Archive for the ‘Eyewitness Identification’ Category

On May 28, 2014, the Pennsylvania Supreme Court published two opinions about whether juries could hear from experts.  In Commonwealth v. Walker, the Court said an expert could tell the jury about the traps and weaknesses involved in eyewitness witness testimony.   But in Commonwealth v. Alicia, the Court  said an expert could not tell the jury about problems with the truthfulness of confessions.   Each opinion featured a different 4-2 vote among six of the Court’s Justices.

This juxtaposition seems so bizarre that the press in our state has noticed.  How does one court, on one day, publish two opinions on virtually the same issue, and end up facing in two completely opposite directions?

The explanation that jumps out from the opinions is the recognition of research and science.    In the Walker majority opinion, Justice Debra Todd discusses how much scientific work on eyewitness testimony has taken place in the last twenty years, since the Court last examined the issue.  In those decades, she says, “advances in scientific study have strongly suggested that eyewitnesses are apt to erroneously identify a person…when certain factors are present, ” and she cites the 2011 opinion of the New Jersey Supreme Court in New Jersey v. Henderson, widely regarded as one of the leading cases on the subject.  The scientific consensus is strong:  “it is beyond serious contention that the statistical evidence on eyewitness inaccuracy is substantial…”  Over the course of the last thirty years, forty-four states have recognized this scientific work and decided that trial courts may allow experts to testify about the problems of eyewitness identification.  Ten of the eleven U.S. Courts of Appeals have done so as well.  (False eyewitness identifications have occurred in almost seventy-five percent of all DNA-based reversals of wrongful convictions.)  Thus Justice Todd and the other Justices in the majority in Walker said that Pennsylvania must recognize the evidence that science has brought to bear.

In contrast, what does the Alicia opinion say about the science on the interrogation of suspects, and how various interrogation techniques may lead to false confessions? (False confessions have occurred in about twenty-five percent of all DNA-based reversals of wrongful convictions.)   Considerable scientific consensus exists on these issues as well; the leading scientists who work on the issue of false confessions published a standard-setting white paper, “Police-Induced Confessions: Risk Factors and Recommendation”  for the journal Law and Human Behavior five years ago. Somehow, this scientific work never even comes up in the Alicia opinion.  It is as if the research on false identifications simply does not exist.

Chief Justice Ronald Castille is at least consistent in these two cases: he wants no experts in court on either eyewitness identifications or confessions, so he dissented in Walker and joined the majority in Alicia.  Castille’s dissenting opinion in the Walker case gives us some insight into the thinking of those who want none of this science before the jury.  In his dissent, Chief Justice Castille scorns the very idea that there could be real scientific work on these subjects.  Indeed, he puts the words science or scientific in derisive quotation marks eight times in just the first three pages of his opinion.  While he admits that “a properly trained chemist can usually reliably explain the chemical composition of a substance,” science about human beings cannot claim any reliability: “studies of human beings, human nature, human perception, and human recollection inevitably have a heavy dose of subjectivity.”   Justice Castille’s position betrays a strong misunderstanding of what science is, and of the process of doing scientific work and of using the scientific method to test a hypothesis.  The real question is not whether we are dealing with so called “hard science” — work done with beakers and bunsen burners.  Rather, the question is whether the work is done reliably and rigorously, according to the scientific method, and in accord with principles and protocols that protect against human biases in the laboratory.  So for now, juries in Pennsylvania will have to do without the best scientific knowledge on why we might have false confessions.

But, even more important, the issue of whether experts can testify on these subjects is not a new one; courts in most jurisdictions resolved this issue years ago.  Instead, every institution that forms part of our criminal justice system in PA should be focused on prevention: establishing protocols for police to follow that will keep dangerous mistakes  eyewitness testimony and interrogation from happening in the future.   The research is there; the best practices are known.  Nevertheless, we remain stuck in the past, arguing about expert witnesses.


For those in the Chicago area, I’ll be speaking about my book “Failed Evidence: Why Law Enforcement Resists Science” on Wednesday, October 16, at 5 p.m. at Gage Gallery, 18 S. Michigan Avenue.  The event is free and open to the public.  The event is sponsored by Roosevelt University’s Joseph Loundy Human Rights Project, and is part of their annual speaker series.  The link to the event is here.

The next public event for “Failed Evidence” will be in Pittsburgh on November 6, at noon at the Harvard Yale Princeton Club, 619 William Penn Place, The talk will be sponsored by the Allegheny County Bar Association and the Pitt Law Alumni Association.  More details to follow.

It has been almost fifteen years since the National Institute of Justice (NIJ) recommended comprehensive changes to the ways that police conduct identification procedures for witnesses.  Yet USA Today reports that a new NIJ report reveals that 84 percent of U.S. police departments still have no policy to govern how live lineups are conducted.

Readers of Failed Evidence know that almost three quarters of wrongful convictions  overturned through DNA feature incorrect eyewitness identifications; it’s the single largest source of error in these cases.  Readers also know that a growing number of jurisdictions (for example here and here,) have adopted changes to eyewitness identification procedures to eliminate these errors, through blind administration of lineups, use of sequential lineups, and other procedures that decades of research has proven to reduce these errors greatly.

But the NIJ study concludes that most of agencies have no policies for how officers conduct these crucial procedures.  Eighty-four percent of the responding police departments said that they had no policy for live lineups, and sixty-four percent said they had no policy for conducting photo lineups.

The study was conducted for NIJ by the Police Executives Research Forum, “a police research organization and a provider of management services, technical assistance, and executive-level education to support law enforcement agencies.”   PERF surveyed hundreds of U.S. police agencies, large and small, to determine how many had at least taken the step of creating a policy to require officers to use at least some of the recognized best practices for eyewitness identification.  PERF researchers found larger police agencies more likely than smaller ones to have a policy in place, but even among the large agencies, fully twenty-five percent did not have a policy.

For those who follow these issues, as I do, the fact that so much of the law enforcement establishment has failed to make changes in basic investigative procedures is not entirely shocking; the resistance to science and the changes it points to is the central theme of my book Failed Evidence.  But the results of the study are nonetheless disappointing, if only because eyewitness identification reform is one of the areas in which there is the greatest consensus that basic changes are needed; there is also broad consensus one what those changes needed most are: blind lineups, sequential lineups, valid ways of dealing with witness confidence statements, instructions to witnesses, and the like.

A decision by Oregon’s Supreme Court on eyewitness identification procedures has re-set the way that juries and courts in that state will think about eyewitness identification.

According to the New York Times editorial on the case, the ruling shifts the burden of proof to prosecutors to prove that eyewitness identifications are reliable before they can be admitted in court.  Before last week’s decision, the rule had been that identifications were generally admitted; it was up to the defense in individual cases to prove that an identification was not reliable.

But at least as important as the new rule itself was the reason that the Oregon court abandoned its old precedent:  the court had concluded that the old rule was based on assumptions about eyewitness testimony no longer supported by the science.  Thus the new case represents a textbook case of a court forcing law enforcement away from the failed evidence of discredited methods, and toward methods that accord with what science teaches us now.

Under the old rule, Oregon judges looked at five factors when evaluating an eyewitness identification: opportunity to view the alleged perpetrator, attention to identifying features, timing and completeness of description given after the event, certainty of description and identification by witness, and lapse of time between original observation and the subsequent identification.  Looking at these factors from the vantage point of the present day, the Oregon court found them “incomplete and, at times, inconsistent with modern scientific findings.”  Given the science on eyewitness identification that is by now well established, the court prescribed a new approach, including the change in the burden of proof.

That’s what the Oregon Supreme Court did, but here is why they did it:

…[W]e believe that it is imperative that law enforcement, the bench, and the bar be informed of the existence of current scientific research and literature regarding the reliability of eyewitness identification because, as an evidentiary matter, the reliability of eyewitness identification is central to a criminal justice system dedicated to the dual principles of accountability and fairness.

It’s hard to imagine a better summing up of the ideas behind Failed Evidence, and why the fight to overcome law enforcement’s general resistance to science is so important.

In connection with my talk today, Nov. 8, on Failed Evidence at the University of Minnesota Law School, I’ve been interviewed by Minnesota Public Radio.  The interview is posted today as part of The Daily Circuit program. You can get to it here.

I was a guest on WYPR Public Radio’s “Midday” program today, discussing Failed Evidence.  Today was the monthly “Midday on Science” show, and host Dan Rodicks and regular science contributor John Monahan asked great questions on everything from DNA to more traditional forensic sciences to eyewitness identification and false confessions.  Listeners asked terrific questions too.

You can hear the whole show by clicking here and clicking on the audio button.

Why has the LAPD chosen to remain resistant to science-based eyewitness identification practices?

The title of an August 24 LA Times article says it all: “LAPD Reluctant to Change Its Handling of Photo Lineups.”  Having lineups conducted by someone other than the detective assigned to the case, who knows which photo shows the person they suspect, is backed by decades of science, but the LAPD opposes it.

The reason for having someone other than a person who knows the “right” answer administer the lineup is based on a well-established scientific protocol to avoid unconscious human bias.  Researchers long ago established that when the administrator of an experiment gives the subject choices from which to pick, the subject picks the right choice more often when the administrator knows which choice is the right answer.  This is called the experimenter-expectancy effect.  It is a form of cognitive bias, in which the person running the experiment who knows the preferred answer gives the subject making the choice subtle but unmistakable cues, without either intending to or realizing it.  To avoid this, scientists :”blind” the person administering the experiment and communicating with the subject.  The administrator therefore does not know the right answer, and therefore cannot communicate it.

Blind experiments are among the most basic requirements of science; we would never take an over-the-counter medicine if we knew that it hadn’t been tested this way.

And yet, with jurisdictions as different as New Jersey, Connecticut and Texas going with eyewitness procedures based on science, including blind lineup administration, LA Police Chief Charlie Beck continues to resist.  According to the Times,  Beck said “if you don’t adhere to the rules, either process is flawed. It’s more important to do them correctly than it is which process you use.”

No — wrong.  I’m an admirer of Chief Beck (and his predecessor, William Bratton), after seeing the transformation of the formerly-broken  LAPD into one of the better large police departments in the country.   But in this instance, Chief Beck is wrong.  What’s important is to move to a process that produces measurably fewer errors.  Sure, people can screw up anything by doing it wrong.  But the evidence is not in doubt: putting a person not involved in the case in charge of lineups eliminates bias — not because police are bad or unfair, but because they are human beings, subject to the same cognitive errors all  human beings are.

Here’s a new take on wrongful convictions.  According to Frank Sedita III, the District Attorney of Erie County (Buffalo), NY, “their rate of occurrence has been obscenely exaggerated.”  The real problem, Sedita says, is wrongful acquittals.

Sedita apparently took exception to an editorial in the Buffalo News on August 15, highlighting a considerable number of recent wrongful convictions in New York State.  The paper chastised the state legislature for failing to make basic changes in eyewitness identification procedure and and for failing to require the recording of suspect interrogations.

This was apparently too much for Sedita, who had his own say in an August 26 op-ed.  In his office, he says, he has many layers of review to assure that no innocent defendant ever faces punishment.  And he calls the “obscenely exaggerated” problem of wrongful convictions a  “deliberate deception heaped upon an unsuspecting public.”


Let’s review.  According to the Innocence Project, there have been 297 people have been exonerated nationwide,  as of today.  The National Registry of Exonerations, which uses broader criteria, puts the number at 945.  For the sake of argument, let’s stay with the lower, more conservative number, which counts only DNA-based exonerations.  The fact is that testable DNA is available in only five to ten percent of all cases.  This means that the 297 cases over the last two decades represents only the tip of the proverbial iceberg.  There are undoubtedly many times more cases of wrongful convictions, but no DNA evidence to test in those crimes.

What about “wrongful acquittals” which Sedita says far outpace wrongful convictions?  Sedita uses “wrongful acquittals” to mean one of two things: 1) cases in which defendants were acquitted, “despite overwhelming evidence”  of guilt; or 2) dismissals following indictment, “because of technical procedural issues or because the court suppresses key prosecution evidence at the request of the defense.”  But wait: if there was “overwhelming evidence,” why did the judge or jury acquit?  Perhaps not everyone viewed the evidence the way Sedita did.  And those “technical procedural issues” or the suppression of key prosecution evidence: we also call those things rulings based on the U.S. Constitution and on state and federal law, made by our elected representatives.  In other words these wrongful acquittals are not actually wrongful at all.  It is the way our constitutional legal system is supposed to function.





What do Texas and Connecticut have in common?  No, New Englanders have not suddenly taken to wearing cowboy hats and boots.  As of August 23, both states are among those recently moving toward better practices in handling eyewitness identification and testimony.

In my last post, I discussed how a new Texas requiring some science-based identification practices would take effect on September 1.  This had, of course, happened in a state in which uncovering wrongful convictions seemed almost routine.  Now, Connecticut’s Supreme Court has moved that state a little closer to the proper ways to use and think about eyewitness identification.

As reported in an article in the August 23 Hartford Courant, the state’s Supreme Court announced that based on the current state of science in eyewitness identification, it was throwing out old precedents and would now allow the defense to call an expert witness in any case in which eyewitness identification played a significant role.  The expert’s testimony would be appropriately limited.

An expert should not be permitted to give an opinion about the credibility or accuracy of the eyewitness testimony itself; that determination is solely within the province of the jury. Rather, the expert should be permitted to testify only about factors that generally have an adverse effect on the reliability of eyewitness identifications and are relevant to the specific eyewitness identification at issue.

Basing its decision on what it called “a near perfect scientific consensus,” the court said that the research “convincingly demonstrates the fallibility of eyewitness identification testimony and pinpoints an array of variables that are most likely to lead to a mistaken identification.”

Allowing expert testimony does not mandate a change to police procedures, the way that Texas and some other jurisdictions, like New Jersey, have.  But it is a step in the right direction.

Tell me again: why can’t other states and police departments do these things?  Why do they continue to resist the science?



Texas leads the nation in wrongful convictions.  That’s why it’s interesting to find some of the state’s law enforcement leaders no longer mouthing the tired old tropes about why they can’t change eyewitness identification procedures.  The state mandated some changes with a new law, which takes effect on September 1.    So now some members of law enforcement in Texas say: it’s important, and we’re going to get it done.

A bit of background: last fall, two state senators introduced a bill requiring some basic science-based reforms in eyewitness identification procedures.  The bill passed, and on September 1, police agencies conducting eyewitness identifications must do so using a sequential process (showing members of a lineup or photos of a lineup to a witness one person at at time).  Police also must hold “blind” lineups, by having the lineup or pictures shown to the witness only by someone not involved in the investigation.  And each agency must adopt a written policy on eyewitness identification. The bill requires other reforms, too.

When changes like this are proposed, there is usually wide resistance by police and prosecutorial leadership: these things will cost too much, be too logistically complex, and will be far too difficult for small agencies.

This is what made an article I saw in the Corpus Christi Caller-Times so striking.  The piece concerned training for law enforcement on the new eyewitness procedures, conducted by Nueces County District Attorney Mark Skurka.  According to the article, Mr. Skurka told about eighty officers  that “the changes came about because of an increase in exonerations due to DNA evidence.  Law enforcement and prosecutors previously relied mainly on eyewitnesses, he said, but that practice caused many wrongful convictions.”  This made the reforms necessary, Skurka said, and “whether you have a two-man agency in Driscoll or a 150-man agency in Corpus Christi,” every agency will have to get on board.

Perhaps there’s nothing more to this than the fact that the legislature required the changes.  I have no idea whether D.A. Skurka  believed or articulated these ideas about wrongful convictions and eyewitness testimony before the legislation or not.  But maybe that isn’t the point.  Once there is a requirement in the law, he has to move his people toward science-based best practices.   So, for advocates of change, making allies in the state legislatures to move these practices into law means everything.  (See Chapters six and seven of Failed Evidence, available next month.)