Methodological Foundations

Abstract

In order to test a model with hypothesized relationships between variables, structural equation modeling (SEM) is used [see SPSS [2006]]. Originally described as “a class of methodologies that seeks to represent hypotheses about means, variances, and covariances of observed data in terms of a smaller number of ’structural’ parameters defined by a hypothesized underlying model” ([Kaplan, 2000, p. 1]), this definition has been recently revised as to “a class of methodologies that seeks to represent hypotheses about summary statistics derived from empirical measurements in terms of smaller numbers of “structural” parameters defined by a hypothesized underlying model” ([Kaplan, 2008, p. 1]). Hoyle [1995] presents the alternative characterization, that SEM is “a comprehensive statistical approach to testing hypotheses about relations among observed and latent variables”.