More evidence that perceived behavioural control is multidimensional - A consideration of physical activity

Recently Trafimow and colleagues (2002) presented evidence that perceived behavioural control (Ajzen, 1988, 1991) is multidimensional, comprising two constructs: perceived control (PC) and perceived difficulty (PD). They determined this difference using non-health relevant behaviours. The purpose of this study was to replicate the findings of Trafimow et al. using their reading example and to extend their work by also considering self-efficacy (Bandura, 1986, 1995), and by examining the distinction of these three constructs in the context of exercise behaviour. Trafimow et al. examined reading 1, 30 and 100 pages to represent different levels of behavioural challenge that should reflect increases in perceived difficulty but not necessarily perceived control. In a repeated measure within-subjects design, with a sample of over 220 undergraduate students participating for course credit, we were able to replicate their findings, demonstrating a significant interaction of PC, PD, and SE such that the decline in PD was steeper than for PC, and the decline in SE was steeper than PD, F(4.214) = 50.13, < 0.001, ES = 0.48. We were also able to reproduce this distinction among the constructs contrasting three exercise frequencies: 2, 4, and 6 days per week, with a significant interaction, F(4.217) = 24.29, p< 0.001, ES = 0.31. These results indicate the conceptual and empirical distinction among constructs and between three levels of two behaviours for which perceptions of difficulty, control and efficacy should differ, is supported. The potential role of each control construct in the production of discrete versus long-term sustained behaviour is presented.