Predicting smoking reduction among adolescents using an extended version of the theory of planned behaviour

Abstract

This study tested the ability of the theory of planned behaviour (TPB) to predict adolescents’ intentions to reduce smoking and the subsequent behaviour one year later. In addition, past behaviour (PB), moral norms, self-identity as a smoker, group identification, group norms and action planning were assessed. A prospective sample of 145 adolescents (M = 14 years, Time 1) participated in the study. The TPB provided good predictions of intentions (adjusted R ² = 0.28). An extended TPB model including self-identity, moral norms, and the group identification and group norm interaction accounted for 39% (adjusted R ²) of the variance in intentions. The TPB components did not have a direct impact on subsequent behaviour, while PB and the perceived behavioural control (PBC) intention interaction accounted for 35% of the variance in behaviour after one year. The practical implications of these results for the development of interventions to encourage adolescent smokers to reduce or quit smoking are outlined.

Keywords:

• Smoking reduction
• adolescents
• TPB
• social influence variables
• PB
• action planning

Acknowledgements

The financial support of the Research Council of Norway is gratefully acknowledged. The authors would also like to thank Reidar Ommundsen and two anonymous reviewers for valuable comments on an earlier version of this work.

Notes

Notes

[1] The results from 2000 are based on data (N = 24,127) from the “School Surveys”, which Braverman et al. (2001) also based their analysis on. The same questions have been administered in November every fifth year since 1975.

[2] Self-initiated smoking cessation refers to smokers who quit on their own, without involvement in a formal quit effort.

[3] We applied Royston's (1982) extension of the Shapiro and Wilk's W statistic to test whether the residuals were normally distributed. A Shapiro–Wilk score which is not significantly different from 1 indicates normality. The residuals from the regression analysis were normally distributed (Shapiro–Wilk score: 0.980, p = 0.06). We also tested whether the residuals were heteroscedastic (i.e., whether the variance in the residuals were associated with the predicted value) by making a scatterplot of the standardized predicted value of intention and the standardized residuals. The plots revealed that residuals were homoscedastic, supporting the use of parametric statistics (Hankins, French, & Horne, 2000).