School closures significantly reduced arrests of black and latinx urban youth

Abstract

Background & methods: Youth of color are surveilled and arrested by police at higher rates than their White peers, contributing to racial inequities across the life course and in population health. Previous research points to schools as an increasingly relevant site for youth criminalization, but existing studies emphasize within-school mechanisms, with limited analysis of policing in surrounding school areas. To fill this gap, we study changes in police arrests of youth after initial COVID-19 school closures in 2020 across four US cities overall and in relation to public school locations. We analyze geocoded arrest records and use interrupted time series negative binomial regression models with city and month fixed effects to estimate change in weekly arrest rates among White, Black, and Latinx youth. We estimate arrest densities within school areas before and after school closures using spatial buffers of 300 feet.

Results: In the immediate weeks and months following COVID-19 pandemic school closures, youth arrest rates fell dramatically and with lasting impacts compared to other age groups. During the period of remote learning, weekly youth arrest rates declined by 54.0% compared to youth arrests rates in 2019, adjusting for city and seasonality (Incident Rate Ratio 0.46, 95% CI: 0.41, 0.52). We estimate Black youth weekly arrests fell from 43.6 to 16.8 per 100,000, vs. 4.6 to 2.2 per 100,000 among White youth. However, Black youth arrest rates during the remote learning period were still nearly 5 times that of White youth pre-pandemic. We also find that youth arrest rates declined during two school closure periods: at the start of the pandemic in 2020 and during Summer 2019. A spatial analysis shows Black and Latinx youth arrest densities in the surrounding 300 feet of K-12 schools were at least 15 and 8.5 times that of White youth, respectively, in both pre- and remote-learning periods.

Conclusions: Black and Latinx youth face a higher likelihood of being arrested near a school than do White youth and older age groups, and racial inequities in arrests remains after school closures. Our findings show school closures significantly reduced arrests of urban youth of color, and policies addressing youth criminalization and structural racism should consider the joint spatial context of schools and policing. Although school closures may have resulted in learning loss and harms to youth wellbeing, closures interrupted comparatively high levels of arrest for Black and Latinx youth.

Fig 1. Temporal trends in youth arrest rates in Boston, Charleston, New York City and Pittsburgh, 2019–2020.

Weekly averages smoothed using a loess function (span = 0.25).

Fig 2. Percentage change in rates of arrest by age group before vs. after school closures in Boston, Charleston, New York City, and Pittsburgh, 2019–2020.

Estimates from four age-specific negative binomial models predicting rates of arrests with a population offset and fixed effects for city and month. The referent time period for these models is 2019-March 2020. Only New York City and Charleston contribute data to the “Partially Returned” period because Boston and Pittsburgh remained remote after September 2020. Complete regression results reported in S1 Table.

Fig 3. Temporal trends in youth arrest rates by race in Boston, Charleston, New York City, and Pittsburgh, 2019–2020.

Weekly averages smoothed using a loess function (span = 0.5).

Fig 4. Weekly youth arrest density (arrests/km²) in school areas (300-foot buffer) and surrounding cities in pre- and post-remote learning periods, New York City, New York, 2019–2020.

Fig 5

Percent of Arrests in School Areas (A), Weekly Density of Youth and Young Adult Arrests in school areas per km² (B), and Percent (C) and Absolute (D) Change in Arrest Density in School Areas by Time Period, Age Group, and Race/Ethnicity.