Medical Education 2011: 45: 792–806
Objectives In this review, we portray the process of problem-based learning (PBL) as a cognitive endeavour whereby the learner constructs mental models relevant to problems. Two hypotheses are proposed to explain how learning is driven in PBL; an activation–elaboration hypothesis and a situational interest hypothesis.
Methods Research relevant to these hypotheses is discussed. In addition, research studying the effects of various support strategies used in PBL is reviewed. Finally, we summarise a number of recent studies in which a new ‘micro-analytical’ methodology was used to trace the process of PBL in the natural classroom setting.
Conclusions We conclude that there is considerable support for the idea that PBL works because it encourages the activation of prior knowledge in the small-group setting and provides opportunities for elaboration on that knowledge. These activities facilitate the comprehension of new information related to the problem and enhance its long-term memorability. In addition, there is evidence that problems arouse situational interest that drives learning. Flexible scaffolding provided by cognitively and socially congruent tutors also seems to be reasonably effective, as opposed to ‘hard’ scaffolding represented by, for instance, worksheets or questions added to problems. Small-group work protects against dropout and encourages students to study regularly. Initially, students do not study much beyond the learning issues generated; the development of personal agency in self-study needs time to develop. The extent of learning in PBL results from neither group collaboration only (the social constructivist point of view) nor individual knowledge acquisition only; both activities contribute equally to learning in PBL.