Loading [Contrib]/a11y/accessibility-menu.js
Bates, Nancy, and Joe Zamadics. 2021. “COVID-19 Infection Rates and Propensity to Self-Respond in the 2020 U.S. Decennial Census.” Survey Practice 14 (1). https://doi.org/10.29115/SP-2021-0002.
Download all (4)
  • Table 1:** COVID-19 county-level infection rate quartiles by 2020 Census self-response rates.\
  • Table 2: OLS regressions of 2020 Census Self-Response Rates (census tract level) on specified predictor variables
  • Figure 1: Graph of COVID-19 county-level infection regression coefficient from the Cumulative 2020 Census Self-Response OLS Model
  • Table 3: Contribution of independent variables in the cumulative response rate model.

Abstract

Survey methodologists have acknowledged that the social environment may influence survey and census participation—both at the societal level and at the community level. The 2020 COVID-19 pandemic affected daily life throughout the entire U.S., but to differing degrees depending upon the particular neighborhood or community. In this article, we use U.S. Centers for Disease Control county-level COVID-19 infection data coupled with U.S. 2020 Decennial Census response rate data to explore whether this societal level pandemic influenced participation in the census. We found that even when controlling for covariates predictive of COVID-19 infection rates (e.g., percent minority population, age 65+), infection rates were found to be significantly (negatively) associated with self-response.

Accepted: January 26, 2021 EDT

Appendix

List of Variables Included in the Model

Dependent Variable

Cumu. RR: Cumulative response rate. Percent of eligible households in a tract that had self-responded to the census as of July 25, 2020.

Independent Variables

Cases-rate: The tract’s county cumulative positive tests (as of July 25, 2020) divided by the 2017 reported county population (Source: CDC).

% Hispanic alone: The PDB reported percentage of persons in a tract reporting to be Hispanic alone.

% Black African American in combo: The PDB reported percentage of persons in a tract reporting to be Black/African American alone or in combination with another race.

% Asian in combo: The PDB reported percentage of persons in a tract reporting to be Asian alone or in combination with another race.

% American Indian or Alaska Native in combo: The PDB reported percentage of persons in a tract reporting to be American Indian/Alaska Native alone or in combination with another race.

% White low education: The PDB reported percentage of persons in a tract reporting to be White alone or in combination with another race and no college education.

% Spanish speakers: The PDB percent of households in a tract reporting to speak Spanish, with no one aged 14 or over speaking English very well.

% Vacant units: The PDB percent of households in a tract reported as vacant.

% Renter occupied households: The PDB percent of occupied households in a tract not owner occupied (including those rented for cash and those occupied without rent payment).
Median HH income: The PDB Census tract median household annual income.

% Age 65+: The PDB percent of persons in a tract reporting as aged 65+.

% Age 18–24: The PDB percent of persons in a tract reporting as aged 18–24.

% Female House: The PDB percent of households in a tract reporting as female headed, no spouse present.

Density: ([The PDB 2017 population in a tract /100] / the PDB tract land area in square miles).

Internet Choice: If tract has distinction as Internet Choice or Internet First (Internet choice set as reference segment). Source: Census Bureau.

States: Fixed effects (factor variable) for each state including DC (Alabama set as reference state).


  1. 78.8 percent of Census tracts were designated as this type of enumeration area. Examples of areas not designated to receive mailed materials include those that either do not have mail delivered to the physical location of the housing units and extremely remote areas such as southeast Alaska and select American Indian reservations that requested personal visit enumeration.

  2. As of July 25, 2020, 62.5% of households responded to the census. Of the responses, 80% were completed on the Internet, 19% were paper responses, and 1% were completed by telephone.