The Impact of Childhood Environment on Personality and Mental Health in Adulthood
Personality traits predict key aspects of human wellbeing, including economic outcomes (e.g. Bowles et al. 2001; Heckman et al., 2006; Heineck and Anger, 2010), mental (e.g. Colvin and Block, 1995) and physical health (e.g. Friedman 1990; Denollet et al. 1996; Weston et al. 2015). But what determines personality in adulthood and how is it affected by childhood environments? This fundamental question has drawn interests from various disciplines. A substantial fraction of variability is due to environmental factors (e.g. Bouchard and Loehlin, 2001; Rushton et al., 2008), but it is unclear what these are.
Several strands of literature show the importance of childhood environments on various outcomes both during early years and in adulthood. In the following paragraphs we provide a short overview of their key findings as well as different strategies used to identify these effects, including lab experiments, randomized controlled trials, instrumental variables approach and panel data studies. We first focus on outcomes that have traditionally been studied by economists including educational attainment, earnings and physical health. We then move on to more recent work that indicate the impact of past environment on personality traits and mental health. Additional attention is given to preliminary studies that suggest an important role for the school and classroom environments, which are key variables for this proposed project.
The impact of childhood environment on education and labor market outcomes is well established. Credible large-scale studies include the effect of neighborhoods on earnings (Chetty and Hendren, 2018), socioeconomic status of peers on student achievement (Ewijk and Sleegers, 2010), childhood health and circumstances on adult health and labor market outcomes (Case et al., 2005). We think that personality could be an important mechanism behind these findings and use personality and mental health measures as key outcome variables.
Randomized controlled trials have also been used to evaluate the long-term impact of early childhood environment. Heckman and co-authors (2006, 2007) argue that early interventions can affect formation of non-cognitive skills in children. Preschool programs for disadvantaged children were shown to positively impact education attainment and earnings (Garces et al., 2002; Campbell et al., 2002). A comprehensive overview of the literature on skills formation is provided by Cunha et al. (2006).
Multiple longitudinal studies have focused on the impact of specific environmental factors on personality development. Work experience has been associated with personality development in adolescents (Roberts et al., 2003), while marriage or remarriage and death of a spouse has been linked to personality trait changes in adults (Mroczek and Spiro, 2003). Others looked at perceived support from parents and number of significant life events such as death of relatives (Branje et al., 2007; Kettlewell, 2019). These papers primarily focused on associations between personality changes and experiences, whereas we will exploit random variations in childhood experiences in the Dutch setting to identify their causal effects.
More recent work has revealed the importance of childhood economic conditions. Parental socioeconomic status has been shown to be correlated with personality (Ayoub et al., 2018). Adhvaryu et al. (2019) used variation in cocoa price to prove that income shocks during early childhood can lead to increased mental health issues in adults. We expand on this literature by considering additional environmental factors during childhood, most notably the school and classroom environment. The classroom setting can conceivably impact a child’s development as children spend a lot of time at school and it is also the setting where most of their social interactions happen.
Experimental evidence show that school environments can also affect the development of personality. Interventions in elementary school aimed at changing non-cognitive skills in children can affect behavior and test scores years after the intervention (Alan et al. 2019), indicating that what happens in school could play a major role in the development of personality. Targeted classroom programs were also successful in reducing behavioral problems in children (Barnett et al. 2008; Bierman et al. 2010). More recent work indicates that educational tracking affects non-cognitive skills with differential effects depending on age (Cotofan et al., 2019). Another strand of literature demonstrates the impact of classroom peers on educational outcomes (e.g. Sacerdote, 2001; Lavy and Schlosser, 2011; Booij et al., 2017) and behavior (Gaviria and Raphael, 2001).
Our study aims to link adult personality and mental health outcomes from Lifelines with registry data provided by CBS to analyze the impact of childhood environmental factors. While the papers cited above mainly focused on personality and behavior in school, we are interested in how these effects are carried into adulthood. It is clear that school environments affect longterm outcomes: class size has a significantly negative effect on wages decades after graduation (Fredriksson, Ockert and Oosterbeek, 2013). This study proposes to evaluate the impact of educational tracking, school starting age, school cohort gender composition and other childhood environmental factors for which quasi-random variation allows the identification of causal effects. The large sample size of Lifelines would allow us to study variations in these environmental factors to identify causal effects in non-experimental settings.
We also contribute to the literature by analyzing a wider variety of personality outcomes. The Big Five personality structure is by far the most popular personality structure. However, recent research has highlighted the importance of other personality dimensions on health and labor market outcomes. We will analyse the impact of environmental factors on mental health, including stress (Rosmalen et al., 2012) and Goal Adjustment Scale, a measure of suicidal thinking (O’Connor and Forgan, 2007).