Overestimation of peer drinking: error of judgement or methodological artefact?

AIMS To examine whether inclusion of both self- and peer-referent items in the context of a single social norms drinking questionnaire plays an active role in producing the much-reported tendency for young people to overestimate the extent of peers' alcohol-related behaviour and the permissiveness of their attitudes towards alcohol. DESIGN, SETTING, PARTICIPANTS AND MEASUREMENTS In a between-subjects design pupils attending two Scottish secondary schools (n=1074; 12-18 years; 52.5% male) completed one of three questionnaires designed to measure a range of alcohol-related behaviours, attitudes and perceptions: a paradigmatic multiple-target questionnaire included self- and peer-referent items while two single-target questionnaires included self-referent or peer-referent items only. FINDINGS Pupils' self-reported drinking behaviours and attitudes were similar, regardless of whether multiple or single-target versions of the questionnaire were used, as were perceptions of peers' frequencies of alcohol use and drunkenness. In contrast, by comparison with pupils who responded to a single-target version that omitted self-referent items, use of a multiple-target questionnaire was significantly more likely to result in reports that peers would consume alcoholic drinks when with friends and hold more permissive or liberal attitudes towards alcohol. CONCLUSIONS Social norms research and related health promotion programmes that seek to reduce the extent of overestimation of peer drinking norms are heavily reliant upon multiple-target drinking questionnaires. The use of such a questionnaire may lead to more distorted or extreme perceptions being reported by pupils compared to single-target versions, which omit self-referent drinking items. By implication, use of multiple-target questionnaires may encourage young people to 'over-overestimate' peer drinking norms.