Modeling Incomplete Data in Exercise Behavior Research Using Structural Equation Methodology

Abstract

Exercise behavior research typically suffers from attrition and other forms of missing data. In studies that suffer from this common malady, several researchers have demonstrated that correct maximum likelihood estimation with missing data can be obtained under mild assumptions concerning the missing data mechanism. Model estimation with distinct missing data patterns can, in many cases, be carried out utilizing existing structural equation modeling software that allow for the simultaneous analysis of mean and covariance structures for multiple groups. Findings are discussed in relation to the utility of latent variable structural equation modeling techniques for analysis with incomplete data in the study of social-psychological determinants of exercise behavior.