Student performance of natives and immigrants differed greatly in the Programme for International Student Assessment 2000 in Germany. This paper analyses the gap in test scores by estimating educational production functions, using an extension study with imputed data. The difference in test scores is assigned to various effects, using a Juhn-Murphy-Pierce decomposition. The analysis reveals that German students have a more favourable family background, particularly in the lower part of the test score distribution. The later enrolment of immigrant students and preferences of parents are more important than parents' education or the family setting for explaining the score gap. Differences in returns have no significant effect. (Contains 1 figure and 5 notes.)