Translating Achievement Tests for Use in Cross-National Studies.

The specific purposes of this paper are to address four important problems that arise when translating tests and how these problems might be resolved, to identify and review judgmental and statistical methods for establishing the equivalence of scores from the same test presented in different languages, and to provide some preliminary guidelines for persons doing test translations and equivalence studies.

Abstract

Translating achievement tests and questionnaires prepared in one language and culture for use in other languages and cultures has been a longstanding practice. Unfortunately, there ia considerable technical evidence which suggests that the quality of test translations vary considerably and too often the translations are not very good, thus reducing the validity of any results produced with the translated tests and questionnaires. The specific purposes of this paper are (1) to address four important problems that arise when translating tests and how these problems might be resolved, (2) to identify and review judgmental and statistical methods for establishing the equivalence of scores from the same test presented in different languages, and (3) to provide some preliminary guidelines for persons doing test translations and equivalence studies. U.S. DEPARTMENT OF EDUCATION Office 01 Educattonal Research and Improvement EOUCI.TIONAL RESOURCES INFORMATION CENTER (ERIC) Thts doct:ment has been reproduced as received from the person or organIzatton originating Msnor changes have been made to improve reproductrOn quality Points of view or opnionS stated ,n this docu. mast do not necessarily represent official OE RI position or poky 2 PERMISSION TO REPRODUCE THIS MATERIAL HAS BEEN GRANTED BY Ravt94.2, ghltiziLEToAi TO THE EDUCATIONAL RESOURCES INFORMATION CENTER (ERIC). Translating Achievement Tests for Use in Cross-National Studies3 Ronald K. Hambleton University of Massachusetts at Amherst, USA Translating achievement tests and questionnaires prepared in one language and culture for use in other languages and cultures has had a long history in educational and psychological testing. One of the earliest examples was the translation of the Binet-Simon Intelligence Scale for Children from the French language into the English language in 1911. By 1916 the Binet-Simon intelligence test had been translated into seven languages. Today, the practice of translating tests and questionnaires from one language and culture to others is widespread (Oakland & Hu, 1992). Tests of intellectual ability such as the Wechsler Intelli ence Scale for Children and the Stanford-Binet Intelligence Scale and measures of personality such as the Thematic Apperception Test and the Minnesota Multiphasic Personality Inventory are among the instruments that have been most frequently translated (Oakland & Hu, 1992). There appear to be at least two reasons for translating tests. First, the practice is economical: For those who want to assess a particular construct, it may be far less expensive and considerably faster to translate a test or questionnaire than to construct a new instrument to measure the construct of interest in a second language. Sometimes, too, the technical 'Support for the preparation of this paper was provided by the Internatitinal Association for the Evaluation of Educational Achievement (IEA) and the National Center for Educational Statistics (NCES) of the United States Department of Education. The opinions here, however, are those of the author and not necessarily these of IEA or NCES. 2Laboratory of Psychometric and Evaluative Research Report No. 241. Amherst, MA: University of Massachusetts, School of Education. 3To appear in the European Journal of Psychological Assessment.