Association between COVID infection at the time of delivery and preterm birth among persons with a singleton birth in California from June 22, 2020, to November 28, 2021. Predicted probabilities of preterm birth were obtained from linear probability models with fixed effects for hospital/place of birth and week of birth. Solid symbols indicate parameter estimates, and vertical bars depict 95% confidence intervals. Results are from a linear probability model with fixed effects for birth facility and week of birth. Covariates include the mother's age, age squared, birth parity, education, race/ethnicity, foreign-born status, and ZIP code of residence socioeconomic quartile. In panel b, the model specification includes a cross-product term between the mother's race/ethnicity indicators and an indicator for COVID infection during pregnancy. In panel c, the model specification includes a cross-product term between the mother's educational attainment and an indicator for COVID infection at the time of delivery. SARS-CoV-2 infection was measured at the time of delivery hospitalization. HS = high school. BA = bachelor's degree. Source: California vital statistics.