When the going gets tough: the "why" of goal striving matters.
Bottom Line: In Study 1, self-reported autonomous goal motives predicted goal persistence via challenge appraisals and task-based coping.In Study 2, primed autonomous (compared to controlled) goal motives predicted greater persistence, positive affect, and future interest for task engagement.The findings underscore the importance of autonomous motivation for behavioral investment in the face of increased goal difficulty.
Affiliation: University of Birmingham.Show MeSH
Mentions: We tested the hypothesized model with SEM, utilizing the single-indicator approach described in Study 1. This model showed excellent fit: χ2(4) = 1.35, p = .85, CFI = 1, NNFI = 1.19, RMSEA = .00, SRMR = .02 (Figure 3). Both the autonomous versus controlled (β = .38, p < .01) and the neutral versus controlled (β = .27, p = .02) contrasts predicted persistence, although the latter effect was possibly due to suppression (see the correlation between the neutral vs. controlled contrast and persistence reported in Table 2). Persistence predicted positive affect change (β = .42, p < .01), which consequently led to greater interest in future study participation (β = .47, p < .01). The hypothesized pathway from the autonomous versus controlled contrast to future interest was significant (β = .22, p = .02), but the pathway from persistence to future interest was not significant (β = −.07, p = .95). We obtained, however, an indirect effect of persistence on future interest via positive affect change (β = .20, p < .01, BC-CI = .07 to.32). We also obtained significant indirect effects from the autonomous versus controlled contrast (β = .16, p = .01, BC-CI = .09 to.49), and marginal effects for the neutral versus controlled contrast (β = .11, p = .06, BC-CI = .04 to.46) on positive affect change through persistence. In an exploratory analysis, we specified a pathway from the neutral versus controlled contrast to future interest; this was not significant (β = .10, p = .45) and had minimal impact on the model fit. The model remained unchanged when we added gender, hours of cycling, hours of training, goal difficulty, and efficacy as control variables. We depict the final model in Figure 3.