Working Paper

During the COVID pandemic, schools were widely believed to be a venue that facilitated disease transmission. While some worried Black and Hispanic students would disproportionately face greater risk of COVID exposure at school (school COVID risk) – given well-documented racially disparate school environments and resources – little empirical evidence exists documenting or refuting potential disparities in school COVID risk. This paper begins to fill this gap by examining racial disparities in the school environments that contribute to COVID risk at school—specifically, vaccination rates and classroom space per pupil (indoor density) - in New York City elementary schools. We then document the relationship between school COVID risk and absenteeism, to shed light on potential academic consequences of this risk, and explore the extent to which this risk mediates racial disparities in absenteeism. Finally, repeat the analyses for middle and high school students to assess consistency of the findings in other school levels. While Asian and White students, on average, attend schools with higher vaccination rates, they also face more crowded classrooms than Black or Hispanic students. Perhaps surprisingly, White students are more likely than Black students to attend the riskiest schools – schools with both indoor density above and vaccination rates below expert recommended critical thresholds. Importantly, we find that much of the racial disparity in school COVID risk – roughly half – is attributed to differences across neighborhoods rather than within them. Controlling for school COVID risk narrows racial gaps in attendance and chronic absenteeism – suggesting variation in risk partially explains racial gaps in attendance rates and chronic absenteeism. We find similar results for the other school levels, although magnitudes vary. Our findings highlight the need to consider multiple dimensions of school health risk and distinguish between risks rooted in physical environments and those driven by behavioral responses.