This coming term, in Perry v. New Hampshire, the U.S. Supreme Court will consider an important question that has split the lower courts: If an eyewitness first identifies a perpetrator under highly suggestive circumstances that seriously compromise reliability, but the police did not orchestrate those circumstances, should a court exclude the identification evidence?
Or, stated differently, is police misconduct necessary to the successful due process exclusion of unduly suggestive eyewitness identification evidence?
Research Shows That Eyewitness Identifications Play a Significant Role in Wrongful Convictions
The question of unreliable eyewitness identification is very timely, because empirical research has now exposed the central role of mistaken eyewitness identifications in wrongful convictions.
A combination of factors contributes to the eyewitness problem. First, juries place a great deal of weight on eyewitness identifications. When a disinterested witness who was actually at the crime scene swears to the jury that he saw the defendant commit the offense, the jury finds such testimony very compelling, especially if the witness speaks with confidence and does not waver.
When people say disparagingly, “All we have is circumstantial evidence,” what they mean is that we lack precisely the sort of direct observations that an eyewitness can provide, where no inferences need to be drawn, because someone without an agenda was right there when it happened, and can report the events that took place.
We tend to view the statements of eyewitnesses in the way in which they are often portrayed in movies—as if eyewitnesses were live recording devices who can play back the video of the crime in their brains and tell us exactly what they see there.
A second factor contributing to eyewitnesses’ role in convicting the innocent is the fact that despite its perceived accuracy, the process of identifying a person after only a brief opportunity for observation is, in reality, vulnerable to errors that take place outside of the witness’s own awareness.
One common error involves the witness’s selecting as the perpetrator the person who, among all those in the lineup, most closely resembles the actual perpetrator. Facing a lineup, witnesses frequently assume that one of the people there must be the correct choice and thus, they simply select the one who best matches their recollection of the criminal.
The same tendency also affects a “showup”—in which only one suspect is brought before the witness, and the witness is then asked “Is this him?” In a showup, the witness will be inclined to think that the answer must be yes, so long as the suspect looks like the real criminal.
Another common error that compounds the others is that once a witness has identified a person as the perpetrator, the image of the identified person can replace the image of the original criminal in the mind of the witness. That is because when one retrieves a memory to compare it to a present stimulus, the original memory undergoes modification in one’s brain. Far from the fantasy in which we retrieve a static video stored in our brains, in other words, an eyewitness who looks at the defendant from the witness stand is often retrieving a “video” that has already changed in response to the witness’s prior identification.
In effect, this means that when an eyewitness remembers the crime in her mind, the memory will have superimposed the image of the person whom she identified in the lineup upon the witness’s older image of the perpetrator from the original crime scene. This memory “update” is not conscious, and it is nearly impossible for a witness to detect. As a result, a witness might be absolutely—and very genuinely—certain that an innocent defendant is the one who robbed the bank, because the witness has already identified the defendant in a lineup, and her image of that innocent defendant has now been retroactively incorporated into her older memory of the crime itself.
Understanding to some extent the risks of suggestive identification contexts and retroactive memory adjustment, the Supreme Court ruled over thirty years ago, in Manson v. Braithwaite, that when police subject an eyewitness to an unduly suggestive identification procedure resulting in a very substantial likelihood of irreparable misidentification, the Due Process Clause of the Fourteenth Amendment may require exclusion of the witness’s identification at trial.
As research has continued in this area, we have learned that even procedures that appear innocuous may still move an eyewitness to identify an innocent person. The Supreme Court is therefore wise to take up this important issue and consider it in light of recent developments in memory research.
Unintentional Unreliability and the Constitution
Under current law, there is little controversy regarding the proposition that when a police officer deliberately arranges for a highly suggestive lineup, the witness’s identification may be excluded from trial.
Imagine, for example, that an eyewitness describes the perpetrator of a grand larceny as a slim Latino male of approximately 30 years of age, measuring 6-foot-3-inches in height, with short medium brown hair. Imagine further that the police then arrange a lineup in which the arrested suspect (#3 in the line) matches the witness’s description, but in which each of the other five people in the lineup is an obese 20 year-old who measures no taller than 5-foot-6-inches in height and has long blonde hair and blue eyes. Imagine, in addition, that three of the people in the lineup are not even male.
It is quite obvious that the police in these circumstances are broadcasting to the witness from whom they are seeking an identification that the right person to choose is #3. This highly suggestive lineup would place a burden on the government to persuade the trial judge of the reliability of any subsequent identification.
Most of the time, however, police do not orchestrate such blatantly inappropriate identification procedures. The Supreme Court, moreover, has articulated a demanding standard for exclusion of identification evidence: Was the procedure employed “so unnecessarily suggestive and conducive to irreparable mistaken identification that [the defendant] was denied due process of law”?
In answering this question, the Supreme Court has directed judges to perform a “totality of the circumstances” test. In Neil v. Biggers, the Court explained the essence of the due process right at issue, saying that “the primary evil to be avoided is ‘a very substantial likelihood of irreparable misidentification.’” Emphasizing this point, the Court added that “[i]t is the likelihood of misidentification which violates a defendant’s right to Due Process . . . . Suggestive confrontations are disapproved because they increase the likelihood of misidentification, and unnecessarily suggestive ones are condemned for the further reason that the increased chance of misidentification is gratuitous.”
The Court thus recognizes the defendant’s strong interest in avoiding conviction based on an identification that is likely to be wrong, and the Court finds especially disturbing the occasions when police gratuitously undermine this interest. There are then two main objectives in play: to protect the defendant from procedures posing a high risk of producing erroneous identifications; and to guard, especially, against unnecessarily suggestive procedures.
The Facts of Perry v. New Hampshire, the Identification Case the Court Will Hear This Term
Let us turn now to the facts of Perry v. New Hampshire, the identification case currently before the Court. In Perry, the police received a report that “a black male was trying to gain entry into vehicles located in the back parking lot of an apartment building.” A uniformed officer, Nicole Clay, responded to the report by traveling to the parking lot in question in a marked police car. Upon her arrival, Officer Clay saw an African-American man, Barion Perry, standing between two cars and holding two sound-system amplifiers. Officer Clay asked Perry to put down the amplifiers and come over to speak with her. Perry did as instructed, and Clay and Perry spoke with each other about the suspected crime.
In the meantime, another man approached the officer and Perry. This other man had complained about his car being broken into, and he explained to Officer Clay that although he did not witness the break-in, his neighbor did. Clay then asked another uniformed officer to stay with Perry, and Clay accompanied the car owner to his neighbor’s apartment.
Officer Clay then interviewed the victim’s neighbor, Nubia Blandon, in the open doorway of the latter’s apartment. Blandon gave Clay a description of the crime that she had witnessed but, when asked to describe the perpetrator, said only that he was a “tall black man” and gave no more detail.
When Clay requested further description, Blandon pointed to her window and said that it was the man in the back parking lot standing with the police officer.
A Hypothetical Case Related to—But Different From—the Factual Scenario in Perry v. New Hampshire
For our analysis here, let us change the factual scenario slightly. Remember that Officer Clay had asked the neighbor to describe the thief, and that Blandon had said he was a “tall black man” without offering any further description. Now, however, imagine that Officer Clay had said at the time, “Look out the window and tell me whether you see the thief in the parking lot now. If Blandon had responded, “Yes, it is the man standing with a police officer,” then we would appear to have a classic, highly suggestive identification for due process purposes. The fact that the eyewitness was apparently unable to tell the officer anything more specific about the perpetrator’s appearance than his race and overall height would also tend to undermine our confidence that Blandon might truly have an independent recollection of the perpetrator whose acts she had witnessed. The defendant might accordingly succeed in persuading the judge to suppress evidence of Blandon’s identification.
The Key Difference Between Perry and the Hypothetical Case
In the actual facts of Perry, though the circumstances surrounding the identification were highly suggestive, the police did not arrange or otherwise orchestrate that suggestiveness. Indeed, Officer Clay spoke with Blandon at the apartment’s doorway, from where she “was not able to see out of the apartment window down to the parking lot where Mr. Perry was standing.”
Unlike in the case of suggestive lineups and showups, then, the police concededly did not engage in misconduct here. The New Hampshire Supreme Court found, on this basis, that the identification was not subject to due process suppression. In the absence of any state misconduct inducing an unreliable identification, the court ruled that introducing Blandon’s identification into evidence would not even implicate—let alone violate—the defendant’s due process rights.
Other courts (not to mention Perry himself) see the matter differently, and suggest that the key interest at stake is in avoiding a highly unreliable identification that could put an innocent person behind bars. What ought to count, given that interest, is whether the circumstances under which the witness identified the defendant created a substantial risk of producing an erroneous identification, regardless of whether the police played a role in arranging for those circumstances.
Is the Right at Issue One Against the Introduction of Evidence, or One Against Police Misbehavior?
On the petitioner’s approach, the right at issue guards against the introduction of highly unreliable evidence, rather than against police misbehavior. As the petitioner explains in his brief in Perry, we keep out such evidence in order to protect the evidentiary integrity of the trial, not to deter police misconduct. Unlike the Fourth Amendment exclusionary rule, Perry argues, the due process right is a trial right, and the governmental conduct that triggers the Constitution’s application is thus the government’s introduction at trial of unreliable evidence arising from a suggestive identification procedure, not police use of a suggestive identification procedure.
This argument is powerful in one sense, because the Court has specifically stated that “[i]t is the likelihood of misidentification that violates a defendant’s right to Due Process.” If that is so, then the police officers’ lack of culpability would seem immaterial to the violation. After all, if we want to avoid false convictions, our focus is appropriately on the conditions under which innocent people are likely to be convicted, rather than on whether the police conducted themselves in a culpable manner.
The difficulty in making this argument successfully, however, is that a ruling for the defendant would call into question much of the evidence that is currently—and, it has been thought, legally—introduced at criminal trials. If the government violates the defendant’s due process rights simply by introducing unreliable evidence, then many government witnesses will be constitutionally barred from testifying.
Are Eyewitnesses Generally Trustworthy? Why Perry Might Lead to an Exception That Swallows the Rule
On what basis do I make this claim about government witnesses? I rely on the fact that much of the evidence offered by the government (and the defendant) during a trial is of highly questionable reliability.
Quite apart from issues regarding the suggestiveness of identifications, eyewitnesses are prone to a variety of errors when they recall events and recount the specifics of who did what to whom at an earlier time.
Studies, for example, have demonstrated that it is not that unusual for an eyewitness to remember a bystander as the perpetrator and vice-versa. Anecdotally, most of us can probably think of a situation in which we have heard another person describing an event at which we were present, and have realized that the description is inconsistent with what we ourselves remember occurring. Though inconsistencies of this sort are generally inconsequential, they give us a small glimpse of how unreliable eyewitnesses can be, even if no one is encouraging them to err in a particular direction.
Beyond human weakness in accurately recalling the details of prior events, people are also quite capable of dishonesty. There are those who lie more than others, but most people lie sometimes, and some lie with regularity.
State witnesses in a criminal case can be especially prone to lying on the stand, both because such witnesses may have been involved in the crime or some other wrongdoing for which they hope to receive lenient treatment from the government, and because these witnesses’ own prior criminal activity may itself have habituated them to shading the truth. Furthermore, a given witness might dislike the defendant or have a racial bias or simply enjoy being in the spotlight provided by a criminal trial.
To suggest that eyewitnesses are generally untrustworthy may sound like hyperbole, but it is worth remembering that in actual trials, the government will frequently introduce the testimony of witnesses who pose a considerable risk of offering unreliable and/or deceitful statements. It is not that unusual for a person who is awaiting trial and who is correspondingly vulnerable to government pressure to appear in court to testify for the government against a defendant, and he may even be the only available witness.
Another type of person, one who has been repeatedly convicted of crimes, may also testify for the government. In neither case will the defendant be able to invoke the Due Process Clause successfully to keep the witnesses off the stand. Our system permits testimony by highly unreliable witnesses, and we rely on the jury to assess witness credibility, by observing witness demeanor and watching cross-examination and other forms of impeachment. Unreliable testimony is, thus, a routine part of criminal trials.
Perry, of course, can credibly argue that eyewitness identifications under suggestive circumstances are different from most types of unreliable testimony, because juries appear systematically to overvalue such identifications, notwithstanding their unreliability. I suspect, however, that this overvaluation occurs in large part because juries simply do not know how easily a seemingly credible and confident witness can come to hold memories of innocent defendants committing crimes.
A Possible Remedy for Unreliable Eyewitness Testimony After Suggestive Identifications: Educate Jurors About the Perils
When we have no government misconduct infecting a trial, then, the best way to approach this form of unreliable testimony may be to do what we do when other unreliable witnesses testify: educate the jury about how the circumstances under which the identification was made distort memory. Just as we tell the jury of a witness’s motive to lie, or of his history of perjury, we could also tell the jury about suggestive identifications.
When the police intentionally orchestrate the problem, then it becomes appropriate to exclude from trial the highly unreliable identifications that can result. At that point, police misconduct is itself responsible for tainting the memory of the eyewitness, and admitting the witness’s testimony would accordingly permit the State to introduce evidence with which it has essentially tampered.
On the other hand, when police do nothing improper to bring about the highly suggestive circumstances, admitting the identification seems no more of a due process violation than admitting the testimony of a convicted criminal or an interested party.
The Supreme Court appears to concur in the important role of police misbehavior in due process violations when it uses the word “unnecessarily” to modify “suggestive.” “Unnecessarily” suggestive implies that police should have, and could have, employed a less suggestive procedure and still obtained the witness’s identification.
A test that includes a standard of necessity, as this one does, evaluates the legitimacy of choosing a particular course of conduct. If unreliability were all that mattered, as Perry suggests in the case before the Court, then the absence of any better alternative means of obtaining a witness’s identification would be irrelevant to the due process analysis. Only if police wrongdoing matters, do we care that the suggestive procedure was not only harmful but also unnecessary.
When, through no fault of the government, identification occurs under highly suggestive circumstances, a judge may legitimately conclude that the probative value of the evidence is sufficiently compromised by the danger of unfair prejudice to be excluded for that reason. This is how courts generally handle evidence that is likely to be overvalued by jurors, including hearsay and inflammatory photographs. When courts do admit such identifications, moreover, the defendant may have a Sixth Amendment right to confront the eyewitness testimony with his own evidence of how and why eyewitnesses predictably and systematically err.
In the end, for better or for worse, we give the jury the job of weighing direct evidence. Courts will best enable the jury to do its job by making sure that jurors understand when and why there is reason to doubt an eyewitness’s sworn statements. The Constitution need not always be involved.