This site uses cookies

Some of these cookies are essential, while others help us to improve your experience by providing insights into how the site is being used.

For more detailed information please check our Cookie notice

Necessary cookies

Necessary cookies enable core functionality. This website cannot function properly without these cookies.

Cookies that measure website use

If you provide permission, we will use Google Analytics to measure how you use the website so we can improve it based on our understanding of user needs. Google Analytics sets cookies that store anonymised information about how you got to the site, the pages you visit, how long you spend on each page and what you click on while you’re visiting the site.

How middle school policy affects primary school outcomes

Author: Kristian Koerselman
Institution: University of Helsinki
Type of case study: Research

About the research

The Tripartite System common in England and Wales between 1944 and the 1970s allocated students to different types of secondary schools based on their performance in the Eleven Plus examination, which was administered at the end of primary school. Such ability-based classification of students is called curriculum tracking. The Tripartite System was gradually replaced by the Comprehensive System in which all students attended the same type of middle school. With the demise of the Tripartite System, the Eleven Plus examination lost its importance.

Tripartite System students were under strong pressure to do well on the Eleven Plus examination, while similar students facing the Comprehensive System did not. We should therefore expect test scores of primary school students facing a Tripartite middle school to be higher. This project has looked for empirical evidence for such differences, i.e. whether age 11 test scores declined in England and Wales after the abolishment of the Tripartite System.

The abolishment of the Tripartite System indeed caused a marked decrease in age eleven test scores. This should however not be interpreted as an argument for reinstating the Tripartite System. It is unclear whether the test score difference reflects a difference in learning. Even if it does, it is an open question whether the effect is permanent. Rather, the point of this project is to illustrate that later age policies can affect earlier age outcomes through the impact of incentives.

The so-called incentive effects are important for two reasons. When looking for the determinants of school outcomes at a certain age, we should not only look among the policies at that age and before, but also at policies that happen after that age. For example, the OECD Programme for International Student Assessment (PISA), tests students at the end of middle school. The widespread attention given to country rankings of PISA test scores has prompted many countries to critically re-evaluate their middle school policies. Later age factors – such as the selectivity of high schools and universities as well as the state of the labour market – have been given much less consideration.

Second, incentive effects imply that earlier age exam scores cannot be blindly used as control variables when evaluating later age policies. For example, countries with comprehensive rather than tracked middle schools experience larger test score growth during middle school. It would therefore seem natural to conclude that comprehensive middle schools are doing a better job. In truth, however, countries with comprehensive middle schools start lagging behind countries with tracked middle schools during the last years of primary school, then catch up with the latter group during middle school. In this light, it’s more questionable whether comprehensive middle schools are doing a better job.

In general, we are used to thinking that causality does not run backward in time. Human beings are however forward looking, and can anticipate things that happen in their future. A policy that affects individuals tomorrow can change their behaviour today. This study is an example of why policy researchers and makers need to take this fact into account more consistently.


The analysis methods included regression of age 11 exam scores on middle school tracking status and of age 7 test scores as well as a wide range of other controls. Various robustness checks suggest that this is an adequate method for recovering the causal effect of middle school tracking policies for this particular dataset. A multilevel model was used, with tracking counted as a school level variable. Inflation of age 11 test scores takes place in order to take into account measurement error in this dataset. The resulting estimates should thus be interpreted as expressed in standard deviations of the latent trait.


Koerselman, K. (2012) ‘Incentives from curriculum tracking,’ Economics of Education Review, 32, pp. 140–150 doi: 10.1016/j.econedurev.2012.08.003

Koerselman, K. (2012) Incentives from curriculum tracking, Helsinki Center of Economic Research Discussion Paper No. 357, University of Helsinki. Retrieved 2 September 2013 from