Abstract
In this article, we review applications of covariance-based structural equation modeling (SEM) in the Journal of Advertising (JA) starting with the first issue in 1972. We identify 111 articles from the earliest application of SEM in 1983 through 2015, and discuss important methodological issues related to the following aspects: confirmatory factor analysis (CFA), causal modeling, multiple group analysis, reporting, and guidelines for interpretation of results. Moreover, we summarize some issues related to varying terminology associated with different SEM methods. Findings indicate that the use of SEM in the JA contributes greatly to conceptual, empirical, and methodological advances in advertising research. The assessment contributes to the literature by offering advertising researchers a summary guide to best practices and a reminder of the basics that distinguish the powerful and unique approach involving structural analysis of covariances.
SUPPLEMENTAL DATA
Supplemental data for this article can be accessed at www.tandfonline.com/ujoa.
Notes
1. We limit discussion to the software used by the JA authors. AMOS and LISREL are most widely used and are available for use by purchasing a license (as are Mplus and EQS). Many other programs exist now including SEM packages within R, which are free to use. Perhaps the most promising is Lavaan (latent variable analysis).
2. Convergent validity, the extent to which multiple measures converge on a consistent meaning, and discriminant validity, the extent to which a measure is unique and not confounded by another, are both necessary elements of the broader concept of construct validity, the extent to which a measure truly represents a construct.
3. We decided not to detail the distinction between reflective and formative indicators. As is expected, and likely is appropriate given the perceptual nature of most of the research, the vast majority of measures involve reflective indicators. Fewer than five studies state some type of formative measure. However, one misnomer is that SEM is not appropriate for formative indicators. Formative indicator models present problems with statistical identification unless formulated in a manner as to avoid these problems. Thus, caution is advised to make sure overidentifying assumptions are in place. For a more comprehensive explanation, see MacKenzie, Podsakoff, and Podsakoff (Citation2011).