Census 2020: Current Response Rates and the Risk of an Undercount

Census 2020: Current Response Rates and the Risk of an Undercount

By Deborah Kennedy, president, National Coalition for Literacy, and
Mary Margaret Kraut, affiliated faculty, Union Institute and University

Here’s the good news: As of yesterday morning, the national self-response rate for the Census was 53.4 percent. According to the Census Bureau’s Self-Response Rankings website, the rate is above 60 percent in Minnesota, Iowa, Nebraska, and Wisconsin. Los Alamos County in New Mexico has the highest rate of response by county at 73.8 percent, and North River, North Dakota has the highest rate of response by city at 100 percent.

These numbers are encouraging, but the data also shows some troubling trends. Table 1 provides comparative data on total response rates, including online, by phone, and on paper, for the Census tracts with the highest (top 20%) and lowest (bottom 20%) rates of self-response. [Note that the table specifies “population” or “household” to reflect the way the Census Bureau reports the data.] As the table indicates, the tracts in the bottom 20% contain substantially higher proportions of hard-to-count populations, including minorities, non-native English speakers, those living below the federal poverty line, and those with no household internet access, than the tracts with the highest response rates. The risk of an undercount for these populations is evident in the data.

The data in Table 1 represents the majority (95.45%) of households: those that have received Internet First and Internet Choice mailings. These two contact strategies take place where mail is delivered to the physical location of the housing unit.

  • In the Internet First contact strategy, a household receives a letter asking the respondent to go online to complete the Census.
  • In the Internet Choice strategy, a household receives a paper questionnaire along with the invitation letter; the Census Bureau uses this strategy in areas that it considers less likely to respond online.
  • In tracts where the Bureau recognizes at least 20% of households as needing “Spanish assistance,” Internet First and Internet Choice mailings are bilingual (Spanish and English).

To date, response rates for Internet Choice tracts are not as robust as those for Internet First tracts. As of April 23, the average response rate for Internet First tracts was about 4 points above the overall national rate, whereas the average response rate for Internet Choice tracts was about 6 points below the overall national rate. Table 2 provides a demographic comparison of the two types of tracts. The comparison shows the relatively higher proportion of hard-to-count populations in the Internet Choice tracts and demonstrates the risk of an undercount for these populations.

To gather information from households that have not responded, the Census Bureau conducts the Non-Response Follow-Up Operation (NRFU), in which Census enumerators visit households in person. Census Outreach [https://www.censusoutreach.org/census-timeline] summarizes the NRFU process this way:

  • If no one answers when an enumerator visits a household, a “Notice of Visit” will be left at the door. This notice will include an online response code to encourage households to self-respond.
  • After one unsuccessful attempt to contact a household, the Census Bureau will determine if the household can be counted using high quality federal administrative records. Households that do not meet this standard will receive additional visits from enumerators.
  • After the 3rd unsuccessful attempt, enumerators can ask nearby reliable “proxy” (for example, a landlord, neighbor, caregiver, letter carrier, or on-site utility worker) for details about the household. If there is no proxy available, enumerators will continue to visit the household up to 6 times.
  • After the 6th unsuccessful attempt, some special case households will still be eligible for additional visits through the end of October.

The 2020 Census Self-Response Operation will continue until October 31, 2020, and the Census Bureau now plans to conduct NRFU from August 11 through October 31 (these dates have been recently adjusted, and may change again depending on COVID-19 developments). The uncertainties associated with face-to-face interviews, particularly in the context of COVID-19 concerns and restrictions, make self-response a far more reliable way to ensure that all people are counted.

The hard-to-count populations that are at risk of being omitted from the Census count have substantial overlap with those who participate in, or could benefit from, adult education. These adults need to recognize the relationship between their participation in the 2020 Census and the availability of equitable federal funding for schools, roads, public assistance, and health services for themselves and their children. Such funding can give adult learners the opportunity for literacy instruction, classes to obtain a high school equivalency or diploma, English for non-native speaker classes, or community college courses to obtain a certificate required for employment. The commonality across all of these programs is for the community to receive its fair share of federal funds.

Adult education practitioners are well connected with the communities where hard-to-count populations reside, and as trusted community resources they can be influential. The challenge for them is to reach out actively to their adult learners in all Census tracts and use that influence to ensure that every individual is included in the 2020 Census count. 



To see current response rates and rankings, see the Rankings Dashboard

To see current response rates by state, county, and Census tract, see the Response Rates Map and the Census 2020 Hard to Count Map

For information on Internet First and Internet Choice contact strategies by state and Census tract, see the Census Bureau’s Mail Contact Strategies Viewer

One thought on “Census 2020: Current Response Rates and the Risk of an Undercount

Comments are closed.

Comments are closed.
%d bloggers like this: