Measurement invariance across face-to-face and telephone modes: The case of minority-status collectivistic-oriented groups

Abstract

This article tests the implications of mode selection on measurement equivalence in interview-administered surveys among minorities using a Multi-Group Multiple Indicators-Multiple Causes Model. A measurement invariance test procedure comparing face-to-face and telephone modes was performed on an attitude scale concerning social integration between Arabs and Jews. The scale was measured in two surveys conducted simultaneously among the Arab minority in Israel. Hypotheses regarding mode effect on the measurement model parameters (mean, intercept, loading, and error variance) were formulated, with reference to three mediators: Cognitive burden, motivation, and threat of disclosure. The results partially supported the hypotheses: Face-to-face mode was fully superior with regard to item random error (reliability), was partially superior with regard to item systematic error, but was not superior with regard to item validity. The configural, metric, partial scalar, and latent factor mean invariance found in this study are an indication that such data sets may be combined cautiously. Implications are discussed for researchers using telephone or mixed modes for data collection in surveys dealing with sensitive topics among collectivistic-oriented minority-status groups. © 2011 The Author. Published by Oxford University Press on behalf of The World Association for Public Opinion Research. All rights reserved.