The relative effectiveness of alternative reward structures in strengthening group performance on a cooperative math task was investigated by operationalizing four reward structure treatments--one non-contingent pay treatment and three differential pay allocation treatments including two levels of inter-group competition and a group piece-rate pay contingency. In each of the four treatments eight male and eight female dyads worked on cooperative math tasks for 10 performance-pay trials. Although male dyads outperformed female dyads, group performance did not vary by pay allocation treatment, even though subjects in the non-contingent pay treatment indicated the pay system was less motivating than did subjects in the three differential pay allocation treatments. Although no main effects of treatment were indicated on various satisfaction measures, members of higher performing dyads reported higher satisfaction than members of lower performing dyads on several measures, particularly when higher performance was associated with higher pay. Feasible interpretations for why performance and differential rewarding were unrelated included uncontrolled or extraneous sources of motivation in the experimental setting and the brief duration of the work session. (Author)