Path analysis or structural equation modeling is a technique for testing the consequences of proposed causal relationships among a set of variables. The technique rests on specific procedures and important assumptions (uncorrelated residuals, one-way causality, linearity, additivity, interval measures). These assumptions, along with the problem of multicollinearity, are discussed and specific techniques offered to deal with them. Path analysis studies from the industrial/organizational psychology literature are reviewed to illustrate the specific consequences of disregarding these assumptions and procedures. The relationship between cross-lagged correlation and path analysis is explored, particularly with regard to the differing but complementary purposes of each. Finally, the problems due to both shared and random measurement error and the application of path analysis to experiments are discussed.Path analysis is a technique that uses ordinary least squares regression to help the researcher test the consequences of proposed causal relationships among a set of variables. Used most specifically, path analysis can test an a priori causal hypothesis against a set of observed correlations. At the most general level, path analysis can be used to test a number of alternative causal sequences against one another. In any application of path analysis, very specific and important assumptions underlie the technique; if any of these assumptions are violated, the causal inferences will very possibly be incorrect.This article explores these assumptions in detail and offers specific advice about how to deal with them. Our purpose is not to add to the techniques of path analysis, but rather to translate the procedures, problems, and possible solutions into operational terms that can be more easily understood and used by industrial/organizational researchers. In doingThe authors would like to thank Tony Greenwald and two anonymous reviewers for very useful comments.

No abstract