Campaigns designed to stop young people "bolting" drinks can be ineffective and can even make them more likely to do it, new research suggests.
Scientists from the University of Exeter and the University of Queensland examined reactions to a poster warning of the consequences of bolting (downing an alcoholic drink in one) and found it had no effect on people's future intentions.
And when a statement was added saying other people disapproved of bolting, study participants reported stronger intentions to bolt in the future.
However, changing this to a message saying most people "do not bolt drinks on a night out" was effective.
"Many young people overestimate the extent to which their peers both approve of and engage in risky drinking behaviours," said study author Dr Joanne Smith, of the University of Exeter.
"One way to tackle risky drinking is to try to correct these misperceptions through health campaigns, such as posters.
"In our research, we wanted to explore what kinds of messages are more effective in changing people's intentions to bolt.
"Our results highlight the potentially harmful effects of exposure to what's called an 'injunctive norm' - a message about the approval or disapproval of others.
"Meanwhile, a 'descriptive norm' - telling people what others do rather than what they think - had a positive impact."
The study is published in the journal Addiction Research and Theory.
Professor Charles Abraham, of the University of Exeter Medical School, said: "This demonstrates how careful we need to be in selecting the right message in campaigns, and evaluating them before wider dissemination, as poorly designed campaigns, however well-intentioned, can backfire."
The research consisted of three studies, in which volunteers (221 in total) saw the poster or did not, and then either received or did not receive messages about what their peers thought or how they behaved.
In one study, some participants received an accurate message saying 70% of their peers "disapprove of bolting", and in another some received an accurate message saying 65% of their peers "do not bolt drinks on a night out".
They all then completed identical questionnaires to measure their perceptions of group norms related to bolting, and their own intentions to do it in the future.
The researchers point out that beliefs about how other people behave are often the "best predictor" in terms of general drinking behaviour and binge drinking, but note that using these beliefs to change behaviour needs to be done carefully to ensure campaigns have the desired effect.
The paper is entitled: "Normative feedback in alcohol-related health promotion: When and how does normative feedback reduce intentions to drink irresponsibly?"
The study was funded by the Australian Research Council.