Cross-Language and Cross-Cultural Comparisons in Scale Translations

Abstract

High fidelity translations of measurement instruments from one language to another are required for meaningful comparisons between and within cultures and language groups. Methods from Item Response Theory (IRT) were used to assess the degree to which a French Canadian translation of the Job Descriptive Index (JDI) (Smith, Kendall, & Hulin, 1969), a job satisfaction inventory, provided measurement equivalence across source and target languages and cultures. Data were collected using the translated instrument administered to Francophone Canadians. Responses to the English version of the JDI were also obtained from (1) Francophone Canadians, (2) Anglophone Canadians, and (3) English-speaking Americans. These data allow for multiple comparisons within each item and test different hypotheses about sources of nonequivalence in the translation; the roles of language and culture in item bias are examined independently. Interpretations for the patterns of bias displayed by some items are presented.