Whether increasing resource use in schools has a positive effect on pupil performance has occupied governments, parents and researchers for decades. A main challenge when trying to answer this question is to separate the effects of school resources from the effects of pupils’ family background, since resources may be allocated in a compensatory manner, and pupils may sort into schools. We address these issues using a comprehensive dataset for two cohorts of pupils graduating from lower secondary school in Norway. The dataset is rich in performance measures, resource use variables and family background variables. As performance measures we use results at age 16 across 11 subjects, and we exploit the fact that we have both information from results from national exams and from continuous assessment in class. Controlling for family background, we find a positive but modest effect of resource quantity such as teacher hours per pupil, on pupil achievement. Observable teacher qualifications, within the variation present in lower secondary school in Norway, do not appear to have significant effects on school results. Resource quality as measured by teacher characteristics does not appear to have a significant impact on pupils’ marks. We find clear evidence of compensating resource allocation and teacher sorting as well as relative setting of marks.