A consideration of issues related to the confirmatory factor analysis of the evaluation use and evaluation involvement scales

Abstract

Factor analysis as a methodological technique has been continually improved and updated for use with data with a variety of characteristics, though the default settings on most software packages assume the use of continuous, normally distributed data. Evaluators planning to use confirmatory factor analysis (CFA) with ordinal survey data measured with intensity response format items must be aware of the special characteristics of their data as well as the decisions these characteristics necessitate in the CFA process. While completing CFAs on data collected using the Evaluation Use and Evaluation Involvement Scales, data characteristics and analysis criteria specific to ordinal data were considered. Guidelines for evaluators interested in completing CFA of data with similar characteristics are presented.