An extended version of the theory of planned behavour: Prediction of intentions to quit smoking using past behaviour as moderator

Abstract

It was hypothesised that: (i) intentions to quit smoking were predictable from group identity, self-identity, moral norms and past quit attempts, beyond the components of the Theory of Planned Behaviour (TPB), and (ii) that past experiences with the behaviour (quit attempts) would increase the predictive utility of the extended TPB model on intentions. The data was collected among 357 daily smoking students (M = 24 years). The TPB components accounted for 12.3% of the variance in quitting intentions, while the extension variables added 16.5% to the explained variance in intentions beyond the impact of the TPB. Past behaviour had the strongest impact on intentions (α = 0.30), followed by moral norms (α = 0.25), perceived behavioural control (PBC, α = 0.20), attitude (α = 0.18) and group commitment (α = −0.11). By splitting the sample into three categories of past quit attempts the picture changed, revealing that the predictive utility of the TPB increased with the number of quit attempts (no past quit attempt, R² = 1%, ns and several past quit attempts, R² = 12.3%, p < 0.001). Moreover, PBC and group commitment were significantly stronger predictors of intention among those who had several previous quit attempts compared with those who never had tried to quit smoking. The extended TPB model explained 1% and 28% of the variance in intentions among those with no and several previous quit attempts, respectively. The practical implications of these results for the development of interventions to encourage smokers to quit smoking are outlined.

• TPB
• past behaviour
• group identification
• moral norms
• intention to quit smoking

Notes

Notes

1. We thank an anonymous reviewer for this idea.

2. (1) A cognitive component which reflects the awareness of the membership of a social category (i.e. self-categorisation), (2) an evaluative component which represents the value attached to the group (i.e. group esteem) and (3) an emotional involvement in the group (i.e. group commitment).

3. Three factors with eigenvalue above 1, explaining 62% of the variance among the items emerged, roughly corresponding with Ellemers’ components. Cronbach's alpha was 0.79 for the cognitive component, 0.61 for the evaluative component and 0.51 for group commitment. In order to increase Cronbach's alpha for the last mentioned component item h was removed (r = 0.75).

4. We thank an anonymous reviewer for this idea.