Servers inadvertently pour 20 to 30 per cent more alcohol into short, wide glasses than into tall, slender ones of the same volume, even if they are professional bartenders.

Studies show two perceptual biases lead to the problem:

• Servers perceive tall glasses to hold more liquid than wide glasses.
• Attention is focused on the height of the liquid without compensating enough for the width of the glass.

Researchers in the United States designed an experiment to see if college students and bartenders pour different amounts into short, wide tumblers compared with highball glasses.

Study participants prepared gin and tonic, vodka tonic, rum and Coke and whiskey on the rocks.

The 198 college students poured drinks from liquor bottles filled with water or tea instead of alcohol into tall and short glasses.

After 10 rounds of practice at pouring four mixed drinks, the students reduced their tendency to overpour into tall glasses but not into short, wide glasses.

The team also tested 82 bartenders who had an average of six years of experience. They also poured 20 per cent more into short, wide glasses.

Bias difficult to overcome

When the experimenter encouraged bartenders to "please take your time" before pouring each drink, they took twice as long to pour but the bias effect was not eliminated, according to the study.

The results are important for policymakers, law-enforcement officials and groups aiming to promote responsible drinking, as well as those in the hospitality industry who want to reduce costs without decreasing customer satisfaction, the researchers found.

"If short tumblers lead even bartenders to pour more alcohol than tall highball glasses, the way to better control alcohol consumption is to use tall glasses or to use glasses with the alcohol level marked on them," the researchers wrote in the light-hearted holiday issue of British Medical Journal.

The study's authors, Georgia Tech marketing professor Koert van Ittersum and Brian Wansink, a professor of marketing, applied economics and nutrition science at Cornell University, also suggest alcohol consumption studies should include questions about the shape of the glass.