Multilevel Structural Equation Modelling in Marketing and Management Research

摘要

Multilevel (or mixed linear) modelling (MLM) either simultaneously tests hypotheses at several levels of analysis or controls for confounding effects at one level while testing hypotheses at another level. Advances in multilevel structural equation modelling (MSEM) enable the specification of latent variables, which are more common in marketing research than the manifest variables employed in hierarchical linear modelling (HLM), and open new conceptual possibilities. However, MSEM involves several challenges and is not frequently used. The authors therefore outline key methodological requirements, options, and challenges regarding MSEM and provide a systematic approach for its use. To achieve this goal, multilevel modelling and the advantages of MSEM over HLM are illustrated, followed by a literature review of marketing and management studies, to determine how MSEM and HLM are differentially applied. The requirements, options, and challenges of MSEM are systematically illustrated by elaborating current knowledge in the literature and by presenting an empirical study describing the sampling, measurement, and methodological issues for three model types: cross- level effects, cross-level interactions, and cross-level effects and interactions. Promising directions and major challenges for future research are examined.