Determining Nonequivalent Measurement in Cross-Cultural Family Research

Abstract

English and Spanish questionnaires designed to measure family and social-psychological variables were randomly distributed to bilinguals to determine the proportions of items and scales which produced nonequivalent scores after using back translation procedures in the questionnaire construction. The nonequivalent rate for demographic items was below that expected by chance while the nonequivalent rate for family and social psychological items was greater than chance with approximately 25 percent of items and scales producing nonequivalent measures. The data point to the need for cross-cultural family researchers to pretest their instruments on a group of bilinguals and then discard those items and/or scales which produce nonequivalent measures before the research is carried out. This is seen as a necessary step in the research process in order to increase the probability of equivalent measurement across cultures.