From Testimony to Judicial Error: When the Eye Sees What Memory Invents
- Caroline Hébert

- Jan 12
- 5 min read
In courtrooms as in the media space, the eyewitness often takes center stage: we listen to them devoutly, dissect their words, cling to their account like a lifeline in a case where material evidence is sometimes lacking. Their face marked by emotion, their voice that trembles or, conversely, their cold certainty, can sway a jury, influence a judge, or justify an arrest in mere minutes.
For a long time, the justice system as well as the general public have nurtured the idea that "someone who saw it with their own eyes" represents the strongest proof there is, almost beyond any suspicion. Yet, behind this reassuring image lies a reality far less comfortable for the serious investigator: human memory is neither a camera nor a recorder, but a fragile system, easily influenced, and profoundly reconstructed after the fact.
Psychological research, just like wrongful conviction files, has repeatedly demonstrated that sincere, convinced, and good-faith witnesses can be grossly mistaken about a suspect’s identity, the chronology of a scene, or even central elements like the presence of a weapon. Experimental studies have highlighted the effect of stress, poor observation conditions, suggestive questions, and post-event discussions on the quality of memory, to the point of creating false details, or even complete false memories in some individuals. Similarly, the analysis of numerous revised conviction cases has revealed that erroneous eyewitness identification ranks among the major causes of judicial errors in several countries, including Canada.

In this context, one question arises for any investigation professional: can we still trust eyewitnesses, and under what conditions? This article first proposes to examine why the credibility of eyewitnesses often leaves much to be desired in light of the main scientific studies on memory and identification. It will then address concrete cases where a witness’s word contributed to sending an innocent to prison, before outlining the reflexes and best practices that a seasoned investigator should adopt to exploit these testimonies without falling into the trap of blind trust.
The Myth of the Good Witness
In the popular imagination, the human eye functions like a camera that faithfully records the scene, ready to be replayed in court. Cognitive psychology shows the exact opposite: memory is reconstructive, filtered by stress, expectations, biases, and information received after the facts.
Several meta-analyses and review articles emphasize that identification errors are frequent, particularly in real judicial contexts. The American Psychological Association estimates that about one in three witnesses makes an erroneous identification in experimental protocols.
The Major Lessons from Scientific Studies
The work of Elizabeth Loftus has revolutionized the understanding of witness memory. In a classic study, Loftus, Miller, and Burns (1978) showed that a simple piece of misleading information given after viewing a film of an accident (for example, mentioning a "yield" sign instead of a "stop" sign) was enough to replace the true memory with a false one in many participants.
In another famous experiment, Loftus and Palmer (1974) demonstrated that the wording of questions alters memory itself: when asking subjects at what speed the cars had "collided" versus "smashed violently," speed estimates increased, and witnesses more often reported broken glass … which did not exist! This type of result illustrates the malleability of memory, even in good-faith witnesses.
Errors Leading to Wrongful Convictions
Witness errors are not just an academic problem: they literally send innocents to prison. In North America, the Innocence Project has shown that about 70% of exonerations obtained through DNA involved at least one erroneous eyewitness identification.
In Canada, several emblematic cases illustrate this danger: David Milgaard, Leighton Hay, Steven Truscott and Romeo Phillion were all wrongfully convicted of serious crimes, then released after years behind bars. In these files, witness identifications or accounts had played a key role in the conviction, before being discredited by subsequent objective evidence.
Organizations like Innocence Canada remind us that erroneous identification is one of the main known causes of recorded judicial errors in the country. In other words, a sincere, convinced, emotional witness can nonetheless contribute to a dramatic verdict … simply because their memory betrayed them.
Factors Undermining Witness Reliability
From an investigator’s perspective, several variables must immediately raise red flags around an eyewitness testimony. Among the most important:
Observation conditions.
Poor lighting.
Significant distance.
Very short exposure duration.
Presence of a weapon that draws the gaze to the danger at the expense of the perpetrator’s face ("weapon focus" phenomenon).
Stress, Emotion, and Eyewitness Credibility

Crime scenes are often experienced in a state of acute stress, which disrupts memory encoding.
The witness then fills in the gaps with their expectations, general knowledge, or what they are told afterward.
Biases, Group Effects, and Judicial Error
Cross-racial identification ("cross-race effect") is statistically less accurate: people generally recognize faces from an ethnic group different from their own less well.
Discussions between witnesses, media, or social networks can also homogenize initially different accounts, creating a contaminated collective memory.
Poorly conducted photo or lineup identification sessions (involuntary cues, overly distinctive suspect, lack of adequate fillers, group presentation rather than individual) significantly increase the risk of error.
To this is added a major trap: the confidence displayed by the witness. Juries give enormous weight to a confident witness who points out the accused without hesitation. Yet, the literature shows that the correlation between subjective confidence and actual accuracy is far from perfect.
How a Professional Investigator Should Handle a Testimony

For an investigator, the eyewitness testimony remains an important piece of the puzzle, but it should never be the single cornerstone. The professional reflex must be to treat the witness’s memory like a crime scene: exploitable, but extremely sensitive to contamination.
A few key operational principles:
Protect memory from the start.
Interview the witness as soon as possible, with open-ended questions rather than suggestive ones.
Avoid providing details that the witness might later incorporate into their account as if they came from them.
Document observation conditions.
Precisely note the distance, lighting, duration, viewing angle, and witness’s emotional state.
Be wary of late reconstructions, especially if the witness has been exposed to media or other versions of events.
Use rigorous identification procedures.
Resort to "blind lineups" (the agent does not know who the suspect is) to limit influences.
Specify to the witness that the perpetrator may very well not be present in the series of images, to reduce the pressure to absolutely choose someone.
Systematically seek corroboration.
Corroborate the testimony with video, phone data, material traces, objective chronology.
Treat any major contradiction with physical evidence as an alarm signal, not as a minor detail.
Finally, when a case rests essentially on a single eyewitness, especially under questionable observation conditions, maximum caution is required. In this type of situation, a seasoned professional will not ask "Is this witness lying?" but "Under what conditions could their memory have been altered?".
In summary, an eyewitness’s word should never be sacralized, even when delivered with assurance and emotion. The role of a professional investigator is to respect what the witness brings, while keeping in mind that the eye sees, but memory reconstructs!
Sources:
Némésis offers criminal profiling services, polygraph testing, investigation, and investigative analysis to Quebec police forces, legal professionals, as well as any entity or citizen requiring specialized expertise in investigation.
