This study investigated the effects of working in dyads and their associated gender composition on performance (solution rate and time) and process variables (number of impasses, number of passed solutions, and number of problem solving suggestions and interactions) in a set of classic insight problem solving tasks. Two types of insight problems were used: multiple moves available (MMA) and few moves available (FMA) problems. The results showed no effects on number of problems solved or solving time, but group and gender composition effects emerged when looking at process variables such as number of impasses, time to impasse, number of passed solutions, number of problem solving suggestions, and number of interactions between dyad members. Results support prior evidence that process loss in creative problem solving may be due to differences in interpersonal interactions during group problem solving that do not necessarily affect overall problem solving success.