Cross-Validation of the Self-Motivation Inventory

Adherence for up to 6 months to an exercise regimen is only about 50% among the general population (e.g., Haynes, 1979), suggesting that there are important individual differences among people embarking on a training program. One such difference is level of self-motivation as measured by the Self-Motivation Inventory (SMI; Dishman & Ickes, 1981). Self-motivation is defined as the tendency to engage in a behavior regardless of extrinsic reinforcement (Dishman & Ickes, 1981). Dishman and Ickes demonstrated that those with high scores on the SMI, and presumably with strong self-motivation, are more likely to adhere to an exercise program. Though there are mixed findings with the SMI (e.g., Ward & Morgan, 1984), others have reported that subjects often give poor motivation as a reason for dropping out of an exercise program (e.g., Oldridge, Wicks, Hanley, Sutton, & Jones, 1978). Since self-motivation is a potential predictor of exercise adherence that may be amenable to training, it is important to understand the correlates of this measure in order to establish its divergent and convergent validity and to direct the focus of any attempt to increase self-motivation and exercise adherence. Dishman (1982) suggests that self-reinforcement skills partly constitute the characteristics of self-motivation. One component of the self-reinforcement process (e.g., Fuchs & Rehms, 1977) is that of accurately evaluating one's own behavior. In the exercise setting, this may include accurately attributing the benefits of exercise to the exercise process. Several studies have shown that belief in the effects of exercise is related to adherence (e.g., Dishman & Gettrnan, 1980) and that an improvement in self-reinforcement skills improves exercise adherence (e.g . , Keefe & Blumenthal, 1980). Although the effects of health locus of control beliefs (i.e., taking responsibility for maintaining one's health) upon adherence has mixed support (e.g., Haynes, 1979), it follows that if individuals attribute health to factors beyond their control, then it is unlikely they would attempt to control such factors. Self-control training has also been found to reduce anxiety (Meichenbaurn, 1977) and depression (Fuchs & Rehm, 1977). Therefore the construct of