In both Management and I/O Psychology, contributions to theory remain an important, and in many cases, sole criterion for evaluating submissions to top journals. In many ways, the definition of theory and the primacy of theory in the organizational sciences is an outlier
in most sciences, articles rarely even mention theories, much less build themselves around advancing theory. We propose that the classic description of the scientific methods provides a better guide to understanding the relationships between data, methods and theory than our current model, which often starts and ends with proposing a theory, which may never again be referenced or tested. We describe a pyramid of types of evidence that is useful for assessing the reliability and worth of particular sorts of data and show how this approach to evidence informs the scientific method and assists in identifying and building useful theories.