Evaluating the Measurement Invariance of MMPI-2-RF Restructured Clinical Scale 4 (Antisocial Behavior) between American and Korean Clinical Samples: Exploring Cultural and Translation Issues Affecting Item Responding

Abstract

Establishing the cross-cultural measurement invariance of psychometric scales is considered an essential step before scale means are compared across cultures. Although the MMPI instruments have been extensively researched, few studies have examined the measurement equivalence of MMPI scales in cross-cultural research. This study examined the measurement invariance of MMPI-2-RF Restructured Clinical Scale 4 (RC4; Antisocial Behavior) using multi-group confirmatory factor analysis with American and Korean clinical samples by (a) comparing a rationally-derived four-factor model (School Problems, Substance Abuse, Family Problems, and Violation of Social Norms) with a one-factor model, and (b) examining the measurement invariance of the RC4 four-factor model. After adjusting for age and gender, partial scalar invariance was achieved, and six non-invariant items were identified, most of which centered around substance abuse. Results support the generalizability of the four factors across cultures; however, special attention is needed when using substance abuse items with Korean clinical populations. Plausible sources of item non-invariance were explored in the context of translation challenges and observed patterns of relationship with external measures.

Acknowledgments

The authors thank Amy S. Dykhouse, Allison L. Gilson, Andy Heft, Brian J. McCabe, and Joseph K. McLaughlan for their help in identifying the RC4 scale rational clusters.

Data availability

Data from the American sample is property of University of Minnesota Press and can be obtained by contacting University of Minnesota Press, Test Division. Data from the Korean sample is property of Maumsarang Inc. and can be obtained by contacting Maumsarang Inc.