ABSTRACT
Confirmatory factor analysis is a special case of structural equation modeling. Although its conception has the same mathematical model as exploratory factor analysis, it differs considerably from exploratory factor analysis in how the model is used and the mathematics of its estimation. There is no searching of data for common factors via computing the eigenvalues and eigenvectors to get the estimates of factor loadings, factor correlations, and unique variances. Instead the researcher begins with a conception of a set of latent exogenous causal variables having specified effects on a set of endogenous manifest “indicator variables” and seeks to test his/her concept of the relation of the latent causes to the manifest indicators as a hypothesis. The hypothesis is formulated before seeing data on the manifest indicators. Usually the hypothesis is based in part on prior knowledge about variables such as the manifest indicators and their possible causes, but also possibly involves a new concept of how the causal variables combine together to be common causes of the indicators.
