The lack of research to develop further a theory of evaluation is aglaring shortcoming for human resource development (HRD). In this paper, 1 argue that the four-level system of training evaluation is really a taxonomy of outcomes and is flawed as an evaluation model Research is needed to develop a fully specified and researchable evaluation model. Such a model needs to specify outcomes correctly, account for the effects of intervening variables that affect outcomes, and indicate causal relationships. I propose a new model based on existing research and accounts for the impact of the primary intervening variables such as motivation to learn, trainability, job attitudes, personal characteristics, and transfer of training conditions. A new role for participant reactions is specified. Key studies supporting the model are reviewed and a research agenda proposed.