Can We Compare Attitudes Towards Crime Around the World? Assessing Measurement Invariance of the Morally Debatable Behavior Scale Across 44 Countries

Abstract

Abstract Objectives We aim to encourage scholars who conduct cross-national criminological studies to routinely assess measurement invariance (MI), that is, verify if multi-item instruments that capture latent constructs are conceptualized and understood similarily across different populations. To promote the adoption of MI tests, we present an analytical protocol, including an annotated R script and output file. We implement the protocol and, doing so, document the first test of configural, metric, and scalar invariance of the three-factor Morally Debatable Behavior Scale (MDBS). Methods We worked with data from wave seven of the World Values Survey (WVS). Applying multi-group confirmatory factor analyses, we, first, explored invariance of the MDBS in 44 countries (N = 59,482). Next, we conducted analyses separately for seven South-american, six South-east Asian, six East-asian, two North American and Australasian, and all four Anglophone countries. Results The MDBS displays an overall lack of invariance. However, we confirmed configural invariance of the MDBS for the South-east Asian sample, metric invariance in the sample of Anglophone countries, and scalar invariance for the Australasian and North American countries. Conclusions Wave seven of the WVS can be used for latent mean score comparisons of the MDBS between the Australasian and North American countries. Associative relationships can be compared in the larger Anglophone sub-sample. Taken together, MI must be tested, and cannot be assumed, even when analyzing data from countries for which previous research has established cultural similarities. Our protocol and practical recommendations guide researchers in this process.