Dutch Institute for Alcohol Policy STAP; 2-11-2026
Alcohol increases the risk of cancer and liver disease. A warning on the packaging or bottle can alert consumers to this risk. But which type of warning works best?
Researchers from the University of Amsterdam, led by Prof. Dr. Bas van den Putte, conducted a study—commissioned by the Dutch Institute for Alcohol Policy (STAP)—to examine various warning messages, with and without pictograms, among 1,027 Dutch adults (aged 18 and older) who consume alcohol.
The study was conducted as an online experiment. Participants were shown an image of a bottle of the alcoholic beverage they reported consuming most frequently: beer, wine, spirits, or mixed drinks. The front of the bottle featured a warning consisting of one of two different text messages, either with or without one of two different hazard pictograms. Additionally, some participants were shown only one of the two pictograms, while another group was shown neither a text message nor a pictogram. After viewing the bottle, all participants answered a series of questions.
The text messages used were: “Alcohol consumption can cause liver damage” and “Alcohol can cause cancer, including breast and colon cancer.” The two hazard pictograms were: an exclamation mark inside a red triangle—a symbol used across Europe to indicate a “health risk”—and a human silhouette inside a red triangle, representing a “health hazard.”
Key findings:
• The clearest outcome of the study is that warnings containing text—regarding the risk of liver damage or cancer—were understood significantly better than warnings consisting solely of a pictogram.
• Hazard pictograms alone are insufficient; however, the exclamation mark performed better than the human silhouette. Participants shown the pictogram with an exclamation mark were more likely to perceive the consumption of alcohol as less normative. The human silhouette pictogram was not correctly identified by the majority, which limits its potential effectiveness.
• A warning regarding liver damage led to a slightly higher perceived risk than a warning regarding cancer. Beer drinkers were more responsive to the liver damage warning than wine drinkers; for them, such a warning evokes a stronger sense of risk.
• Women also responded differently to a liver damage warning than men. After viewing a liver damage warning, women perceived alcohol consumption among family and friends as less normative and less accepted than they did when no warning was presented.
• The reception of the warnings was positive across the panel. In all cases, participants were more in favour than opposed placing health warnings on alcoholic beverages. There were no differences in this regard between the specific types of warnings. No significant effects were found—neither regarding the type of pictogram or text, nor their interaction—on the participants’ intention to reduce their alcohol consumption.
With support from Us Blau Hiem, STAP was able to commission this study. Wim van Dalen, director of STAP, finds the results promising—particularly the fact that the warnings are well-received across the panel. Van Dalen states: “It is clear that a warning on bottles must always include text; the two hazard pictograms tested here are not sufficient on their own.” Consequently, Van Dalen suggests exploring whether other, more effective pictograms might be used. “The message is being accepted,” he notes, “so now we need to identify the pictograms that work best.”
Mollen, S., Smit, E., Dyment, G., Van den Putte, B. From Label to Effect: An Experimental Study on the Impact of Alcohol Warning Symbols and Texts on Consumer Responses. University of Amsterdam, Department of Communication Science, 2026.