Media Briefings

Pupils Taught In Smaller Classes Get Higher Exam Marks

  • Published Date: June 2010

Class size has a very significant effect on examination marks at the end of compulsory schooling, according to new research by Dr Eskil Heinesen. His analysis of data on the educational outcomes for pupils in Denmark shows that the smaller the class, the higher the marks pupils achieve at age 15 or 16.

But the study, which is published in the June 2010 issue of the Economic Journal, also finds that the effect of a reduction in class size is not the same for all pupils:

  • Boys benefit more from a reduction in class size than girls – the effect of reducing class size is about 50% larger for boys than for girls.
  • Academically weak pupils benefit more than academically strong pupils.
  • Pupils whose mothers did not go through further or higher education benefit more – again the effect is about 50% larger than for pupils whose mothers have attended institutions of further or higher education.

Class size is generally considered to be a central aspect of school quality. It is of great concern to parents and is directly affected by political choices. The extent to which class size matters for pupils’ educational outcomes is an issue of great importance since it is expensive to reduce class sizes.

But estimation of the causal effects of class size is difficult because pupils are not randomly distributed across small and large classes. For example, parents who feel strongly about their children’s education and well-being in school may move their children to other schools if they are not satisfied with school quality, perhaps because of poor teachers or bad relationships between pupils in the class. And school administrators may often place less able pupils in smaller classes.

For these reasons, smaller classes may often be small because they are of low quality or have many disadvantaged pupils. To overcome these problems, this analysis exploits an institutional feature of the Danish education system that provides random variation in the size of subject-specific classes – French and German classes.

About 22% of Danish pupils attend schools that offer a choice between German and French from seventh grade (age 13-14). The size of German and French classes at a given school in a given year depends on how many pupils choose one or the other. All subjects other than German and French are typically taught in basic classes. Because of data issues, the analysis focuses on class-size effects in French.

It is reasonable to suppose that pupils who choose French (for example) will not change school because of a large class size or low teacher or peer group quality in the French class specifically, since the quality of basic classes is far more important.

Also, class-size effects are estimated from variation over time within schools only, thereby taking account of time-constant unobserved differences between schools and catchment-area populations.

The analysis is based on Danish administrative data for three cohorts of pupils. These data contain individual examination marks at the end of ninth grade (that is, at the end of compulsory school in Denmark – age 15-16) and detailed information on pupils’ parental background.


Notes for editors: ‘Estimating Class-size Effects using Within-school Variation in Subject-specific Classes’ by Eskil Heinesen is published in the June 2010 issue of the Economic Journal.

Eskil Heinesen is at AKF, the Danish Institute of Governmental Research.

For further information: contact Eskil Heinesen on +45 4333 3426 (email:; or Romesh Vaitilingam on 07768 661095 (email: