Background: In fall 2020, all public K-12 schools reopened in broadly 3 learning models. The hybrid model was considered a mid-risk option compared with remote and in-person learning models. The current study assesses school-based coronavirus disease 2019 (COVID-19) spread in the early fall using a national data set. Methods: We assess COVID-19 case growth rates from August 10 to October 14, 2020 based on a crowdsourcing data set from the National Education Association. The study follows a retrospective cohort design with the baseline exposures being 3 teaching models: remote learning only, hybrid, and in-person learning. To assess the consistency of our findings, we estimated the overall, as well as region-specific (Northeast, Midwest, South, and West) and poverty-specific (low, mid, and high) COVID-19 case-growth rates. In addition, we validated our study sample using another national sample survey data. Results: The baseline was from 617 school districts in 48 states, where 47% of school districts were in hybrid, 13% were in remote, and 40% were in-person. Controlling for state-level risk and rural-urban difference, the case growth rates for remote and in-person were lower than the hybrid (odds ratio [OR]: 0.963, 95% confidence interval [CI]: 0.960-0.965 and OR: 0.986, 95% CI: 0.984-0.988, respectively). A consistent result was found among school districts in all 4 regions and each poverty level. Conclusions: Hybrid may not necessarily be the next logical option when transitioning from the remote to in-person learning models due to its consistent higher case growth rates than the other 2 learning models.