Theoretical Constructs that Predict Women's Exercise
Although research has examined the determinants of physical activity, this research has focused primarily on men and few efforts have been made to explain the interrelationships between commonly used predictors of physical activity. Descriptive data and regression analyses were conducted with 193 female students, faculty, staff and community members of a southwestern Virginia university town. Variables that were entered into the regression included age, body mass index, exercise knowledge, self-esteem, depression level,self-efficacy, stage of change, exercise goals, outcome expectations and outcome values. Because of the high correlations between some of these variables, a principal components factor analysis was conducted. The factor analysis indicated significant overlap between items on the stage of change and knowledge measures with self-efficacy. Some knowledge items were dropped to create a more succinct measure and self-efficacy and stage of change were combined into one measure. Results indicated that the variables listed above significantly predicted physical activity level as measured by kilocalories expended per day on the Aerobic Center Physical Activity Questionnaire (Kohl, Blair, Paffenbarger, Macera, & Kronenfeld, 1988) for this sample of women (R2 = .346, p < .001), but that self-efficacy was the only significant single predictor. Additional regressions were conducted to examine the indirect relationships between these variables and physical activity level. The constructs of selfesteem, depression, and knowledge seemed to have indirect, rather than direct, effects on physical activity that were mediated by self-efficacy. Goals and outcome expectations also did not seem to directly predict physical activity level but were related to some of of the other variables. Implications of the interrelationships between these variables for planning physical activity interventions include enhancing self-efficacy for exercising while at the same time assessing for depressive symptomatology and using incentives and motivators that are age appropriate.