The causal predominance of performance affecting later cohesiveness that has been shown in previous studies was examined by means of a series of statistical analyses designed to assess influence in a longitudinal panel design. Male students (N = 44) participating in a basketball league were administered cohesiveness and participation motivation scales at early, mid, and late season. In contrast to previous findings, the cross-lagged correlations showed that performance and cohesion were significantly related to each other with no causal predominance of one over the other. With the exception of the friendship measure, the cross-lagged correlations were no longer significant when earlier measures of the effect variable were controlled through partial correlation and path analysis techniques. In contrast to previous research, midseason cohesion, as measured by friendship, was a significant (p < .04) predictor of late season performance. The importance of interpersonal attraction in the recruitment and maintenance of intramural team members is discussed along with the necessity for determining, in future studies, the reliability of cohesiveness measures.