Class Size

In two studies, we use international student achievement tests to estimate the effect of smaller classes on student achievement in many countries. In nearly all countries, we can reject that class sizes have a statistically and economically significant effect on student achievement. Noteworthy class-size effects are observed only in countries with relatively low teacher salaries. The big challenge in estimating class-size effects is to get from correlation to causation, as students are not randomly allocated to schools and classes. In a first identification strategy, we thus use variation in class sizes between consecutive grades within the same schools that emerge from natural cohort fluctuations. A second identification strategy uses discontinuities that emerge from maximum class-size rules, with very similar results.

Here you can find a short non-technical overview on this topic.

Research papers:

Class-Size Effects in School Systems around the World: Evidence from Between-Grade Variation in TIMSS (with M.R. West). European Economic Review 50 (3): 695-736, 2006

Educational Production in Europe. Economic Policy 20 (43): 445-504, 2005

Which School Systems Sort Weaker Students into Smaller Classes? International Evidence (with M.R. West). European Journal of Political Economy 22 (4): 944-968, 2006

Two-pager:

Class Size: Does It Matter? (with E.A. Hanushek). EENEE Policy Brief 2/2011

Non-technical contributions:

International Evidence on Expenditures and Class Size: A Review. Brookings Papers on Education Policy 2006/2007: 245-272, 2007

Crowd Control: An International Look at the Relationship between Class Size and Student Achievement (with M.R. West). Education Next 3 (3): 56-62, 2003

Material available only in German

A contribution about our research results:

Kleine Klassen machen auch nicht schlauer. Sonntagsökonom in der Frankfurter Allgemeinen Sonntagszeitung vom 18.9.2005 über unsere Forschung zum Einfluss von Klassengrößen auf Schülerleistungen

Non-technical contribution:

Kleinere Klassen = bessere Leistungen? ifo Schnelldienst 58 (17): 6-15, 2005