Case File 264 · Investigation Psychology

How Does Alcohol Affect Eyewitness Memory?

By Murder Book Editorial TeamPublished by Gungrounds, the studio making Murder BookPublished 3 October 2026Updated 3 October 2026

Alcohol impairs eyewitness memory, but the relationship is more nuanced than a simple "drunk witnesses are unreliable." Research shows that the effects depend on how much was consumed, when it was consumed relative to the event, and how memory is later tested. At low to moderate levels of intoxication, witnesses can still provide broadly accurate accounts, though they tend to recall fewer details. At higher levels, free recall becomes significantly less complete, with peripheral and fine-grained details suffering the most. Intoxicated witnesses do not necessarily produce more false details; they mainly produce fewer correct ones. Understanding these effects matters because a substantial proportion of crimes are witnessed by people who have been drinking, and dismissing their accounts entirely would mean losing evidence that may still be valuable if gathered carefully.

Key takeaways

Alcohol primarily reduces the completeness of recall rather than increasing the rate of false details.
Low to moderate intoxication has a smaller effect on accuracy than commonly assumed.
Higher intoxication impairs free recall significantly, with fine details lost before the general outline.
The timing of consumption matters: alcohol consumed after witnessing an event may actually protect the memory from subsequent misinformation.
Careful interview techniques, such as the Enhanced Cognitive Interview, can improve recall from intoxicated witnesses.

What does the research show about accuracy?

A meta-analysis of studies on acute alcohol intoxication and witness recall found that alcohol had a significant, moderate-sized effect on the number of correct details recalled. Intoxicated witnesses reported fewer correct details than sober ones. However, the effect of alcohol on incorrect details was not significant, meaning that drinking did not cause witnesses to fabricate more false information. They remembered less, but what they did remember was not notably less accurate.

This distinction matters for investigations. An intoxicated witness's account may be incomplete, but the details they do volunteer are not automatically wrong. The practical question is whether those partial details can be supplemented by other evidence, a principle explored in why does corroboration matter?.

Does the level of intoxication matter?

Yes, substantially. Research has found that at low to moderate levels of intoxication, recall was not significantly impaired when tested through free recall or recognition tests. The measurable deficits appeared primarily at higher intoxication levels, and even then, the impairment was mainly to the completeness of free recall rather than to its accuracy.

Within higher intoxication, the type of detail affected also varies. High-salience details, the central actions and main features of an event, were more resistant to alcohol's effects than low-salience peripheral details. This mirrors the general pattern seen with other memory impairments such as stress and delay: the gist of what happened tends to survive better than the specifics.

At very high levels of intoxication, memory can fail substantially or produce blackouts, periods where no continuous memory is formed at all. These represent a qualitatively different problem from the graded impairment seen at moderate levels.

When does timing of consumption matter?

One of the more counterintuitive findings in this area concerns alcohol consumed after the event. Research has shown that when witnesses drink after witnessing a crime, the alcohol can actually protect the memory from subsequent misinformation. The mechanism appears to be that alcohol impairs the encoding of new information, including misleading details encountered after the event. Because the misinformation effect depends on new information being incorporated into the existing memory, blocking that incorporation can preserve the original account.

This does not mean post-event drinking improves memory. It means it may shield a memory from one specific threat, contamination by later information, while still carrying the other costs of intoxication such as reduced clarity and impaired recall of new details. The finding is a useful reminder that the relationship between alcohol and memory is not straightforward.

How should intoxicated witnesses be interviewed?

Research suggests that interview technique makes a meaningful difference to the quality of evidence obtained from intoxicated witnesses:

The Enhanced Cognitive Interview (ECI) has been shown to improve the correct recall, accuracy and completeness of moderately and severely intoxicated witnesses. The ECI uses techniques such as mental reinstatement of context, reporting everything, and recall from different perspectives to help witnesses access more of what they encoded.
Free recall first. Letting the witness describe what they remember in their own words, without interruption, tends to produce the most reliable information. Specific questions can follow, but they should be open-ended rather than leading.
Interview timing. There is a practical tension between interviewing early, when the memory is freshest, and waiting until the witness is sober enough to give a coherent account. Research does not offer a single answer, but recording a brief initial account and then conducting a fuller interview when the witness is more alert is a reasonable approach.
Do not dismiss the account. The biggest risk is not that the intoxicated witness will lie, but that their account will be discarded entirely when it could still contribute useful information, especially about central details.

What about identification?

Alcohol's effects on eyewitness identification follow a similar pattern to its effects on recall. Intoxicated witnesses are less accurate at identifying faces, particularly at higher levels of intoxication, and the effect is worse for faces seen briefly or under poor conditions. The combination of alcohol and a long delay before an identification attempt compounds the problem, since both factors independently weaken the memory trace.

This makes it especially important that identification procedures are fair and conducted carefully when the witness was intoxicated at the time of the event. A witness whose memory was weakened by alcohol is more susceptible to suggestive procedures, since they have less confidence in their own recollection and may be more willing to choose the option that seems most plausible rather than the one they genuinely recognise.

How does Murder Book handle uncertain accounts?

Murder Book: The Art of Interrogation asks players to weigh accounts from people whose clarity and reliability vary. The game, developed by Gungrounds, an independent studio in Zagreb, Croatia, is inspired by real police footage and psychological interviews. Just as an investigator must judge how much weight to give a witness whose perception was impaired, players examine unreliable stories, behavioural changes and narrative gaps to work out what is supported by the evidence and what is not.

The game does not hand players a clean, authoritative account. It gives them fragments, and the question is always how much each fragment deserves to be trusted. Cases are designed for focused 5–10-minute sessions. Murder Book is in development for PC (Steam) and mobile, with release currently planned for the end of Q2 2027.