To maximize the impact of research on programs, this article proposes a ‘reaching-down—reachingup’ perspective in evaluation design, whereby it serves two functions simultaneously: the program improvement function, reaching down, and the knowledge development function, reaching up. This proposal frames applied research as a particular species of evaluation. As validity is a fundamental assessment of how well a design supports particular inferences related to specific uses, the article subsequently summarizes and integrates disparate validity perspectives to develop an argument-based approach to validity in a ‘reaching-down—reaching-up’ evaluation model case study. Lastly, the location(s) of validation is considered in order to highlight evaluator responsibilities.