The impact of poorly translated items on measurement invariance
MetadataShow full item record
The effect of improperly translated instruments on measurement invariance has been a key concern of cross-cultural researchers. However, no research to date has investigated how poorly translated items affect statistical techniques for identifying measurement invariance across different cultural groups. The purpose of this study is to examine this issue by including both poorly and correctly translated items on questionnaires administered to American and Taiwanese students. Mean and covariance structure (MACS) analyses and likelihood ratio (LR) tests are used to determine which parameter(s) are most affected by the poorly translated items. Results show the test statistics of the hypotheses about measurement invariance or DIF were not particularly affected by the mistranslated items across both approaches. Furthermore, different types of translation errors did affect the invariance of factor loadings, and intercepts in MACS invariance analyses, and differential item functioning (DIF) detection of item discrimination, and item difficulty in LR tests, especially the not faithful poorly translated items.