Does education protect against depression? Evidence from the Young Finns Study using Mendelian randomization

Prev Med. 2018 Oct;115:134-139. doi: 10.1016/j.ypmed.2018.08.026. Epub 2018 Aug 23.


Using participants (N = 1733) drawn from the nationally representative longitudinal Young Finns Study (YFS) we estimate the effect of education on depressive symptoms. In 2007, when the participants were between 30 and 45 years old, they reported their depressive symptoms using a revised version of Beck's Depression Inventory. Education was measured using register information on the highest completed level of education in 2007, which was converted to years of education. To identify a causal relationship between education and depressive symptoms we use an instrumental variables approach (Mendelian randomization, MR) with a genetic risk score as an instrument for years of education. The genetic risk score was based on 74 genetic variants, which were associated with years of education in a genome-wide association study (GWAS). Because the genetic variants are randomly assigned at conception, they induce exogenous variation in years of education and thus identify a causal effect if the assumptions of the MR approach are met. In Ordinary Least Squares (OLS) estimation years of education in 2007 were negatively associated with depressive symptoms in 2007 (b = -0.027, 95% Confidence Interval (CI) = -0.040, -0.015). However, the results based on Mendelian randomization suggested that the effect is not causal (b = 0.017; 95% CI = -0.144, 0.178). This indicates that omitted variables correlated with education and depression may bias the linear regression coefficients and exogenous variation in education caused by differences in genetic make-up does not seem to protect against depressive symptoms.

Keywords: Depression; Education; Instrumental variables; Mendelian randomization.

Publication types

  • Research Support, Non-U.S. Gov't

MeSH terms

  • Adult
  • Depression / genetics*
  • Educational Status*
  • Female
  • Finland
  • Genome-Wide Association Study / methods
  • Humans
  • Male
  • Mendelian Randomization Analysis / methods*
  • Middle Aged
  • Risk Factors