open menu

Selective schooling and long-term health

Research overview

Education has long been regarded as important for adult health and health inequalities. Given that policy influences the amount, quality, and distribution of education, it is important to study to what degree education impacts health.

One difficulty is that controlled experimental studies for long-term outcomes like adult health are impractical, and this means we rely on observational studies. However, because background socio-economic characteristics are key drivers of education, confounding is a major issue in observational studies that is difficult to account for and because those attending different types of school will differ it may be their background rather than schooling that affects adult health.

Natural experiments provide a way around this by controlling for confounding more through design rather than statistical control. Historically, state selective schools (grammar and senior secondary in England and Scotland, respectively) chose students based on scores from tests given at the end of primary schooling, around age 11. This opens up the possibility of a natural experiment, as those with similar test scores who fell one side or the other of the cut-off are likely to have a similar range of background characteristics, but attended different schools as if randomised.

Such a situation is ideal for a regression discontinuity study. Previous research has used this method with the Aberdeen Children of the 1950s cohort to assess selective schooling’s impact on economic outcomes. It showed that selective schooling increased the length of education, the likelihood of getting higher qualifications, and the likelihood of having a professional occupation. However, selective schooling only slightly increased income for women and did not increase income for men.

The Aberdeen Children of the 1950s cohort is a large population with extensive early-life records of schooling and socio-demographics, and follow-up later in life. This allows evaluation of the natural experiment in schooling assignment, and here we propose to analyse the effect of this assignment on long-term health.


Grammar schooling is a perennial topic of interest, and at the time of writing (Autumn 2016) it is back on the political agenda in England. This study is about the impact of increased length of education, with the selection process of grammar schooling in the 1960s providing a natural experiment and allowing us to get closer to a causal effect of increased schooling.

Recent reforms in England mean that pupils are expected to continue in education or training up to 18, an extension of the school leaving age by 2 years. This is coupled with the huge increase in recent years in higher education participation.

Knowing whether these increases in education may have benefits for health is an important question as they affect large numbers of people. Moreover, if education does have benefits for health, then investing in more education in those presently disengaging early may be one possible way of reducing health inequalities. Of course one study cannot answer this question definitively, but this study adds to the literature using natural experiments to assess education’s impact.

Date Approved

February 2017

Project team

Frank Popham (lead researcher)               University of Glasgow

Corri Black                                              University of Aberdeen

Michelle Hilton Boon                                 University of Glasgow

Jessica Butler                                          University of Aberdeen

Peter Craig                                              University of Glasgow

Chris Dibben                                           University of Edinburgh

Ruth Dundas                                           University of Glasgow

Marjorie Johnston                                    University of Aberdeen

Page last updated: 05/01/2018