By on 6.29.23 in Carolina Demographics, Census 2020

Achieving a complete and accurate count in the U.S. Census is not simple. The biggest challenge in conducting the census is getting households to mail back or complete their forms online or by phone, a process known as self-response. Higher self-response rates reduce the overall costs of conducting the census. And, when individuals self-respond, the data is more accurate.

Before the 2020 Census, Carolina Demography partnered with the NC Counts Coalition to create a map that identifies communities across North Carolina that were most at risk of being undercounted in the 2020 Census.

Who is more likely to be undercounted?

Historically, certain population groups, referred to as historically undercounted communities, have not been fully counted and represented in the Census count. There are four main reasons why a population may be hard to count:

  1. They might be hard to locate, like people who are more transient and have incomplete address listings.
  2. They might be hard to contact, like people who live in gated communities.
  3. They might be hard to interview, meaning they have low literacy or Limited English Proficiency (LEP).
  4. They might be hard to persuade, meaning they are suspicious of the government or don’t see a benefit in participating in the Census.

Historically undercounted populations in North Carolina include:

  • Young children under age 5
  • Hispanic or Latin-X individuals
  • American Indian/Alaska Native individuals
  • Black/African American individuals
  • Frequent movers
  • Renters

In an analysis written for EdNC, Carolina Demography’s former director Rebecca Tippett noted that “some groups—like college students or people who own two homes—are more likely to be counted more than once, while other groups—like young children or people who have recently moved—are less likely to be counted at all.”

Looking at county-level data from the 2010 and 2020 Censuses

We recently looked at data from the most recent two Census counts to assess which counties in North Carolina had persistently low self-response rates.

There are 100 counties in North Carolina. Of the 10 counties with the lowest self-response scores in the state as reported in the 2010 Decennial Census, eight remain in the bottom 10 percent of self-response scores in the 2020 Decennial Census. The counties include Graham, Montgomery, Jackson, Swain, Avery, Macon, Alleghany and Dare.

These counties are primarily located in the western and coastal regions of the state. Graham, Jackson, Swain and Montgomery counties experienced an increase in self-response rates between the 2010 and 2020 Census, but remain in the bottom 10 percent of self-response rates as of 2020.

The top self-response rates by county in 2020 were Union (74.8%), Wake (73.3%) and Orange (72.9%).
Of the eight counties that have had persistently low self-response rates, some have characteristics that increase the likelihood of an undercount, like hard-to-count housing. For example, people living in overcrowded housing units are commonly undercounted.

The ACS uses two criteria to define overcrowded housing:

  1. More than 1 person per room in a housing unit.
  2. The housing unit has fewer bedrooms than the ACS considers appropriate for the size and composition of the household.

Residents of overcrowded housing units tend to be undercounted because the "Head of Household" typically responds for everyone living in the household. If there are multiple families occupying one housing unit, who may or may not be related, it creates more risk for occupants of these dwellings to not be enumerated in the Census.

Historically, the Census Bureau has also found it more difficult to identify addresses for certain housing unit types, like mobile homes because they may have incomplete address listings and residents may change addresses more frequently. Renter-occupied housing units have a historically higher undercount for similar reasons.

Both Swain and Montgomery counties surpass the state average in the 2010 Census and the 2020 Census in overcrowded rental, rental, and mobile home occupancy in both the 5-year American Community Survey (ACS) 2010-14 and the 5-year American Community Survey (ACS) 2016-20. However, the patterns are not always clear. Alleghany County, although in the bottom ten percent of self-response rates in both the 2010 and 2020 Census, has rates of overcrowded rental, rental and mobile home occupancy below those of the state average across years.


Other factors which increase the likelihood of an undercount include demographic indicators such as race and ethnicity, age, and whether someone was born outside of the United States. In a preliminary analysis conducted by Carolina Demography, we found no clear relationship between changes in self-response rates and changes in BIPOC residents, foreign-born residents, and young children in these eight persistently low self-responding counties.

The most common thread between these counties is their rural classification. A Census block, the smallest geographic area for which the Bureau of the Census collects and tabulates data, is classified as rural if it does not meet the criteria for urban. In a recent blog post, we discussed changes to how the US Census Bureau defined urban areas for the 2020 Census.  The definition of an urban area used during the 2020 Census is as follows: a densely settled core of census blocks that met minimum housing unit density and/or population density requirements.

All counties that have had persistently low self-response rates in the 2020 Census, with the exception of Dare County, have rates of the population living in a rural area above the state average of 63.95%. Five (Alleghany, Avery, Graham, Montgomery, and Swain) of these eight counties are designated as having exclusively rural populations.

Republish our content for free under a Creative Commons license.

Need help understanding population change and its impacts on your community or business? Carolina Demography offers demographic research tailored to your needs.

Contact us today for a free initial consultation.

Contact Us

Categories: Carolina Demographics, Census 2020


Featured projects

Your support is critical to our mission of measuring, understanding, and predicting population change and its impact. Donate to Carolina Demography today.