The article will analyse PISA - particularly the part dealing with science - as an example of a major comparative evaluation. PISA will first be described and then analysed on the basis of test theory, which will address some detailed technical aspects of the test as well as the broader issue of validation. The purpose of this is to illustrate how the technical aspects of evaluations are not neutral practices, but rather a part of the fundamental value system on which the evaluation is based. Some apparently objective choices must necessarily be made which have consequences for the theoretical basis of the evaluation, and the technique thereby becomes part of the fundamental value system. These considerations form the basis for an evaluation of PISA's predicative power in a national context - in this case, that of Denmark. On this basis, the analysis will focus on the relationship between PISA's fundamental assumptions and the national consequences of participation. Finally, I will conclude with some reflections on how PISA may be utilised and developed.