Objectives. There is a debate on the determinants of smoking behaviour, their relative impact, and how impacts are exerted. This longitudinal study is on the relations among social influence, intention to smoke, and smoking behaviour, controlling for attitude and self-efficacy. Design and Methods. A model combining parents and peers with subjective and descriptive norms, resulting in four factors, was used to assess social influence. Data were collected at the beginning of the 7th(-T1), 8th(-T2), and 9th(-T3) school years, concerning 578 students (Mage= 13.04 at T1). Structural Equation Modelling was used to test longitudinal effects. Results. Variances explained by the model were high: R2intention-T2= .65, R2behaviour-T2= .67, and R2behaviour-T3= .76. Longitudinal analyses confirmed the effects of social influence on intention and behaviour. These effects on behaviour were direct and indirect (peers’ and parents’ descriptive norms in both cases). Descriptive norms had a stronger effect on behaviour than subjective norms. Peers’ effect on behaviour was stronger than parents’, but peers’ effect was exerted only through descriptive norms while parents’ effect was exerted through both norms. The intention effect on behaviour was not as detached as expected and its role of full mediator between other variables’ effects on behaviour was not confirmed, since descriptive norms and self-efficacy had also a mediation role. Conclusions. Results show direct and indirect effects of social influence on behaviour. Descriptive norms are an important variable to operationalize social influence. Peers and parents exert influence on adolescents’ intention and behaviour through different processes. The impact of intention on behaviour is not as important as expected.

--

• Medicina Básica - Ciências Médicas