Glass, J., & Cumming, G. (2000). Incubation in problem solving. In C. Davis, T. van Gelder, & R. Wales (Eds.) Cognitive science in Australia 2000. Proceedings of the Fifth Biennial Australian Cognitive Science Conference (pp. I-133--I-141). Adelaide, SA: Causal Productions. (ISBN 1876346264)

Abstract:
Incubation is defined as an improvement in problem solving performance after temporarily putting a problem aside. This may occur either spontaneously or when the subject consciously returns to the particular problem. However only a handful of controlled experiments on incubation have been carried out providing some trends but few substantial findings. A meta-analysis was carried out to assess evidence for the existence of incubation and the possible role of a number of influencing variables. Only 15 papers were identified as having sufficient information to generate effect sizes for the meta-analysis. Given the variation in tasks, time allowed and other conditions, it was difficult to draw any firm conclusions. There were a number of substantial effect sizes, suggesting that incubation does exist. Overall the mean main effect size was found to be 0.16 (S.D of 1.23) (small), while the mean simple main effect size was found to be 0.48 (S.D of 0.53) (medium). However, there were insufficient data to be confident about the possible role of the potentially fascinating independent variables.