be innocent. However, the serologist presumably
allowed for the fact that the suspect may have been among the many
innocent people that a blood-type analysis failed to exclude. The same
argument may be mounted in defense of hair comparison. Microscopic hair
comparison is not considered to be very discriminating.98 Thus, it should
not be surprising that many suspects might be included by microscopic hair
comparison, but excluded by DNA profiling.99
This is a sound argument if the expert witness properly conveys the
probative value of the evidence to the fact-finder. In serology, that would
mean conveying not only that the analysis failed to exclude the defendant,
but also the proportion of innocent persons the test would also be expected
to fail to exclude. For microscopic hair comparison, the problem would be
trickier because responsible estimates of the population frequencies of the
attributes analysts rely upon are not available. Therefore, the analyst
would be forced to employ vague verbal characterizations of the
discriminating power of the observed features. The testimony would be
something like: These attributes of the crime scene hairs were shared by
the sample hair from the suspect; however, many people in the population
also share these attributes.
However, Garretts findings show that the forensic expert witnesses
did not properly convey the probative value of the evidence to the fact-
finder.100 Rather, as noted above, Garrett shows that most of the serology
cases were not simply failures to exclude but cases in which masking
problems meant that the evidence should have been reported as having no
probative value whatsoeverthat the test failed to exclude one-hundred
98 See Walter F. Rowe, The Current Status of Microscopial Hair Comparisons, 2001 SCI. WORLD
868, 876.
99 See Max M. Houck & Bruce Budowle, Correlation of Microscopic and Mitochondrial DNA
Hair Comparisons, 47 J. FORENSIC SCI. 1, 964, 966 (2002); Max M. Houck, Letter to the Editor,
Forensic Science, No Consensus, 20 ISSUES SCI. & TECH. 6, 6-7 (2004); D. Michael Risinger &
Michael J. Saks, A House with No Foundation, 20 ISSUES SCI. & TECH. 35, 35 (2003); Michael J.
Saks & D. Michael Risinger, Baserates, the Presumption of Guilt, Admissibility Rulings, and
Erroneous Convictions, 2003 MICH. ST. L. REV. 1051, 1057.
100 See GARRETT, supra note 7, at 92-93.
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
percent of the population.101 Or, put another way, that there was a one-
hundred percent probability that the test would fail to exclude an innocent
person. In an additional set of cases, exculpatory evidence was reported as
either non-probative or inculpatory.102 Similarly, Garrett shows that in
many of the hair comparison cases, the probative value of the evidence was
exaggerated.103 It is misleading to categorize cases in which the probative
value of the evidence was skewed against the defendant as [n]on-specific
science failed to exclude the defendant.104
The above cases probably account for the bulk of the difference
between Garretts and Collins and Jarviss counts of the numbers of
wrongful convictions attributable to forensic science. Collins and Jarvis
also exclude cases in which the forensic evidence proffered against the
defendant derived from disciplines . . . rarely, if ever, practiced in
Americas crime laboratories.105 These include voice-print analysis and
dog-scent tracking.106 In my opinion, these analyses fit better in a category
called forensic science than in any other commonly used category of
evidence.107 Certainly, in the categorization typically used for the analysis
of wrongful convictions, dog scent fits better with forensic science than
with any of the other categories. But, in any event, the number of these
cases is small, so this difference in classification accounts for little of the
discrepancy between Garretts numbers and those of Collins and Jarvis.
108 See Collins & Jarvis, The Wrongful Conviction of Forensic Science, CRIME LAB REP. 12 (July
instead, that one can only understand forensic science in the context of
its use in law, and one can only understand lawyering in the context of
the evidence that lawyers use to support proof. The disparity between
Garretts position and that of Collins and Jarvis, therefore, like the earlier
dispute between Saks and Koehler and defenders of forensic science,113
illustrates the limitations of blaming discrete causes like eyewitness
identification, false confession, prosecutorial misconduct, bad
lawyering, and forensic science for miscarriages of justice.114 Further,
this dispute illustrates the limitations of trying to quantify the impact of
forensic science on wrongful convictions relative to other causes.115 As
Professor Natapoff notes: This posture assumes that if those discrete
pieces of evidence were stronger, the convictions would be sound. . . .
Rather than focusing on discrete pieces of evidence such as confessions or
fingerprints, innocence skepticism should be aimed at the entire procedural
apparatus.116
To be clear, I am not questioning the value of what we might call the
categorizing project. This project has certainly succeeded in giving us a
rough sense of what the contributors to miscarriages of justice are, and the
data is sufficient to warrant the inference that our criminal justice system
has weaknesses in each of these areas. It is understandable that authors of
books about miscarriages of justice find that the standard categories offer a
convenient way of organizing the material. So, certainly, this categorical
way of thinking about miscarriages of justice has its uses. But even those
who find it useful presumably understand that it is, ultimately, a
simplification.117 It is well understood, moreover, that data limitations
make drawing further inferences about the relative role of each of these
areas or their general accuracy problematic. In real cases, these discrete
causes interact with one another in complex and contradictory ways.118
Finding patterns in these interactions is far more challenging than counting
up cases in which a particular cause was present or contributed to a
113 Beatrice Schiffer, The Relationship Between Forensic Science and Judicial Error: A
Study Covering Error Sources, Bias, and Remedies 124 (2009) (unpublished Ph.D. dissertation,
University of Lausanne) (on file with the New England Law Review), available at
http://www.unil.ch/webdav/site/esc/shared/These_Schiffer.pdf.
114 See id. at 132.
115See id. at 32, 132 (studying the influence of various causes on wrongful convictions and
then concluding that such influence cannot be quantified or empirically determined).
116 Natapoff, supra note 86, (manuscript at 49).
117 See Richard A. Leo, Rethinking the Study of Miscarriages of Justice: Developing a
Criminology of Wrongful Conviction, 21 J. CONTEMP. CRIM. JUST. 201, 213 (2005).
118See George Castelle & Elizabeth F. Loftus, Misinformation and Wrongful Convictions, in
WRONGLY CONVICTED: PERSPECTIVES ON FAILED JUSTICE 17, 30-31 (Saundra D. Westervelt &
John A. Humphrey eds., 2001).
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
miscarriage of justice.
CONCLUSION
123Well, we do not actually know, but we have very strong reason for believing, and, even
if we are wrong about a handful of cases, it is extremely unlikely that we are wrong about
many of the 250 post-conviction DNA exonerations.
124 See supra note 81 and accompanying text.
125Should be may be more appropriate than is because most crime laboratories in the
United States are part of law enforcement agenciesdespite the recent recommendation of a
National Research Council committee that crime laboratories be independentand this is
unlikely to change. See COMM. OF IDENTIFYING THE NEEDS OF THE FORENSIC SCI. CMTY., NATL
RESEARCH COUNCIL, supra note 68, at 23.
COLE - FINAL (DO NOT DELETE) 6/2/2012 3:56 PM
126Peter Neufeld & Barry C. Scheck, Commentary in EDWARD CONNORS ET AL., CONVICTED
BY JURIES,EXONERATED BY SCIENCE: CASE STUDIES IN THE USE OF DNA EVIDENCE TO ESTABLISH
INNOCENCE AFTER TRIAL, at xxviii (1996) (noting that out of 8000 sexual assault cases with
conclusive DNA test results, approximately 2000 excluded the primary suspect).