EVIDENCE LIBRARY
Can Expecting Alcohol Change How People Value Gains and Losses?
A 75-student experiment separated monetary valuation from perceived probability. Its low alcohol target was calculated, not measured in the main participants.
Yes, an alcohol-like drinking context can affect ratings of hypothetical gains and losses, even when the beverage contains no alcohol. In a 75-student experiment, both a low-dose wine condition and a placebo condition differed from an explicitly nonalcoholic control on some valuation measures. That does not mean either condition established unsafe driving, or that alcohol changed every aspect of risk perception.
The most important exposure limitation is easy to miss: the study’s 0.02 g/100 mL blood alcohol target was calculated. Blood or breath alcohol was not measured in the main experimental participants.
Three beverage groups answered different kinds of questions
Tsurugizawa and colleagues randomly assigned healthy university students to wine, placebo or control groups. There were 23 participants in the wine group, 23 in the placebo group and 29 in the control group.
The wine contained 14% alcohol. The other groups received an alcohol-free, wine-flavored beverage. In the placebo condition, a real labeled wine bottle provided alcohol-related cues, and the researchers reported that participants believed they had consumed wine. The control participants were told their beverage was nonalcoholic.
About 45 minutes after drinking, participants completed paper-and-pencil ratings. Some questions concerned how likely a hypothetical gain or loss seemed. Others concerned perceived waiting time until an outcome, or the subjective value of gaining or losing different amounts of money. These are related psychological concepts, but they are different measured outcomes.
Valuation changed without a general beverage effect on probability ratings
For hypothetical amounts from 10,000 through 50,000 yen, subjective valuation ratings in the alcohol group were significantly lower than in the control group. The placebo group also had lower ratings than control for the two smallest amounts, 10,000 and 20,000 yen.
The investigators additionally fitted a mathematical power function to the ratings. Its valuation exponent differed significantly between placebo and control, while the alcohol-versus-control comparison did not show that same significant effect. This model result should not be expanded into a claim that placebo always has a larger effect than alcohol.
The beverage groups did not show a significant general effect on subjective likelihood or waiting-time ratings. A finding about the value assigned to an outcome therefore cannot simply be restated as increased perceived crash risk, reduced perceived risk or a greater willingness to drive.
The stated alcohol level was an estimate, not a specimen result
The wine dose was tailored using body weight, sex and a Widmark calculation to target 0.02 g/100 mL. That is a target of 0.02%, not a measured BAC in these participants.
Preliminary breath testing had been performed separately. The investigators omitted breath testing during the main experiment to avoid revealing the placebo condition. They expressly acknowledged uncertainty about whether the intended alcohol level was actually achieved.
This limits concentration-specific interpretation. The experiment cannot establish a measured-BAC threshold at which monetary judgment changes, compare blood with breath at that threshold, or reconstruct an individual’s concentration at driving.
Sex differences require separate interpretation
Some ratings differed between women and men. For example, women rated low-probability losses as more likely than men did. Such group differences are not evidence that alcohol caused the difference, particularly when the beverage effect on that outcome was not significant.
The sex-by-beverage subgroups were small: the wine group included nine women and the placebo group eight. These results do not supply a reliable prediction of a particular person’s decisions based on sex.
What this contributes to a DUI evidence review
The experiment supports keeping three questions separate: what alcohol exposure was established, what the person believed they consumed, and what behavior was actually measured. Here, the behavioral evidence concerned hypothetical money ratings. There was no driving task, crash outcome or person-specific impairment assessment.
Nor can an expectancy effect explain away a valid chemical alcohol measurement. Expectancy concerns psychological responses; the presence and concentration of ethanol remain analytical questions.
A related 64-person placebo study of lightheadedness measured a different outcome and included repeated breath testing. Reading the two studies together makes clear why reported sensations, monetary judgments and measured alcohol exposure should retain their own evidential meaning.
Primary source
Tsurugizawa T, Tokuda S, Harada T, Takahashi T, Sadato N. Pharmacological and Expectancy Effects of a Low Amount of Alcohol Drinking on Outcome Valuation and Risk Perception in Males and Females. PLOS ONE. 2016;11(4):e0154083. https://doi.org/10.1371/journal.pone.0154083.