Social cognitive determinants of exercise behavior in the context of behavior modeling: a mixed method approach

Research has shown that persuasive technologies aimed at behavior change will be more effective if behavioral determinants are targeted. However, research on the determinants of bodyweight exercise performance in the context of behavior modeling in fitness apps is scarce. To bridge this gap, we conducted an empirical study among 659 participants resident in North America using social cognitive theory as a framework to uncover the determinants of the performance of bodyweight exercise behavior. To contextualize our study, we modeled, in a hypothetical context, two popular bodyweight exercise behaviors - push ups and squats - featured in most fitness apps on the market using a virtual coach (aka behavior model). Our social cognitive model shows that users' perceived self-efficacy (β_T = 0.23, p < 0.001) and perceived social support (β_T = 0.23, p < 0.001) are the strongest determinants of bodyweight exercise behavior, followed by outcome expectation (β_T = 0.11, p < 0.05). However, users' perceived self-regulation (β_T = -0.07, p = n.s.) turns out to be a non-determinant of bodyweight exercise behavior. Comparatively, our model shows that perceived self-efficacy has a stronger direct effect on exercise behavior for men (β = 0.31, p < 0.001) than for women (β = 0.10, p = n.s.). In contrast, perceived social support has a stronger direct effect on exercise behavior for women (β = 0.15, p < 0.05) than for men (β = -0.01, p = n.s.). Based on these findings and qualitative analysis of participants' comments, we provide a set of guidelines for the design of persuasive technologies for promoting regular exercise behavior.