Empirical chapters

In this study, published in European Sociological Review, we (with Antonie Knigge, Ineke Maas, Eveline de Zeeuw, and Dorret Boomsma) used Dutch twin data to answer the question of whether classrooms equalize or amplify educational inequality. By comparing twins in the same vs. different classrooms, we disentangle four sources that contribute to differences in performance: genetic (A), shared environmental (C), non-shared environmental (E), and classroom differences (CL). Our results show that the classroom environment contributes to performance differences to a small extent and that its influence is dependent on family background. The influence of the classroom was larger when the level of parental SES was lower, indicating a compensatory effect.

• Read the whole article here (open access)

• See here for a blog post on the article by Jeroen Janssen (in Dutch)

• See here for an article in Didactief by Bea Ros (in Dutch)

Do schools increase or decrease environmental and genetic influences on educational performance? With Antonie Knigge and Ineke Maas, we investigated this in our study published in npj Science of Learning. Building on behavioral genetics literature on gene-environment interactions and sociological literature on the compensating and amplifying effects of schools on inequality, we investigate whether the role of genes and the shared environment is larger or smaller in higher-quality school environments. We apply twin models to Dutch administrative data on the educational performance of 18,384 same-sex and 11,050 opposite-sex twin pairs, enriched with data on the quality of primary schools. Our results show that school quality does not moderate genetic and shared-environmental influences on educational performance once the moderation by SES is considered. We find a gene-environment interplay for school SES: genetic variance decreases with increasing school SES. This school SES effect partly reflects parental SES influences. Yet, parental SES does not account for all the school SES moderation, suggesting that school-based processes play a role too.

• Read the whole article here (open access)

• See here for an interview about this article by the EUI

In this study, co-authored by Kristian Karlson and published in Social Science Research, we investigated the role of gender, family SES, school SES, and their intersection in educational achievement using a twin design. Drawing on theories of gene-environment interaction, we test whether high-SES environments compensate for genetic risks or enhance genetic potential, and if this is dependent on gender. Using data on 37,000 Danish twin and sibling pairs from population-wide administrative registers, we report three main findings. First, for family SES, but not for school SES, we find that the relative contribution of genes is slightly smaller in high-SES environments. Second, this relationship is moderated by child gender: in high-SES families, the genetic influence is considerably lower for boys than for girls. Third, the moderating effect of family SES for boys is almost entirely driven by children attending low-SES schools. Our findings thus point to significant heterogeneity in gene-environment interactions, highlighting the importance of considering the multiplicity of social contexts.

• Read the whole article here (open access)

In this final study of my dissertation, I investigate if differences in educational performance (i.e., dispersion) and the underlying genetic and environmental differences are reproduced, exacerbated, or compensated over the primary school career. I use longitudinal reading and mathematics test score data of ~5,500 same-sex and opposite-sex twin pairs, identified in the Netherlands Cohort Study on Education (Grades 1-5, age 6-11). Employing biometric latent growth models leads to three main conclusions. First, most dispersion is already present at the start of formal education and is accounted for by genetic variance. Second, initial performance differences at the start of formal education are compensated over time. The extent to which this is due to decreasing genetic and environmental differences varies by domain. For reading, genetic differences decrease whereas environmental differences are largely reproduced. For mathematics, the opposite is true: environmental differences decrease and genetic differences are reproduced. Third, new influences come into play over the primary school career, mostly new genetic influences. Combining the development of pre-existing sources of dispersion with these new influences results in an increase in the total dispersion in educational performance during primary education.

• Read the pre-print here (open access)