This article briefly summarizes findings from a review of 95 empirical studies of games used in instruction. The article suggests that such efforts are best assessed as transfer from game play to performance on external tasks that are targeted by the instruction. Review findings suggest that such transfer may be expected only if the cognitive processes engaged by games and external tasks overlap. Integrating games into a course of study is likely to facilitate such transfer. Research on improvement in cognitive processes as a result of playing "first-person shooter" games is briefly overviewed, and suggestions for similar research not using aggressive content are made. Minimal overlap between this and another research review of the effects of games used in instruction is discussed, and the need for generally accepted definitions and a taxonomy of games is noted.