Maximum Likelihood Estimation of Factor Structures of Anxiety Measures: A Multiple Group Comparison

Abstract

This paper examines the construct generality of self-report measures of anxiety across male and female samples, and illustrates the use of confirmatory maximum likelihood techniques for examining factorial invariance. Factor structures for each of five anxiety inventories (state and trait forms of Spielberger's State-Trait Anxiety Inventory; Taylor's Manifest Anxiety Scale; two forms of Endler, Hunt and Rosenstein's S-R Inventory of Anxiousness) are identified and used to develop explicit measurement models. The generality of these models is assessed by examining the extent to which the model obtained for each measure is factorially invariant across sex groups. Results indicate that in this sample of 256 female and 135 male college students all measures, except the manifest anxiety scale, obtained statistically invariant factor patterns across sex groups.