Factor analysis in the development and refinement of clinical assessment instruments.

The goals of both exploratory and confirmatory factor analysis are described and procedural guidelines for each approach are summarized, emphasizing the use of factor analysis in developing and refining clinical measures. For exploratory factor analysis, a rationale is presented for selecting between principal components analysis and common factor analysis depending on whether the research goal involves either identification of latent constructs or data reduction. Confirmatory factor analysis using structural equation modeling is described for use in validating the dimensional structure of a measure. Additionally, the uses of confirmatory factor analysis for assessing the invariance of measures across samples and for evaluating multitrait-multimethod data are also briefly described. Suggestions are offered for handling common problems with item-level data, and examples illustrating potential difficulties with confirming dimensional structures from initial exploratory analyses are reviewed.