A significant amount of research has been conducted on diversity "best practices" to understand whether they help reduce discrimination, increase managerial diversity, and enhance performance. Regardless, which diversity practices should be used, how they should be implemented, for what purpose, and to what effect remains unclear due to issues related to inconsistencies across outcomes studied within and between diversity practices as well as a lack of integration of micro research and macro research. In this review, these issues are revealed in an effort to make sense of the array of research conducted since the turn of the century. We first synthesize key research findings from sociological, firm-level research on diversity management practices with research conducted within the psychological tradition using experimental and within-organizational research designs. The results of our review provide the basis for a multi-level process model through which research findings can more easily be interpreted to uncover when and how diversity practices produce desirable results. We discuss how considering the interplay across levels of analysis is critical for enriching our theoretical understanding of the mediating and moderating mechanisms that link the distal constructs of diversity practices and organizational outcomes. We conclude with implications for future research that emerged from an analysis of the literature through our multi-level process model.