There is a vast body of literature on school size but comparatively few high quality empirical studies comparing outcomes in schools of different sizes. This systematic review synthesizes the results of the published research from 31 studies on the effects of secondary school size from OECD countries since 1990. Overall the directions and patterns of effect vary for different outcomes. For pupil attainment measured by exam results, and for attendance, larger schools appear to do better up to some optimal school size but estimates of this point or range are insufficiently precise to be useful. The implications of different school sizes on student behaviours are equivocal, but teachers and pupils at smaller schools are more likely to have a positive perception of their "school environment". Costs per pupil appear to decrease as school size increases. The results of the review suggest that there is little empirical evidence to justify policies that aim to "change" or mandate particular school sizes. However, given the evidence that there do appear to be optimal sizes for some outcomes, stakeholders should be made aware that dramatic changes in a school's size may change the characteristics of a school's learning environment. (Contains 1 figure and 4 tables.)