Cognitive factors related to problem solving have been explored, but affective factors also play an important role in the teaching of mathematical problem solving. This paper outlines the theories of George Mandler and Bernard Weiner, providing a useful background for research related to affect and problem solving. Data related to the two theories are presented. Current issues on affect and problem solving are addressed: the role of belief systems in problem solving; the nature of metacognitive processes; and the influence of particular instructional settings, especially the role of technology in the teaching of problem solving. Finally, some methodological considerations for future research are presented. (MNS)