The article reviews research and findings on problems and issues faced when translating international academic achievement tests. The purpose is to draw attention to the problems, to help to develop the procedures followed when translating the tests, and to provide suggestions for further research. The problems concentrate on the following: the unique and demanding purpose of the translation task, the partly contradictory task specifications and translation instructions, the indecision as to whether to produce one or two target versions, the indecision as to whether to use one or two source versions, inadequate revision and verification, deficient translator competences, and a lack of time. To solve the problems, the article suggests the following: ensuring that the translation guidelines provide a right, unequivocal, and balanced picture of the purpose of the translation task; ensuring the equivalence of the two source versions; putting more emphasis on revision, and ensuring that the verification is sufficiently thorough; using only qualified translators, providing them with training in test translation, and including also subject matter and testing specialists in the translation teams; and allotting sufficient time to the translation work. However, the main lesson from the review is that more research in the field is badly needed. (Contains 1 table and 2 notes.)