Smaller classrooms lead to higher adult incomes

11 March 2014

A class size reduction from 25 to 20 students provides students on average with over three per cent higher wages when they are between 27 and 42 years old.

This comes from a large-scale study based on Swedish data from various sources and archives. The research also confirms earlier findings that students coming from smaller primary school classes score better on cognitive ability and personal skills in their teens. Smaller classes also benefit the level of completed education, according to the research by Hessel Oosterbeek, professor of education economics at UvA's Amsterdam School of Economics together with Swedish academics Peter Fredriksson and Björn Öckert. 

Oosterbeek: "There is a considerable amount of research on the effects of class size in the short term by economists as well as psychologists and educators. Most studies find a positive effect of class size on learning results. This was the first time we could show that the positive effects do not fade out and translate into economically meaningful improvements in labour market outcomes. The effects in the long term are almost as strong as the effects in the short- and medium-term." Smaller classrooms require extra teachers and classrooms but the wage effect is large enough to pass a cost-benefit test. A rough calculation estimates an internal rate of return of almost 20%. 

The study is based on Swedish data about class size in the last three years of primary school. Oosterbeek: "You cannot simply compare small and large classes in any given country. It would be likely that other influences would also play a role. Maybe schools choose to put difficult children in smaller classes. Also smaller classes could be over-represented in the better neighbourhoods or parents could have paid to let their children go to schools with smaller classes. In those cases you can’t draw any conclusions about the effect of smaller classes." In the United States, the state of Tennessee did a huge experiment with the STAR project, a study in which almost 12,000 students in 80 schools were divided randomly into classes of 15 or 22 students. However, the students in this experiment are still too young to show long-term effects. 

Sweden has a maximum class size rule of 30, which offered the scientists unique data. "A class of 31 students is usually split into two. This custom mimics an experiment because whether children are in a large or small class is determined as good as randomly. A school district with 31 students does not differ fundamentally from a district with 30 students," says Oosterbeek. In addition, Sweden keeps track of extensive population statistics, on income, also in combination with the parents’ income, and many students are extensively tested on various skills at ages 13, 16 and 18. 

The results of the study leave little space for opponents of smaller classes. Apparently the smaller Swedish classes on average had less experienced teachers. Oosterbeek: "This seems another result of class size reduction, when new teachers are required. But it emphasizes that our findings were not due to the quality of the teachers." Based on the research Oosterbeek cannot indicate whether a minimum desirable class size exists. 

"We had too few classes with less than 15 students to be able to draw any conclusions. We have focused on classes with 20 to 30 students and within this range it seems ‘the smaller, the better’-rule applies." The research already influenced Dutch education policy. Oosterbeek: "It was picked up in the House of Parliament by the political party D66 and some argumentation has contributed to adjustments in the recent budget agreement."

Amsterdam School of Economics