Although mediational designs and analyses are quite popular in Organizational Behavior research, there is much confusion surrounding the basis of causal inferences. We review theoretical, research design, and construct validity issues that are important for drawing inferences from mediational analyses. We then distinguish between indirect effects, and partial and full mediational hypotheses and outline decision points for drawing inferences of each type. An empirical illustration is provided using structural equation modeling (SEM) techniques, and we discuss extensions and directions for future research. Copyright © 2006 John Wiley & Sons, Ltd.