Rank Order Data and the Choice of a Correlation Input to Confirmatory Factor Analysis

Abstract

The consequences of factoring alternative correlation matrices are investigated assuming ordinal scale data. LISREL maximum likelihood solution algorithm is used for estimation. A single-factor model with a priori assigned loading values is studied via a simulation design. The performances of four selected correlation procedures, product-moment, polychoric, Spearman’s rho, and Kendall’s tau-b are compared under various design conditions. Implications of the study results are discussed.

Publication Title

Developments in Marketing Science: Proceedings of the Academy of Marketing Science

Share

COinS