Abstract
BACKGROUND: In an effort to reduce viral transmission, many schools reduced class sizes during the recent pandemic. Yet the effect of class size on transmission is unknown. METHODS: We used data from Project STAR, a randomized controlled trial in which 10,816 Tennessee elementary students were assigned at random to smaller classes (13 to 17 students) or larger classes (22 to 26 students) in 1985-89. We merged Project STAR schools with data on local deaths from pneumonia and influenza in the 122 Cities Mortality Report System. Using mixed effects linear, Poisson, and negative binomial regression, we estimated the main effect of smaller classes on absence. We used an interaction to test whether the effect of small classes on absence was larger when and where community pneumonia and influenza prevalence was high. RESULTS: Small classes reduced absence by 0.43 days/year (95% CI -0.06 to -0.80, p < 0.05), but small classes had no significant interaction with community pneumonia and influenza mortality (95% CI -0.27 to + 0.30, p > 0.90), indicating that the reduction in absence due to small classes was not larger when community disease prevalence was high. CONCLUSION: Small classes reduced absence, but the reduction was not larger when disease prevalence was high, so the reduction in absence was not necessarily achieved by reducing infection. Small classes, by themselves, may not suffice to reduce the spread of respiratory viruses.