A growing body of research and popular reporting shows racial differences in school modality choices during the COVID-19 crisis, with white students more likely to attend school in person. This in-person learning gap raises serious equity concerns. In our paper, we use unique panel survey data from the Understanding Coronavirus in America Tracking Survey (UCA) to document racial differences in how students from different racial backgrounds engaged in school in the 2020-2021 academic year and to study potential factors associated with the observed racial differences. Using data collected between September 30 and October 27 2020, we corroborate significant racial gaps in mode of schooling. Nationally, a majority (68%) of American parents used some sort of remote learning (fully remote or hybrid) but there exist significant differences by race. 77% of Black and Hispanic parents report using fully remote or hybrid learning options for their children, as compared with only 61% of white parents. Despite school districts' recent re-opening efforts, these observed racial differences have continued into the spring semester. Data from the first federal survey of districts' offerings show that by February 79% of all students have been offered the options to attend school either in-person or hybrid modes by their schools. However, more than 50% of non-white respondents were still learning remotely. Recent UCA survey results corroborate this pattern and further indicate that some portion of this in-person attendance gap is explained by parental preference. In early April, 60% of white parents expressed a preference for in-person learning as compared to approximately 25% of Black and Hispanic parents. We then use this survey responses and multinomial choice models to study the potential role of a myriad of important factors including perceived school quality, trust of media and health messaging, and type of school (public, private, or charter). Additionally, we match the UCA data to respondent's 2020 election polling responses to study the role of individual political leanings. Using a database of over 12,000 district reopening plans along with innovative cell-phone location data, we are also able to estimate school districts' learning option availability for respondents of the UCA at their census tract or county of residence. We also study the role of local COVID-19 infections and deaths using county-level data from Johns Hopkins' Center for Systems Science and Engineering. Using all this information, we are able to study the potential role that these explanatory factors might have had on the observed racial differences in the use of in-person, hybrid, or remote learning school options. Our results so far show that a combination of factors may explain observed racial differences. We do find that for initial decisions in the fall of 2020, political partisanship plays a relatively large, mostly asymmetric role. Respondents indicating they intended to vote for candidate Donald Trump in the 2020 Presidential Election were 11-percentage points more likely to select in-person learning and 14percentage points less likely to select remote learning as compare to undecided voters. Interestingly, we do not estimate any statistically significant relationship between support for Joe Biden and choice of learning modality. We also find that type of school attended had a significant role with those attending a private school having a 41-percentage point higher probability of attending school in-person and a 26-percentage point lower probability of attending fully remotely, as compared to enrollment in a public school. Additionally, parents appear to be responsive to local COVID-19 outbreaks. A one-percentage point increase in local incidence-rate per 100,000 is associated with an 8-percentage point increased probability of remote learning. Finally, we find that the decisions that local public schools make are also very important. It appears that when given the option parents choose in-person learning a higher probability of about 15-percentage points. Living in an area where all public school districts prefer in-person learning is associated with a 19-percentage point increase in the probability of families choosing in-person learning for their students. The racial differences in school modality choices that we observe along with concerns on the quality of remote education raise important equity concerns about the potential differential effects by students' race of the current COVID-19 crisis. Our findings underscore the complexity of returning students to in-person learning. Factors other than the availability of learning options appear to play a role in how families and students participate in school during the 2020-21 school year. As the pandemic subsides and vaccinations are administered, concerted efforts by district and local leaders may be necessary to ensure that all students participate in expanded in-person learning.