EFA, CFA and SEM Applied to Consumer Brand Relationships

Michèle Boulanger, PhD, Assistant Professor, Rollins College
Marc Fetscherin, PhD, Associate Professor, Rollins College

This talk describes the statistical components of a study designed to relate the concepts of brand love to consumer behavior. Our sample consisted of 800 respondents who answered questions regarding their preferred brand and their relationship to it. An exploratory factor analysis (EFA) was conducted to extract the underlying structure in the data and compare it to the structure developed in previous marketing studies. The measurement validation process followed. It is a multistep approach that assessed the validity of each construct from content, convergent, discriminant, and nomological points of view. Construct reliability was also assessed. A confirmatory factor analysis (CFA) was performed for that purpose. We used JSL to calculate some of the metrics not provided directly as standard output from SAS or JMP. The measurement validation process led to changes in the initial measurement system to create the final constructs of brand love, brand loyalty, word of mouth and purchase intention. We conclude this talk with the structural equation modeling (SEM) analysis conducted to test our hypotheses. We draw conclusions regarding the impact that some characteristics have on type and intensity of relationships between brand love and brand loyalty constructs and consumers' behaviors, and we review the consequences our results have for marketing.