Our purpose in this study was to meta-analytically address several theoretical and empirical issues regarding the relationships between safety climate and injuries. First, we distinguished between extant safety climate-->injury and injury-->safety climate relationships for both organizational and psychological safety climates. Second, we examined several potential moderators of these relationships. Meta-analyses revealed that injuries were more predictive of organizational safety climate than safety climate was predictive of injuries. Additionally, the injury-->safety climate relationship was stronger for organizational climate than for psychological climate. Moderator analyses revealed that the degree of content contamination in safety climate measures inflated effects, whereas measurement deficiency attenuated effects. Additionally, moderator analyses showed that as the time period over which injuries were assessed lengthened, the safety climate-->injury relationship was attenuated. Supplemental meta-analyses of specific safety climate dimensions also revealed that perceived management commitment to safety is the most robust predictor of occupational injuries. Contrary to expectations, the operationalization of injuries did not meaningfully moderate safety climate-injury relationships. Implications and recommendations for future research and practice are discussed.