To provide a complete picture of the nation’s population, the American Community Survey (ACS) collects information from populations living in households and group quarter (GQ) facilities. Federal prisons are one type of GQ facility. To collect data from inmates at federal prisons, ACS field staff visit these facilities to drop off paper questionnaires for inmates to complete. Due to the COVID-19 pandemic, federal prisons suspended visits and we were unable to deliver paper questionnaires. Without these data, inmates in federal prisons risked being underrepresented in the 2020 ACS. To help address this, we used administrative data from the Bureau of Prisons (BOP) in lieu of response data for the federal inmates included in our 2020 ACS sample. These data provided the following characteristics: sex, age, race, ethnicity, and citizenship. The BOP data do not cover all the items asked on the ACS; however, we were able to use these five characteristics in our edit and imputation procedures to help impute values for the remaining characteristics for which we did not have administrative data.

The BOP data was not a perfect replacement for ACS response data for some items. For example, the BOP data for race was limited to four racial classifications (White, Black or African American, American Indian or Alaska Native, and Asian), while the ACS includes two additional classifications (Native Hawaiian or Other Pacific Islander, and Some Other Race). Additionally, the ACS allows respondents to choose more than one race and be classified in a seventh category (Two or More Races). The BOP data only classify people in one race category.

This methodology was tested prior to use in the 2020 ACS and was determined to be a better option than using statistical methods alone to estimate the characteristics of federal inmates. Federal prisons continue to suspend visits; therefore, administrative data from the BOP will also be used to capture this population for the 2021 ACS.