Eyewitness memory: stress experiments and the record of an identification
Eyewitness memory did not become reliable simply because the people being tested had military training. In a study reported by Yale in 2004, 509 active-duty personnel underwent survival-school exercises involving interrogation. A day after leaving the mock prisoner-of-war camp, they had to identify the people who had confronted them. Recognition was better after the lower-stress interrogation than after the physically confrontational version. The experiment put a measurable gap between enduring a threatening encounter and later recognizing the person involved.
The training involved 48 hours of food and sleep deprivation before interrogation. Two instructors occupied the room, one acting as a guard and the other as the interrogator. The high-stress condition involved physical confrontation; in the lower-stress condition, the interrogator attempted to trick the subject into disclosing information. Identification took place 24 hours after release, using a live lineup, a photographic spread or sequential photographs. Yale's report says the lower-stress advantage appeared regardless of presentation method. Some participants even misremembered the guard's or interrogator's gender. Those results concern a controlled military-training setting, not a measured error rate for all crime victims.
A weapon adds a related but distinct difficulty for eyewitness memory. The National Research Council's 2014 review, Identifying the Culprit, describes evidence that a weapon can capture visual attention at the expense of other details, including the offender's face. It distinguishes remembering features from selecting a person in a later identification procedure. Its discussion of research syntheses reports variable effects: a 1992 analysis found poorer identification and feature recall, while a later analysis found the identification effect inconsistent and larger in threatening scenarios. A gun's presence is a potential influence, rather than a numerical verdict on an individual witness.
That review also separates circumstances of the encounter from procedures investigators can control. The first group includes stress, exposure time and the interval before identification. The second includes the organization of the identification test. The distinction puts limits on what a later lineup can repair. The review identifies missing details and uncertain ground truth as obstacles in studies of actual crimes, where researchers may not know independently whether a particular selection was correct. Laboratory control and real criminal investigations each supply evidence with different practical limits.
An original eye-tracking study published in 2023 tested the usual explanation of weapon distraction more directly. Participants in Germany watched a man committing a robbery while holding either a gun or a mobile phone. The researchers used videos and slide shows and recorded eye movements. With slides, the gun drew more gaze than the phone, but that extra attention did not reduce viewing time on the perpetrator. With videos, weapon presence did not appear to change how long participants looked at him. The researchers reported no reduction in recalled appearance details in either experiment.
An additional online study using the original videos also found no such memory effect. The investigators concluded that weapon focus was not universal. Their findings limit an automatic sequence in which a visible gun must divert gaze from a face and must then impair recall. The absence of an effect in those experiments concerns their specific viewing conditions. Eyewitness memory research contains both positive findings and unsuccessful replications, with results tied to particular stimuli, procedures and measures of performance.
The legal treatment of an identification involves another distinction. In Perry v. New Hampshire, decided on 11 January 2012, the U.S. Supreme Court considered a witness who identified Barion Perry from an apartment window while he stood beside a police officer in a parking lot. Police had responded to a report of a man trying to break into cars. Perry sought to exclude the identification as unreliable. The Court held that the Due Process Clause did not require a preliminary judicial reliability inquiry when unnecessarily suggestive circumstances had not been arranged by law enforcement.
That holding addressed the trigger for a particular constitutional screening procedure. The majority recognized the fallibility of eyewitness identification and relied on other trial safeguards, including cross-examination, counsel, jury instructions and evidence rules. It did not announce that a witness standing above a parking lot could never be mistaken. The court record illustrates why admission into evidence and scientific reliability are separate matters: a constitutional rule can assign evaluation to the jury without resolving every influence on the original observation or later recollection.
The Justice Department addressed the handling of eyewitness memory in a policy announced on 6 January 2017. Except in exceptional circumstances, federal agents were to use blind or blinded photo-array procedures. In blind administration, the administrator is outside the investigation and does not know the suspect's appearance. In blinded administration, precautions prevent the administrator from seeing the arrangement presented to the witness. The policy also called for confidence to be documented at the initial identification, using recording or a transcription close to the witness's exact words. That preserves an early statement before confidence at trial can become a different piece of evidence.
- Yale News, Eyewitness Memory Poor in Highly Intense and Stressful Situations, 3 June 2004: institutional report of Morgan and colleagues' survival-school experiment — news.yale.edu
- National Research Council, Identifying the Culprit, 2014, chapter 5: applied identification research, weapon focus and stress — www.nationalacademies.org
- Revisiting the role of attention in the weapon focus effect, 2023: original eye-tracking experiments, PubMed Central — pmc.ncbi.nlm.nih.gov
- U.S. Supreme Court, Perry v. New Hampshire, No. 10-8974, slip opinion, 11 January 2012, Cornell Legal Information Institute copy — www.law.cornell.edu
- U.S. Department of Justice, Department-wide eyewitness-identification procedures, 6 January 2017 — www.justice.gov