Second Circuit Criminal Law Blog

Failure to Challenge Eyewitness ID Evidence is Ineffective, Circuit Finds

On April 15, 2020, the Second Circuit (Rakoff, by designation, Sack, Hall) vacated the conviction of Ralph Nolan, who was convicted of conspiracy and attempt to commit a Hobbs Act robbery, on ineffective assistance of counsel grounds.  The panel ruled that Nolan’s trial counsel’s failure to challenge the introduction of eyewitness identification evidence against him or to call an expert witness to guide the jury on evaluating that evidence departed from the standard of reasonable professional care.  The case, United States of America v. Ralph Nolan, No. 16-3423, represents a significant endorsement by the Circuit of research showing that in certain circumstances, such as those present here, eyewitness identification evidence will often be unreliable.  Nolan will be cited frequently by defendants pressing forward to seek post-conviction relief and it will also serve as a wake-up call for defense counsel who might have been unfamiliar with this research or on the fence about whether to call an expert witness to testify on the subject of eyewitness identification.


On December 16, 2013, two individuals committed an armed home invasion of an apartment in the Bronx that was home to number of family members who dealt drugs.  The individuals threatened the victims with guns and demanded drugs and money.  One individual pistol-whipped one of the victims, and ripped cable wires out the wall and used them to bind the hands of the victims.

The five victims—four members of the same family, and the home health aide of one of those family members—reported the crime to the police.  Several of the victims stated that the robbers may have been Hispanic (Nolan is white), and though two of the victims previously knew Nolan, none identified him as one of the robbers.  Three of the victims were shown a photo array that did not include Nolan, and all three were unable to identify the robbers from the array.

Several weeks later, based on a detective’s suspicions that a particular individual was a co-conspirator in the robbery, the police added Nolan’s picture to the photo array.  They showed the photo array to one of the victims again; this time, she identified Nolan, but said she was only 75% sure.  That victim then entered the room where another victim was, and the two discussed whether the photo of Nolan was that of their assailant.  The two victims, in the presence of the NYPD detective, then viewed additional photos of Nolan on his Facebook page.  The victims shared Nolan’s Facebook photos with each other; all four family members (who were offered immunity in exchange for their testimony, we suspect because of the drug-related nature of the robbery) then identified Nolan.  The home health aide, the only victim not offered immunity, was the only victim not to make an identification.

Defense counsel initially made a pre-trial motion to exclude in-court identifications of Nolan by the four family members.  But she inexplicably withdrew the motion, and decided instead to pursue a strategy of impeaching the credibility of those identifications on cross examination.

The government’s evidence against Nolan consisted mainly of the four witnesses’ eyewitness identifications, and a photo taken from Nolan’s Facebook page of him holding a BB gun which the government argued demonstrated “the defendant’s . . . access to firearms and knowledge of them.”  On April 10, 2015, a jury convicted him of conspiracy to commit a Hobbs Act robbery and attempted Hobbs Act robbery.

The Appellate Decision

The Second Circuit, having consolidated Nolan’s direct appeal and appeal of the district court’s denial of his § 2255 motion—which raised the ineffective assistance of counsel argument—held that Nolan’s counsel was ineffective under the long-standing test outlined in Strickland v. Washington: trial counsel’s omissions fell “outside the wide range of professionally competent assistance,” and this ineffectiveness prejudiced his conviction.

First, the Court cited to academic research—and to court opinions, including those of two state supreme courts, adopting that research—casting doubt on the accuracy of eyewitness identifications in certain circumstances that were present here.  For example, the Court noted research indicating that

  • cross-racial eyewitness identifications are notoriously unreliable,
  • individuals who are threatened with a gun often experience “weapons focus” where their attention is focused on the gun in front of them and not on the individual brandishing the gun,
  • witnesses do a poor job of identifying people who wear masks or face coverings,
  • eyewitness identifications are less reliable as more time elapses between the incident and the identification, and
  • identifications that result from police suggestiveness are often unreliable. 

All such “glaring indicia of unreliability” were present in Nolan’s case: the victims were Black and Hispanic; Nolan was white.  The robbers were armed and wore face coverings.  The victims did not identify Nolan until several weeks after the incident occurred.  And the identifications were the result of “the police employ[ing] highly irregular procedures” by allowing the victims to “talk among themselves about Nolan’s identification and allowing them to view his photos on Facebook.”  The sheer number of issues with the reliability of the evidence against Nolan was such that “an effective defense counsel would have vigorously contested the admissibility of the testimony.”   The trial counsel’s failure to do so “fell outside ‘the wide range of reasonable professional assistance.”  The entire case would have fallen apart if the motion had been successful, and even if unsuccessful, a motion would have alerted the district court to the serious problems with the identification.

Next, the Court turned to trial counsel’s failure to call—or even consult—an expert witness who could have raised for the jury, or even for trial counsel, issues with the reliability of eyewitness identifications such as those identified by the Court.  “Strickland ordinarily does not require defense counsel to call any particular witness,” the Court wrote.  But in light of the fact that the questionable eyewitness identification evidence comprised virtually all the evidence against Nolan, the Court held that failure to call an expert was a prejudicial “fail[ure] to render reasonable professional assistance.”

Finally, the Court addressed trial counsel’s failure to move to exclude the photo taken from Nolan’s Facebook page of him holding what looks to be a handgun, but is actually a BB gun.   The Court effectively conducted its own Rule 403 analysis and determined that the photo would be “obviously prejudic[ial]” to Nolan, with little probative value.  Accordingly, the Court saw “no strategic rationale for defense counsel’s failure to move to exclude the photo” or to seek a limiting instruction for the photo.  This, too, was prejudicial error in light of the “emotional reaction among the jurors” that this photo likely elicited.


In the past decade, the Supreme Courts of New Jersey and Oregon have each required courts in those states to consider research on eyewitness identification reliability before admitting such evidence.  The recognition of the limitations of eyewitness identifications has long been understood by psychologists to be a real problem.  When one of the authors of this article took criminal law as a first-year law student, the professor staged a robbery of himself in the classroom and then asked people to identify the person who committed the fake robbery—was he tall or short, light-skinned or dark-skinned, what color clothing, long hair or short?  There was no consensus among the students even though the “robbery” occurred right in front of them and only moments before the discussion.  Lesson learned!

In light of this decision, defense attorneys and prosecutors in the Second Circuit will need to learn the same lesson.  The Court’s adoption of the scientific research, and the panel’s repeated references to the unreliability of the evidence against Nolan in light of that research, sends a clear sign to practitioners and judges that eyewitness identification under similar circumstances may only be admitted sparingly and after a careful evaluation of the circumstances surrounding the identification.  The Court takes care to stress that the case involved “highly irregular procedures” by the police with respect to obtaining eyewitness identification and that the case presented “unusual circumstances,” and so it would not appear that every case in which there is eyewitness identification will necessarily call for an expert witness and a motion to exclude.  That said, it is hard to see how a defense lawyer could avoid taking these steps in a case with cross-racial identification, the use of weapons, and/or the wearing of masks. 

Moreover, the Court’s declaration that Nolan’s counsel was per se ineffective for failing to call an expert witness on eyewitness identification is another warning sign to counsel and for that matter, to trial judges.  Even the jury instruction given by the district court about the perils of eyewitness identification was not enough to save this conviction.  Through an expert witness, practitioners and judges will need to ensure that future juries consider eyewitness identification evidence in the context of the research identified by the Court.

-By Emma Ellman-Golan and Harry Sandick