Generalized Measurement Invariance Tests with Application to Factor Analysis

摘要: The issue of measurement invariance commonly arises in factor-analytic contexts, with methods for assessment including likelihood ratio tests, Lagrange multiplier and Wald tests. These tests all require advance definition the number groups, group membership, offending model parameters. In this paper, we construct based on stochastic processes casewise derivatives function. can be viewed as generalizations test, they are especially useful for: (1) isolating specific parameters affected by violations, (2) identifying subgroups individuals that violated a continuous auxiliary variable. presented illustrated detail, along simulations examining tests' abilities controlled conditions.