EVIDENCE LIBRARY
Did Alcohol Increase Laboratory Risk-Taking and the Urge to Drink?
A randomized 142-student experiment found more balloon-task risk-taking and a greater rise in alcohol urge after drinking. Statistical mediation did not establish the causal sequence or predict a driver’s decisions.
Alcohol increased both balloon-task risk-taking and reported urge to drink in a randomized student experiment. The relationship is relevant to how drinking decisions can change after initial consumption. It does not establish that a laboratory risk score predicts a specific driver’s decision to drive, a crash or loss of control over drinking.
A task with hypothetical rewards
Rose and colleagues randomized 142 university students to alcohol or placebo. The alcohol dose was 0.6 g/kg. Mean breath alcohol concentration before and after the task was approximately 0.36 mg/L, a direct breath unit that should not be relabeled as g/dL blood.
Participants inflated simulated balloons. Each pump added hypothetical money, but a burst erased the temporary reward for that balloon. Participants could stop and collect the accumulated reward. The primary measure was the average number of pumps on balloons that did not burst.
That adjusted score deliberately excludes burst trials. It measures behavior under this task’s incentives and uncertainty, rather than every possible form of risk-taking.
The randomized comparison showed an average effect
The alcohol group averaged 33.37 adjusted pumps, compared with 28.68 in the placebo group. The difference was statistically significant, p = 0.019, with a reported standardized effect size of d = 0.40.
Urge to drink increased in both groups, but increased more after alcohol. The between-group comparison of urge change was significant, p < 0.001, d = 0.81. Expectancy and the experimental setting may therefore have contributed alongside the alcohol exposure.
There was no significant sex effect or alcohol-by-sex interaction on task performance. That null result does not prove that every subgroup responds identically, but it does not support a sex-specific rule from this experiment.
Associations with prior drinking were not forecasts
After accounting for self-reported impulsivity and sensation seeking, balloon-task performance in the alcohol group was associated with reported weekly consumption and binge frequency. Corresponding relationships were not significant in the placebo group.
The drinking measures described participants’ reported habits. The study did not follow them prospectively to establish who later developed hazardous drinking, drove after drinking or experienced a collision. The statistical word “predict” in a regression should not be mistaken for a validated forecast of future events.
Why mediation did not establish the causal chain
The authors’ mediation analysis was consistent with a statistical pathway connecting alcohol, balloon-task behavior and increased urge. However, urge was assessed before the balloon task in the experimental sequence.
Therefore, the study did not demonstrate that the measured task behavior occurred first and caused the previously recorded urge change. Randomizing alcohol strengthens the alcohol-versus-placebo comparison; it does not randomize the proposed mediator or establish its causal direction.
A shared alcohol response, unmeasured individual differences or a different temporal relationship could contribute. The mediation result is a hypothesis about mechanism rather than proof that risk-taking caused additional drinking.
Limits for DUI interpretation
The participants were young students, with average drinking scores already indicating substantial alcohol involvement. Rewards were hypothetical, and each participant completed only one treatment condition. There was no sober baseline balloon score for the same person and no driving task.
The experiment supports an average alcohol effect on a specified laboratory choice, not an individual conclusion about intent, recklessness or why a driver acted. It cannot infer BAC from risky behavior or establish that one risky choice proves intoxication.
The site’s study of voluntary stopping and response inhibition explains why different behavioral-control tasks need separate interpretation. Its real-car parking experiment provides a different, directly driving-related outcome.
Primary source
Rose AK, Jones A, Clarke N, Christiansen P. Alcohol-induced risk taking on the BART mediates alcohol priming. Psychopharmacology. 2014;231:2273–2280. https://doi.org/10.1007/s00213-013-3377-1.