Principal Investigator: Dr. Leandro Carvalho
Department: Center for Economic and Social Research
University of Southern California, Center for Economic and Social Research, 635 Downey Way, Los Angeles CA 90089, United StatesTags: 15666, Compulsory schooling, education, gene-environment, health
Lead Collaborators: 1) Mr Patrick Turley
Collaborating Institutions and Addresses: 1) Harvard University, Department of Economics, 1805 Cambridge Street, Cambridge MA 02138, United States
Funding body: National Institutes of Health
1a: We aim to understand how a range of health-related outcomes are influenced by education and by the interaction of education with genetic variants.
1b: Our research will contribute to a better understanding of the effects of a social policy — a change in the compulsory schooling law — on health-related outcomes and whether the effects of such policy may depend on genotypes. It will also provide a better understanding of gene and environment interactions.
1c: In 1972, education reforms in England increased the minimum age at which a student could drop out of school from 15 to 16. Students turning 15 the year before this policy change and student turning 15 the year after would be approximately identical except that some fraction of the latter group would be constrained to get more education. As such, differences in health outcomes between these two groups may be attributed to the additional year of education imposed. We will also consider how a person’s genetic makeup influences the relationship between this additional schooling, SES, and health.
1d: Full cohort
There is still no consensus on whether the strong association between education and health reflects causality from more education to better health. A recent and growing literature exploits changes in compulsory schooling laws to study the causal effects of education on health (e.g. Lleras-Muney 2005; Oreopoulous 2006; Albouy and Lequien 2009; Silles 2009; Powdthavee 2010; Kemptner et al. 2011; Clark and Royer 2013; Jurges et al. 2013). This literature has produced mixed evidence across countries, time, outcomes, and subpopulations. Reviewing this research, Grossman (2015) concludes that there is enough conflicting evidence (in these recent studies) to warrant more research on the question of whether more schooling does in fact cause better health outcomes. One possible explanation for this conflicting evidence is that education might have a heterogeneous effect on health, affecting individuals from different health, social, and genetic backgrounds differently. In this project we will study the heterogeneous effect of education on health and investigate whether heterogeneous effects can rationalize the conflicting evidence.
There is increasing recognition in the social and behavioral sciences that SES is not determined by nature or nurture but rather by the interplay of the two. However, researchers have struggled to produce evidence about the empirical importance of nature-nurture interactions, because the relationship between SES and environmental circumstances is typically confounded by third factors (e.g., individuals born to higher-SES parent may inherit wealth and family ?connections?). In principle one could overcome this problem by exploiting a natural experiment that generates exogenous variation in environmental circumstances, but it is difficult to come by data containing a natural experiment and genetic data for a large sample. We will combine a well-known natural experiment with the UK Biobank to study how nature and nurture interact to influence SES, using a polygenic predictor (or score) to consider the interaction of nature and nurture.
Project extension approved by UK Biobank 13/02/2017:
This project is investigating whether education has a causal impact on distributions of health outcomes. To estimate the causal impact of education, we take advantage of a school reform that in 1972 increased the minimum school-leaving age in England, Scotland, and Wales from 15 to 16 years. The reform affected only students born on or after September 1 1957, who had to stay in school until age 16; students born before that date could drop out at age 15. As a consequence, students born within days of each other?who would otherwise have had similar education?obtained different levels of schooling. The reform is estimated to have increased average education by 0.14 years of schooling. Using a regression discontinuity design, we will study the causal impact of education on health distributions by comparing the health distributions for those born right after September 1, 1957 to the distributions for those born right before this date. Point estimates from preliminary analysis using month and year of birth (data-fields 52 and 34) suggest that the effects of education on the distribution of BMI, for example, are concentrated in the upper tail of the distribution (reducing the likelihood of BMIs above 30).
The problem is that data-fields 52 and 34 are not sufficient to conduct hypothesis testing. With just data on month and year of birth, it is necessary to allow for correlation between individuals born in the same month (Card and Lee 2008), but unfortunately no method has been developed yet to correct for this correlation when testing differences between two distributions. This issue only arises because data on month and year of birth are not sufficiently detailed.
We would be able to conduct hypothesis testing if we had data on date of birth (datafield 33) because in this case no clustering is needed, allowing us to use existing methods (Shen and Zhang 2015).2 In other words, gaining access to data-field 33 would enable us to generate p-values for a test of the causal effects of education on health distributions, a crucial step for the successful realization of our project.
Project extension requested 26/02/2019:
In our original submission, we proposed exploiting the 1972 school-leaving age reform to study the causal effects of education on cognitive function. We are requesting access to the brain MRI imaging data to further study such effects.
Last updated Jun 10, 2019