Over five decades of research has yielded mixed findings as to whether groups make more risky or more cautious decisions than individuals. We contrast four theories that have been used in the extant literature to explain such a shift in group- as opposed to individual decision making. By means of a meta-analysis of 33 articles and 126 effect sizes, we were able to evaluate the evidence for the occurrence of shifts while considering the potential influence of the various conditions. Our results indicate an overall tendency for a risky shift, which is also affected by a number of conditions. Specifically, the occurrence of a shift is affected by the extent to which members are exposed to each others' preferences, the culture where the study is conducted, the extent to which decision makers are affected by their decision, and the extent to which a task is risk- or caution-oriented. We discuss to what extent our findings support each of the four theories, propose an integrative perspective, and set an agenda for future research.