Educational inequality, manifested in varying levels of academic performance, leads to disparities in later life outcomes such as income, occupation, and health. Understanding the sources of this inequality is crucial for developing effective interventions. While socioeconomic status (SES) is a well-established factor, genetic differences also contribute significantly to performance variations. This study explores the interaction between these genetic and environmental factors within the context of school quality. There are competing hypotheses regarding how high-quality schools affect this interplay. The "cumulative advantage" hypothesis suggests that high-quality schools may exacerbate inequalities, benefiting high-SES students more due to their better preparation and cultural alignment with the school environment. In contrast, the "compensation" hypothesis proposes that high-quality schools could mitigate inequalities by providing support and resources that compensate for disadvantaged home environments. Similarly, within behavioral genetics, there are contrasting views regarding how advantageous environments influence genetic expression, with some suggesting increased realization of genetic potential in enriched environments (bioecological model), while others posit compensation for genetic risks in supportive environments (diathesis-stress model). The study aims to clarify this debate by examining whether school quality moderates the influences of genes and family background on educational performance, thereby informing strategies for reducing educational inequality. Furthermore, the study considers the ethical implications of reducing genetic inequalities in educational performance, acknowledging differing perspectives on whether genetic differences are inherently problematic or represent justifiable variations in potential.

The existing literature on gene-environment interactions in education offers mixed results. Social science research suggests that high-quality schools might either amplify or compensate for the impact of family background. Some studies propose that high-SES children benefit disproportionately from high-quality schools due to their superior prior academic preparation, while others argue that high-quality schools offer compensatory advantages to low-SES students by overcoming unfavorable home environments. Similarly, behavioral genetics research presents competing perspectives. The bioecological model suggests stronger genetic influences in high-quality schools, as they facilitate the realization of genetic potential. Conversely, the diathesis-stress model proposes that supportive school environments in high-quality schools could mitigate the negative effects of genetic risks. Recent studies utilizing polygenic indices (PGIs) have also yielded mixed results, some supporting the compensation hypothesis, where individuals with lower PGIs benefit more from higher-achieving schools, while others show more ambiguous patterns. A critical limitation of previous studies is the insufficient consideration of socioeconomic selection into schools. High-SES children are more likely to attend high-quality schools, confounding school effects with family effects.

This study employs a twin design, leveraging the power of this method to disentangle genetic (A), shared environmental (C), and nonshared environmental (E) influences on educational performance. The researchers used administrative data from Statistics Netherlands (CBS) on 29,434 same-sex and opposite-sex twin pairs (birth cohorts 1994–2007), enriched with school quality indicators from the Dutch Inspectorate of Education and parental SES data. The dependent variable is the Cito test score, a high-stakes national standardized achievement test administered at the end of primary school. School quality is measured as a factor score derived from multiple indicators reflecting school resources and school culture. Parental SES is constructed from parental education and income data, while school SES is calculated as the average parental SES of all children in the school. The researchers utilized structural equation modeling (SEM) within the framework of an ACE model to analyze the data. Since the zygosity (MZ vs. DZ) of same-sex twins is unknown, they used different values for the estimated genetic relatedness among same-sex twins (rSSG = 0.70, 0.75, and 0.80) to assess the robustness of their findings. The analyses involved fitting a series of ACE models, first examining unmoderated genetic and environmental influences and then testing for moderation effects of school quality, school SES, and parental SES on the ACE components. The models accounted for the nested structure of the data (twins within schools) by adjusting standard errors for clustering at the school level. The researchers also performed robustness checks using different values of rSSG and alternative operationalizations of school quality, including separate analysis of school resources and school climate dimensions. Non-parametric analyses were conducted to examine potential non-linear effects.

Initial analyses revealed substantial genetic influence (61–91%) on educational performance, with varying contributions from shared environmental (0–15%) and nonshared environmental (9–24%) influences depending on the assumed genetic relatedness of same-sex twins. School quality was positively associated with performance, but this effect was partially explained by parental SES. The core findings centered on the moderation analyses. Initially, models suggested a significant negative moderation of genetic variance by school quality (i.e., decreasing genetic variance with increasing school quality). However, this effect diminished and became non-significant when school SES and parental SES were included as moderators in the model, indicating that the initial finding was largely confounded by SES. The most robust finding was a statistically significant negative moderation of genetic variance by school SES. This effect persisted even after controlling for parental SES, suggesting that school-based processes beyond the simple reflection of parental SES also contribute to the reduction of genetic variance in higher-SES schools. This moderation effect represents a substantial reduction in genetic influence (40%). The model also revealed a weak moderation of shared environmental variance by school SES after controlling for parental SES, and a decrease in non-shared environmental variance with increasing school SES that was fully explained by parental SES. Robustness checks using different rSSG values and alternative school quality measures generally supported these key findings, though the precision of estimates varied across the different values of rSSG.

The findings challenge the notion that school quality directly reduces educational inequality. Although an initial indication suggested that genetic variance decreases in high-quality schools, this effect proved to be primarily driven by school and parental SES, not inherent school quality characteristics. The negative gene-SES interaction (i.e., decreasing genetic variance with increasing SES) supports the diathesis-stress model, indicating that more favorable environments (higher-SES families and schools) compensate for genetic risks leading to lower performance. This suggests that low-SES children face a double disadvantage, lacking compensatory mechanisms both at home and in their school environments. The study highlights the intertwined roles of family and school contexts in shaping educational outcomes. Future research should explore the interactions between these two levels to better understand their combined effects. The results also suggest that focusing solely on improving school quality may be insufficient to address educational inequality. Reducing school segregation and reforming school funding mechanisms to account for both socioeconomic and genetic factors might be more effective strategies.

This study demonstrates that school quality does not directly reduce educational inequality, either in terms of family background or genetic influences. The observed gene-environment interplay is primarily linked to SES composition effects rather than school quality differences. To effectively address educational inequality, policies should focus on reducing school segregation and reforming school funding mechanisms to compensate for both socioeconomic and genetic factors. Future research should investigate the underlying mechanisms of the gene-SES interaction, potentially focusing on specific learning difficulties rather than general academic performance.

The study's reliance on administrative data without zygosity information for same-sex twins is a limitation, as this necessitates the use of estimated genetic relatedness, which might introduce some uncertainty into the estimates. The use of an overall school quality measure might mask the effects of specific school characteristics. The sample selection, focusing on twins who attended the same primary school and had available Cito test data, could also limit the generalizability of the findings. The possibility of a ceiling effect due to censoring at the upper end of the Cito score scale should also be acknowledged, although the researchers argue that this is unlikely to significantly affect the results.