This paper uses data from the Norwegian elementary school to test whether students from disadvantaged backgrounds benefit from smaller classes. The data cover one cohort of fourth graders who have been treated in small versus large classes for a period of three years. The Norwegian class size rule of maximum 28 students is used to generate credible exogenous class size variation. We find significant class size effects for the subgroup of students with parents who are educated at or below the upper secondary school level, and for the subgroup of students from dissolved families. The estimated effects for the former subgroup are smaller than those reported from STAR, while the effects for the latter subgroup are within the range reported from the STAR-studies.