Examines the methodological problems that arise when using psychological instruments containing figurative or metaphorical language that may have different meanings for researcher and subject. While researchers try to be literal when formulating test items in more than one language, they do not always eliminate figurative expressions that are used to describe everyday experiences but which also can encode contrasting, or nonequivalent, cultural constructs. For the study, more than 1,000 ninth grade students attending public schools in St. Paul, Minnesota, including about 100 Laotian Hmong adolescents, responded to a questionnaire designed to measure well being and depression, in English or in translation. The development and administration of the translated survey instrument revealed significant differences between the way English speakers and the Hmong categorized psychological attitudes and affective states. Thus, a definitive determination of the mental health status of the Hmong subjects was not possible. Moreover, the study demonstrated that direct translation from one language to another will not always produce an interview instrument that is semantically matched item for item. Researchers may need to develop a separate group of culturally appropriate questions for each subject group, ensuring that nonequivalent forms of metaphors are removed.
