In the first paper, titled "The ADHD deficit in school performance across sex and parental education", I investigated the relationship between ADHD and school performance. I analysed register data on practically every Norwegian junior high school student born between 1997 and 2002 (N = 344,152) and found a large difference in school performance between children with and without ADHD. This difference was partly due to ADHD being more prevalent in groups that typically perform worse in school, such as boys and children of less educated parents. When I adjusted for potential confounders, including unobserved familial confounders by comparing siblings, the difference was reduced but remained relatively large. Performance differences related to ADHD was not stable across social background: The differences was larger among children with highly educated parents, despite ADHD being less common in these groups. Finally, I also found that ADHD was associated with large performance differences across all school subjects, indicating that ADHD is associated general factors shared across these subjects.
In the second paper, titled "Parental income and mental disorders from age 10 to 35", I studied the relationship between parental income rank and prevalence of mental disorders using register data on practically every Norwegian resident aged 10 to 35 between 2006 and 2018 (N = 2,112,355). I found large differences in the prevalence of mental disorders across parental income quartiles, with higher prevalences in the lower income quartiles throughout the entire age-range. Applying an extended children-of-twins model, I found evidence suggesting a small, direct effect of parental income among adolescents, but genetic similarity accounted for most of the correlation. In terms of variance explained, high heritability was evident for both income rank (30–40%)and mental disorders (47–70%), with minimal variance explained by shared environmental factors.
Finally, in the third paper, titled "Genetic similarity between relatives provides evidence on the presence and history of assortative mating", I used path analysis to investigate the expected impact of assortative mating (i.e., non-random matching of partners with similar traits) on genetic similarity within families. I found that assortative mating will have a greater impact on similarity between distant relatives compared to close relatives, indicating that genetic variants associated with assorted traits will concentrate in extended families, potentially increasing inequality. I then correlated polygenic scores between 47,135 partner pairs and 1,213,258 dyads of related individuals in the Norwegian Mother, Father, and Child Cohort Study (MoBa) cohort study, and found empirical evidence of assortative mating for several traits including educational attainment, height, intelligence, and body mass index. Surprisingly, I did not find genetic evidence of assortative mating for mental disorders, which could either suggest other causes of phenotypic partner similarity or that the polygenic scores are severely limited. For some of the traits that showed evidence of assortative mating, such as educational attainment, my analysis suggested that genetic variance and similarity in families are still increasing across generations, leading to increasing differences. However, most of the genetic consequences of assortment had already manifested.
In the thesis introduction, I discuss the nature of causality, differences between genetic and environmental influences, and the challenges posed by inferring causality from observational data. This exploration sets the groundwork for the analyses carried out in the research. In the concluding chapters, I further address methodological considerations and potential biases, and provide a general discussion on the potential implications of these findings.
Overall, the findings across these papers reiterates that there is a strong association between socioeconomic position and mental disorders. As to whether this association is causal, the answer depends on whether we are looking at this across generations or within individuals. Across generations, it appears that the association cannot entirely be blamed on confounding factors, but a large part of the association seems attributable to a genetic correlation. Within individuals, the answer remains uncertain, but when my findings are seen in the light of other studies, bidirectional influences between mental health and socioeconomic position seem plausible. Finally, assortative mating seems to increase and maintain the differences caused by intergenerational transmission of socioeconomic differences.