Personality Moderates Change In Exercise Self-efficacy And Descriptive Norms: A Longitudinal Observation

Evidence links personality to physical activity (PA) and PA related social cognitions commonly targeted in behavioral interventions. Frameworks of the personality system propose that traits moderate change in social cognitions which then influence behavior change. PURPOSE: This study aimed to assess whether personality traits moderate organic change in PA related social cognitions, and the relationship between those changes and changes in PA behavior. METHODS: A sample (n = 346) of female undergraduates completed the Big Five Inventory and participated in a 12-week longitudinal observation involving three surveys, spaced six weeks apart. Surveys included measures of exercise related self-efficacy (SE), perceived controllability, affective and instrumental attitudes, injunctive and descriptive norms (DN), and intentions, as well as PA outcomes measure by the GPAQ. Moderated mean level latent growth models tested the effect of the Big Five personality traits on change in social cognitive and behavioral variables. In the case of significant moderation, traits were quartile split, and RM-ANOVA was used to decompose significant effects of time for each quartile of the respective moderator. RESULTS: Significant linear change over time was observed for exercise SE (ηs = 0.25, p < .03) and DN (ηs = 0.21, p < .03). Change in SE was moderated by agreeableness (β = 0.24, p < .03; RMSEA [90%CI] = 0.09[0.03, 0.17], CFI = 0.96, SRMR = 0.03), whereas change in DN was moderated by extraversion (β = 0.20, p < .03; RMSEA[90%CI] <0.01[<0.01, 0.06],CFI > 0.99, SRMR<0.01). SE at week 12 was greater than at baseline (p < .05) and week 6 (p < .05), but only for those in the highest quartile for agreeableness (F[2] = 4.32, p < .02; all other quartiles, F[2] < 2.61,p > .07). DN was greater at week 12 than at baseline (p < .05) among those scoring above the median for extraversion (F[2] > 3.9, p < .03; below the median, F[2] < 1.5, p > .22). CONCLUSION: Personality moderated observed change in social cognitions. Tests of moderation of the relationships between change in social cognitive variables and change in PA by personality were precluded by the absence of change in PA. Future work should use randomized controlled designs to manipulate outcome variables and assess the possibility that personality moderates intervention effectiveness.