Residual centering, exploratory and confirmatory moderator analysis, and decomposition of effects in path models containing interactions

Abstract

Hierarchical moderated regression (HMR) analysis may lead to interpretational problems in tests of moderator (interaction) hypotheses. An alternative, residual- centering approach is described and compared to traditional HMR analysis. Procedures for evaluating interaction hypotheses and general effect analysis procedures are described for path (causal) models containing interactions. Index terms: Confirmatory analysis, Effect analysis in path models, Goodness-of-fit tests, Hierarchical moderated regression, Mediators, Moderated regression, Multicollinearity, Path models, Residual centering.