any procedure used to test for cause-and-effect relationships (as opposed to mere correlation) between multiple variables. Structural equation modeling is the best known example. For a causal model to be a useful evaluative tool, strict conditions concerning the measurement of the variables must be met.