EVIDENCE LIBRARY
Can Faster Answers After Alcohol Hide Less Accurate Visual Judgments?
A visual-number experiment found faster responses but lower accuracy after alcohol. What its breath-derived values and reporting limitations mean for DUI evidence.
Faster answers after drinking do not necessarily mean better performance. In a laboratory study of visual number judgments, alcohol reduced average accuracy from 67% to 64%, even though responses to larger displays became faster. For DUI evidence, the finding illustrates why response speed and response quality need separate assessment. It does not establish that a particular driver was impaired or predict that driver’s crash risk.
Harvey and Seedhouse’s 2021 experiment asked people to judge how many squares briefly appeared on a screen. The study addresses a narrower question than whether alcohol generally slows thinking: can someone respond quickly while processing a complex visual display less accurately?
What participants actually did
The investigators first measured sober working-memory capacity with an operation-span task combining letter recall and arithmetic. Participants then completed an enumeration task while sober and repeated it after a vodka beverage. Each trial displayed between one and nine squares for 200 milliseconds, followed by a visual mask. Participants entered the number they thought they had seen.
Recognizing a few items at a glance is often called subitising. Larger displays require more counting or estimation. That distinction allowed the researchers to examine whether alcohol-related changes appeared mainly in simple recognition or in more demanding judgments.
The alcohol condition always followed the sober condition. It was not a counterbalanced, placebo-controlled comparison. Consequently, practice, fatigue, expectations and the passage of time cannot be separated cleanly from the beverage effect.
The alcohol values were breath-derived estimates
The researchers measured alcohol with a Dräger Alcotest 3000 in milligrams per litre of breath. They converted those readings into BAC percentages using an assumed blood-to-breath ratio of 2,300:1. They reported converted values ranging from 0.03% to 0.08%, averaging approximately 0.05%.
Those percentages were not directly measured whole-blood ethanol concentrations. No serum/plasma, urine or alcohol-biomarker result establishes the exposure in this experiment. The conversion assumption should remain attached to the numerical values whenever the findings are discussed in a DUI context. The study did not validate that ratio for each participant.
Accuracy fell, but the pattern was not uniformly significant
The overall decline in enumeration accuracy was statistically significant (reported p = .007). Precision also worsened overall, with a reported alcohol effect of p = .033. The modest absolute accuracy difference matters: a three-percentage-point average change is not a diagnosis of severe impairment.
The authors interpreted the near-preserved performance on small displays and difficulties with larger displays as more consistent with reduced cognitive resources than with a simple loss of perceptual capacity. However, the alcohol-by-set-size interaction for accuracy was not significant (p = .437). Neither the alcohol-by-working-memory interaction (p = .619) nor the three-way accuracy interaction (p = .276) was significant.
These null findings limit the strength of a claim that alcohol selectively damaged one display size or definitively affected a particular working-memory subgroup more than another. Correlations between sober working-memory scores and large-display accuracy appeared in the alcohol condition, but that association does not establish a personal resistance to alcohol.
Why faster responses were an important surprise
The researchers expected alcohol to slow responses. Instead, participants responded faster to larger displays after drinking. The authors considered increased impulsivity or reduced persistence on the repeated task as possible explanations, while acknowledging uncertainty about the mechanism.
Speed and accuracy therefore pointed in different directions. A person can complete a response promptly without selecting the correct answer. Equally, the study does not show that every fast answer after alcohol is impulsive, or that slower responding necessarily indicates impairment.
This is related to, but distinct from, the broader issue of why passing one cognitive test cannot exclude effects in other domains. Here, the important contrast occurred between two outcomes within a single visual-number task.
Reporting limitations affect the weight of the evidence
The full paper contains an unresolved sample-count inconsistency. Its Methods section describes 40 participants, while the accuracy Results refer to 47 remaining after three exclusions. The reported analysis degrees of freedom do not resolve that discrepancy for a reader. The reported standard deviation for the converted alcohol values also appears inconsistent with the stated range.
These reporting problems are reasons to qualify reliance on the numerical details, not to silently substitute a corrected sample size or dispersion estimate. The participant population described in Methods consisted of young psychology undergraduates, predominantly women. Generalization to older drivers, clinical populations or people with different drinking histories requires additional evidence.
What this means when evaluating DUI observations
The forensic application is an inference about evidence: an observation that someone answered quickly cannot, by itself, establish that the answer was accurate or that complex visual processing was intact. An evaluator should identify the actual task, record errors as well as completion time, and distinguish an observed response from a reconstructed alcohol concentration.
The experiment included no on-road driving, crash outcome or validated roadside impairment classification. Counting flashed squares is not a substitute for driving, and its average changes cannot supply a person-specific BAC, an impairment threshold or a legal conclusion. This study supports careful interpretation of apparently normal response speed; it does not independently resolve a DUI case.
Primary source
Harvey AJ, Seedhouse M. Influence of Alcohol and Cognitive Capacity on Visual Number Judgements. Perception. 2021;50(1):39–51. https://doi.org/10.1177/0301006620984105.