Purpose: Participating in an organized physical activity class or group exercise program is an important vehicle to develop good exercise habits and attitudes toward exercise (Dishman, 1988). Little research, however, has investigated the mechanism of engagement in group exercise programs (Spink & Carron, 1994). Given that the most previous studies of exercise programs focus on individual factors, more research is needed to examine the influence of social constructs on exercise adherence in group settings. Group cohesion has been identified as an important small group variable. This variable includes four related factors that bind members to their group: Individual Attractions to Group-Task (ATG-T), Individual Attractions to Group-Social (ATG-S), Group Integration-Task (GI-T) and Group Integration-Social (GI-S) (Carron et al., 1985). Although group cohesion has been investigated in exercise settings, sparse research has focused on the relationship between cohesion constructs and students' achievement outcomes such as perceived effort and attendance in exercise domain. The main objective of the current study was to examine how measures of group cohesion during the early stages of group development affect subsequent adherence behavior in college female students participating in structured exercise classes.

Methods: Participants were 124 female college students (M age = 21.37) enrolled in aerobics classes at a southeastern university. They completed previously validated questionnaires during the sixth week (group cohesion constructs; Blanchard et al., 2000) and thirteenth week (perceived effort; McAuley & Duncan, 1989) of the 13-week class. The class met three times per week, and attendance was monitored daily. Pearson correlations and multiple regression analyses were used to assess the effects of group cohesion on students' perceived effort and exercise attendance during the program.

Analysis/Results: Pearson correlations indicated that group cohesion correlated with perceived effort and exercise attendance. Multiple regression analyses revealed that ATG-S and GI-T were positive predictors of perceived effort (β = .25, p= .006; β = .22, p = .015), accounting for 9.9% and 4.3% of the variance respectively. For exercise attendance, ATG-T was a positive predictor (β = .36, p= .001), explaining 4.4% of the variance. In contrast, ATG-S was a negative predictor (β = - .28, p= .007), explaining 5.6% of the variance.

Conclusions: The findings show that specific measures of group cohesive were associated with adherence behavior in exercise settings involving female college students. They provide insight into how to design a group environment aimed at the promotion of exercise engagement for female college students.