Since the Covid-19 pandemic in 2020, indoor ventilation rates in classrooms became a major topic. However, the research on the effect of the indoor environment on occupants’ health and cognition is much older. Studies have shown that high concentrations of carbon dioxide (CO2) are negatively associated with the cognition performance of occupants (Du et al., 2020). Air pollution can also cause adverse health reactions (Fuentes-Leonarte et al., 2009). Elementary school children are a particularly vulnerable group because they are still in their growth phase and their school performance early on determines their success in higher education and the labour market later on. Past research found a negative relationship between classroom air quality and school performance (Wargocki et al., 2020). However, there are individual differences in the response to poor air quality which are still unknown. Overall health could be an important factor moderating individual resilience.

Therefore, this study examines if unhealthier school children in terms of overweight and higher sick leave days than their peers are less resilient towards the detrimental effects of indoor CO2 on test scores. Data from two field studies have been merged. The first study found a negative effect of indoor CO2 on test scores for pupils in schools in the south of The Netherlands (Palacios et al., 2022). From this dataset, 5 schools took part in an intervention study aiming to improve children’s dietary behaviour and physical activity levels, which successfully reduced BMI levels (Bartelink et al., 2019). Thus, the health data of 1,149 children could be connected to their test scores and the indoor CO2 concentration of 58 classrooms. The scores of a nationally standardized test have been collected in January 2019 and 2020. 

Linear regression models with various fixed effect specifications on the testing period, test domain and classroom were conducted. Standard errors were clustered on classroom and period levels to account for observation dependency within a classroom. Preliminary results show that higher indoor CO2 concentrations and being more often on sick leave are negatively associated with lower test scores (p < .05). Additionally, being overweight is negatively associated with test scores (p < .05), however, it partially mitigates the negative effect of CO2 (p < .05), contradicting our hypothesis. Being part of an intervention school offsets the negative effects of CO2 on test scores (p < .001). This countereffect is significant for the test domains of math and reading (p < .01) compared to spelling tasks. The regression models corrected for indoor temperature and relative humidity levels, indoor fine particles concentration, children characteristics (age, sex, absence other than due to sickness, parental socioeconomic status) and learning environment characteristics (class size, noise levels). In conclusion, the individual health of a child determined by BMI and sick leave days does not reduce the negative effect of CO2 on academic performance. However, an intervention aiming at the physical activity level and dietary behaviour can partially offset the negative impact of poor air quality on school performance.