Interaction Effects in Multiple Regression

Dr. James Jaccard is Professor of Social Work at New York University Silver School of Social Work. He received his doctoral degree from the University of Illinois, Urbana, in 1976. Dr. Jaccard’s research focuses on adolescent and young adult problem behaviors, particularly those related to unintended pregnancy and substance use, broadly defined. He has developed parent-based interventions to teach parents how to more effectively communicate and parent their adolescent children so as to reduce the risk of unintended pregnancies and problems due to substance use. Dr. Jaccard has written numerous books and articles on the analysis of interaction effects in a wide range of statistical models and teaches advanced graduate courses on structural equation modeling. He has written influential articles on the issue of arbitrary metrics in social science research. Dr. Jaccard also has written about theory construction and how to build conceptual models in a book published by Guilford Press.

• Series Editor′s IntroductionPrefaceChapter 1: IntroductionThe Concept of InteractionSimple Effects and Interaction ContrastsSimple EffectsInteraction ContrastsA Review of Multiple RegressionThe Linear ModelHierarchical RegressionCategorical Predictors and Dummy VariablesPredicted Values in Multiple RegressionTransformations of the Predictor VariablesOverview of BookChapter 2: Two-Way InteractionsRegression Models with Product TermsTwo Continuous PredictorsThe Traditional Regression StrategyThe Form of the InteractionInterpreting the Regression Coefficients for the Product TermInterpreting the Regression Coefficients for the Component TermsSignificance Tests and Confidence IntervalsMulticollinearityStrength of the Interaction EffectA Numerical ExampleGraphical PresentationA Qualitative Predictor and a Continuous PredictorA Qualitative Moderator VariableA Continuous Moderator VariableMore Than Two Groups for the Qualitative VariableForm of the InteractionSummaryChapter 3: Three-Way InteractionsThree Continuous PredictorsQualitative and Continuous PredictorsA Continuous Focal Independent VariableA Qualitative Focal Independent VariableQualitative Variables with More than Two LevelsSummaryChapter 4: Additional ConsiderationsSelected IssuesThe BiLinear Nature of Interactions for Continuous VariablesCalculating Coefficients of Focal Independent Variables at Different Moderator ValuesPartialing the Component TermsTransformationsMultiple Interaction EffectsStandardized and Unstandardized CoefficientsMetric PropertiesMeasurement ErrorRobust Analyses and Assumption ViolationsWithin-Subject and Repeated-Measure DesignsOrdinal and Disordinal InteractionsRegions of SignificanceConfounded InteractionsOptimal Experimental Designs and Statistical PowerCovariatesControl for Experimentwise ErrorsOmnibus Tests and Interaction EffectsSome Common MisapplicationsInteraction Models with Clustered Data and Random Coefficient ModelsContinuous Versus Discrete Predictor Variables The Moderator Framework RevisitedReferencesNotesAbout the Authors