構造方程式モデリングの光と影

抄録

Structural Equation Modeling (SEM) has changed psychological investigation drastically in two ways. First, SEM serves not only as a statistical technique for estimating descriptive statistics and testing causal model hypotheses, but also as a powerful tool to check validity of a theory or even to construct a new theory. Second, thanks to SEM, researchers have begun to think of constructing large-scale and systematic research projects in order to benefit from the advantages of SEM. SEM is, however, not a theory itself, but a tool for theory construction. It is researchers' critical thinking that tests and evaluates a theory. Close collaboration of researchers and statisticians in a research project is desirable, or even necessary.