Translating psychological tests and attitude scales from one language and culture for use in others has been common in the history of educational and psychological assessment. Despite the long history, methods for conducting test and attitude scale translations and establishing equivalence of the multiple versions of an instrument are not well-known or routinely used in practice. The primary purpose of this paper is to focus attention on three major sources of invalidity in test translation work and to suggest solutions whenever possible. The first is cultural differences associated with administrations, item formats, and speededness. The second is technical factors associated with the instrument itself, the selection and training of translators, the translation process, and the judgmental and empirical designs that can be used in translating instruments. The third source of invalidity is factors affecting the interpretations of achievement test results such as similarity of curricula, equivalence of motivational levels, and sociopolitical considerations. (Contains 21 references.) (Author)