空气净化器在学校的作用

Jacopo Bonan, Francesco Granella, Stefania Renna, Luis Sarmiento,and Massimo Tavoni The Effect of Air Purifiers in Schools∗ Jacopo Bonan1,2,3,*Francesco Granella1,2Stefania Renna1,2,5 Luis Sarmiento1,2,4Massimo Tavoni1,2,5 We randomize the installation of air purifiers across primary school classrooms toreduce children’s exposure to air pollution. The intervention reduces indoorP M2.5concentrations by 32% and decreases student absenteeism by 12.5%.Effects arelarger among students with higher pre-treatment absenteeism. The impact is greaterwhen outdoor air pollution is relatively low and diminishes as outdoor pollutionintensifies, consistent with non-linear marginal effects of air quality on health. Thetreatment students report fewer respiratory symptoms and exhibit greater awarenessof air quality. The cost per absence day avoided is approximatelyAC11, resulting ina conservative cost-benefit ratio of one-to-nine. JEL codes:C93, I21, Q53, Q51 Keywords:Indoor air quality, air purifiers, school absences, randomized controlled trial *Corresponding author. E-mail: jacopo.bonan@cmcc.it 1. Introduction Air pollution is a global health issue contributing to child mortality and morbidity worldwide(Institute for Health Metrics and Evaluation, 2019; Annesi-Maesano et al., 2021). High pollution-related morbidity increases school absences (Currie et al., 2009), negatively impacting learningoutcomes and educational activities (Gershenson et al., 2017; Aguilar-Gomez et al., 2022). Al-though various pollution-control measures have reduced average urban exposure, levels oftenexceed World Health Organization (WHO) guidelines (World Health Organization, 2021), dam-aging health and welfare and leading to significant economic losses (Shaddick et al., 2020). Sincetraditional air pollution control mechanisms, such as low-emission zones or industrial policies,can be costly and complex to implement, adaptive measures to reduce exposure are needed. This study evaluates the efficacy and cost-effectiveness of installing portable air purifiers inschool classrooms. Our main hypothesis is that air purifiers reduce absenteeism by improvingindoor air quality and thus promoting children’s health. Using a cluster randomized controlledtrial (RCT) across 5 primary schools in Milan, Italy—an area known for poor air quality (EEA,2022)—we randomly assigned 95 classes to either receive or not receive air purifiers.We alsoinstalled indoor air quality sensors in a subsample of classrooms to collect detailed data on airpollution and environmental conditions. The air purifiers reduce indoor air pollution by 32%.The relative efficacy of the purifiersshows no relation to outdoor air pollution levels and remains rather stable throughout the studyperiod.The treatment decreases school absences by approximately 12.5%, about 1.3 misseddays per year. The effect is more pronounced for students with higher baseline absenteeism. Bystudying dynamic treatment effects we found that that most of the reduction in absences occursin fall and spring, rather than in winter when average pollution levels are significantly higher.Supporting this observation, evidence shows that purifiers do not have statistically significanteffects on absences during periods of high outdoor air pollution. Specifically, the treatment effecton absences is no longer significant when the 10-day rolling average of ambientP M2.5exceeds10 micrograms per cubic meter (μg/m3), or when more than 2 of the past 10 days exceed theWHO threshold of 15μg/m3.These insights align with previous findings indicating that themarginal effect of air pollution on health is concave; that is, improving air quality when pollutionis relatively low has greater health effects than doing so when pollution is high (Berkouwer andDean, 2023; Miller et al., 2024; Weichenthal et al., 2022; Corrigan et al., 2018; Pope III et al.,2015; Arag´on et al., 2017). Using survey data, we find that the treatment students are less likely to report respiratorysymptoms over the past week than control students. This result suggests that reduced absenceslikely arise from improved health.We also observe significant treatment effects in students’perceptions of classroom air quality and preferences for urban policies related to air quality.However, we cannot rule out experimenter demand and priming effects. To examine the potentialconfounding role of behavioral changes in our results, we estimate the treatment effect on proxiesfor opening and closing classroom doors and windows (e.g., sudden shifts in temperature orcarbon dioxide).We find no evidence that the introduction of purifiers significantly altersventilation behavior or classroom occupancy, as indicated by the lack of differences in classroomCO2levels, temperature, and estimated ventilation episodes. Cost-effectiveness calculations indicate that installing air purifiers leads to a cost per avoidedabsence ofAC10.6 and a lower-bound benefit-cost ratio of 9.6. The analysis was pr