School inspections are fundamental to quality assurance and improvement in many countries. Yet the question of the validity of these inspections has not yet been comprehensively addressed. This paper proposes a systematic approach to assessing the validity of school inspections, based on Messick's multifaceted concept of validity. We apply each aspect of Messick's concept to the process of school inspections, surveying research to identify existing (or potential) ways that school inspections might be validated in each case. We uncover several unresolved questions, concluding that empirical data are lacking for nearly all of the aspects of validity addressed by Messick. This contrasts strikingly with both the widespread implementation of school inspection internationally and the importance accorded the results for improvement processes. (Contains 1 figure and 4 tables.)