Despite the known health benefits of regular exercise, American youth and adults are becoming less physically active (U.S.D.H.H.S., 1996). Many attempts have been made to use theoretical frameworks to understand motivational processes and develop effective interventions in the physical activity context. Among them is self-efficacy theory proposed by Bandura (1986, 1997). However, while the role of self-efficacy in influencing physical activity behavior or behavioral intentions has received considerable attention, less research has been conducted with outcome expectations. Within self-efficacy theory, Bandura (1997) defined outcome expectations as a person's beliefs concerning the likely consequences of a behavior. These consequences are classified as physical outcomes of the behavior, social outcomes of the behavior, and self-evaluative reactions to the behavior. Although some researchers have studied the affect of different types of outcomes in exercise settings, more research is needed to identify specifically the outcome expectations that are most influential to physical activity behavior and intentions. Therefore, the purpose of this study was to investigate whether these three types of outcome expectations exert different influences on behavioral intentions and actual behavior in a physical activity class. A total of 137 college students (75women, 62men) enrolled in a beginning weight training class completed the modified outcome expectations scale and behavior intentions scale (Rodgers & Brawley, 1991) during the first week of class. Their attendance for the whole semester was recorded as a measure of their actual behavior. The exploratory factor analyses resulted in three factors (physical, social, and self-evaluative reaction outcomes) with eigenvalues greater than one, accounting for 65.2% of the variance. Cronbach alpha coefficients for these subscales were .76, .79, .78, respectively. It was found that these three types of outcomes were strongly associated with each other (r12=.70, r13=.72, r23=.74, respectively), and moderately positively related to behavioral intentions (r1=.45, r2=.40, r3=.37) and actual behavior (r1=.48, r2=.38, r3=.31) as well. Multiple regression analysis further revealed that the physical outcome expectations were the only significantly positive predictor for behavioral intentions (F=34.78, p<.001), and accounted for 20.49% of the variance. Further, behavioral intentions (F=16.28, p<.001) and physical outcome expectations (F=41.21, p<.001) were the significantly positive predictors for students' participation, explaining 31.69% of the variance. These findings provide empirical evidence on relationships among outcome expectations, intentions and behavior, and also shed new light on the importance of promoting positive physical outcome expectations when providing interventions for physical activity adoption.Keyword(s): college level issues, health promotion, physical activity