This is an overview of what the authors believe to be every empirical research article into the linkages between HRM and performance published in pre-eminent international refereed journals between 1994 and 2003. The analysis covers the design of the study, including the primary level of analysis and the identity of the respondents; the dominant theoretical framework(s) informing the article; how HRM is conceived and operationalised; how performance is conceived and operationalised; and which control and/or contingency variables are incorporated. Finally, the article examines how each study depicts the so-called 'black box' stage between HRM and performance. It reports wide disparities in the treatment of these components, but also some welcome commonalities and indicative trends that point towards a gradual convergence on how future research into this complex relationship might usefully be conducted. The findings are compared with previous reviews of the literature. The analysis should illuminate the ongoing debate about the linkages between HRM and performance, and prove valuable for future research designs.