As explained in a previously released blog, Pandemic Impact on 2020 American Community Survey 1-Year Data, the 2020 American Community Survey (ACS) data collection operations were significantly impacted during the COVID-19 pandemic. These disruptions prevented us from collecting information from certain segments of the population. This resulted in substantial nonresponse bias in the 2020 ACS data, which meant the characteristics of people who responded to the survey were significantly different from the people who did not respond.
To address nonresponse bias in the 2020 ACS, we created and released experimental 1-year weights as part of the 2020 Public Use Microdata Sample (PUMS). However, as shown in the working paper Addressing Nonresponse Bias in the American Community Survey During the Pandemic Using Administrative Data that accompanied that release, the experimental estimates for 2020 were not directly comparable to prior ACS data. To give one example from Appendix Table WXK201501 for educational attainment, using the 2020 experimental weights, we estimated the share of people 25 and over with a bachelor’s degree was 35.1 percent. Using the same weighting methodology (i.e., the experimental weights), the estimated change from 2019 to 2020 in the share with a bachelor’s degree was 0.74 percentage points. However, comparing the 2020 experimental weights to the 2019 production weights, we estimated a change of 1.95 percentage points, of which 1.21 percentage points was due to the difference between the estimate in the 2019 experimental weights compared to the 2019 production weights.
Many comparisons are available in the appendix tables that accompanied the working paper. To allow researchers to make other comparisons, however, we are releasing a public-use experimental weight file for 2019. This file contains only the weights and the identifiers needed to merge the weights to the 2019 ACS PUMS files.
The 2019 experimental public-use weights were generated by weighting the existing 2019 ACS PUMS files in each state to match survey estimates generated from the full ACS sample using the experimental weights. The procedure was also discussed in detail in Section 5 of the working paper on nonresponse bias in the Current Population Survey Annual Social and Economic Supplement (CPS ASEC), Coronavirus Infects Surveys, Too: Survey Nonresponse Bias and the Coronavirus Pandemic. The 2019 ACS 1-Year Public Use Microdata Sample with Experimental Weights can be accessed on the ACS File Transfer Protocol (FTP) server.
The 2019 ACS 1-Year PUMS experimental weights were approved for release by the U.S. Census Bureau, authorization number CBDRB-FY22-232.