Detection of Differential Item Functioning on the Kirton Adaption-Innovation Inventory Using Multiple-Group Mean and Covariance Structure Analyses.

This article demonstrates how Sörbom's (1974) mean and covariance structure analysis (MACS) model can be used to detect uniform and non-uniform differential item functioning (DIF) on polytomous ordered response items that are assumed to approximate a continuous scale. Using item responses from 773 civil service employees, the MACS model was applied to three cognitive style scales on the widely-used Kirton Adaption-Innovation Inventory (KAI, Kirton, 1976; 1977). As predicted, DIF across managerial and staff support occupational groups were detected on each of the scales. As expected, there was no evidence of DIF across gender groups and across two groups formed through a random split on any of the scales. Substantive implications for the use of the KAI as well as methodological implications and future research direction for detection of DIF are discussed.