This paper uses data from the National Pupil Database to investigate the differences in "performance" across the range of science courses available following the 2006 Key Stage 4 (KS4) science reforms in England. This is a value-added exploration (from Key Stage 3 [KS3] to KS4) aimed not at the student or the school level, but rather at that of the course. Different methodological approaches to carrying out such an analysis, ranging from simple non-contextualized techniques, to more complex fully contextualized multilevel models, are investigated and their limitations and benefits are evaluated. Important differences between courses are found in terms of the typical "value" they add to the students studying them with particular applied science courses producing higher mean KS4 outcomes for the same KS3 level compared with other courses. The implications of the emergence of such differences, in a context where schools are judged to a great extent on their value-added performance, are discussed. The relative importance of a variety of student characteristics in determining KS4 outcomes are also investigated. Substantive findings are that across all types of course, science prior attainment at KS3, rather than that of mathematics or English, is the most important predictor of KS4 performance in science, and that students of lower socio-economic status consistently make less progress over KS4 than might be expected, despite prior attainment being accounted for in the modelling. (Contains 4 tables, 2 figures, and 11 notes.)