16.1 Introduction

The big picture regarding the accuracy of the US Census since 1950 is a good news/bad news story. Over the past 60 years, the overall accuracy of the US Decennial Census has steadily improved, but some groups have persistently experienced higher net undercounts than other groups in the Census and some new differential undercounts (for example, young children compared to adults) have emerged over this period.

There is also a good news/bad news story associated with the 2010 Census coverage as well. The good news is, the coverage error for the total population was very small by international and historic standards, but the bad news is there are still large coverage differences among groups. Data presented in this book underscore the extent to which some groups have higher net undercount and omissions rates than others in the Census.

Who is missing? Based on the groups for which the Census Bureau provides net undercount and omissions data, there are a handful of groups that are at exceptionally high risk of being missed in the Census. To put the data below in context, the Non-Hispanic White Alone population had a net overcount of 0.8% and an omissions rate of 3.8% in the 2010 Census The groups listed below are those with the highest net undercount and omissions rates based on Census Bureau data:

  • Young children (age 0–4) had a higher net undercount and higher omissions rate than any other age group in the 2010 Census. In the 2010 Census, the net undercount rate for young children was 4.6% and the omissions rate were 10.3%. The net undercount rate for young children has increased rapidly since the 1980 Census.

  • The Black population had a higher net undercount than any other race/Hispanic group in the 2010 Census. The net undercount rate for the Black population was 2.1% (2.5% based on the Demographic Analysis method) and the omissions rate was 9.3%. Black males age 20–60 have exceptionally high net undercount and omissions rates. The net undercount rate for Black males age 30–49 was 10% and the omissions rate for this group was 16.7%. Historically, the Black population experienced high net undercount rates in the Census. While the net undercount of Blacks has decreased over time, it is still relatively high and the differential undercount between Blacks and Nonblacks has improved little since 1940.

  • Hispanics had a net undercount rate of 1.5% in the 2010 Census and an omissions rate of 7.7%. Hispanic males age 20–50 had very high net undercount and omissions rates. The net undercount rate for Hispanic males age 30–49 was 5.1% and there was an omissions rate of 10.9% in the 2010 Census.

  • In the 2010 Census the net undercount rate for American Indians living on reservations was very high at 4.9%.

One of the themes that flows through the various groups that are most likely to be missed in the Census is social and economic marginalization. Higher income groups like Non-Hispanic Whites are counted accurately or have over counts, while less affluent groups like Blacks, Hispanics, and American Indians on reservations have significant undercounts. In addition, young children, who have a high net undercount, are politically powerless. These groups are often under-represented in terms of political power and/or civic participation as well. In some ways, the Census undercounts of these groups is just one more way they are sidelined and under-represented in society. Given the connection between Census counts and federal funding, the groups most in need of federal assistance are the groups least likely to get their fair share because they have the highest net undercounts and omissions.

16.2 Net Undercounts and Omissions

Another point that was made repeatedly in this book is the fact that net undercounts and omissions are not the same thing. In some cases, a low net undercount rate for a group might lead people to assume no one was missed in the group but that is often not the case. For example, the net undercount for Asians was zero in the 2020 Census, but Asians had a 5.3% omissions rate which is somewhat higher than the rate for Non-Hispanic White Alone (3.8%).

There was a very small net undercount for American Indians and Alaskan Natives Alone or in Combination in 2010 but the omissions rate for American Indians and Alaskan Natives was 7.6% which is double the rate for Non-Hispanic White Alone (3.8%). The count of Native Hawaiian and Pacific Islanders in the U.S. Census is relatively accurate but the omissions rate for Native Hawaiian and Pacific Islanders (7.9%) is about double the rate for Non-Hispanic White Alone (3.8%).

16.3 Cumulative Impact

One point woven into the analysis in previous Chapters is how the accumulation of risk factors has a multiplicative impact on net undercount rates. Table 16.1 shows net undercount rates for four individual factors (age, race, sex, and tenure) and then shows the net undercount rate when all of these factors are combined. None of the net undercount rates for individual factors is more than 2.5%, but collectively the net undercount rate for someone with all four factors is 12.2%. If we had more data, I strongly believe the addition of other factors like poverty, language ability, and household structure would further drive up the net undercount rate for people with multiple risk factors.

Table 16.1 Undercount rates for individual factors and a combination of factors
Full size table

16.4 The 2020 Census

It is worth repeating that Census data are the backbone of our democratic system of government. The significance of the Census envisioned by the founding fathers is reflected in the fact that Decennial Census is mentioned in the sixth sentence of the Constitution and is the first responsibility given the new federal government.

Chapter 2 provides detailed information about the many ways the Decennial Census data are used. Census-related figures are used to distribute more than $850 billion in federal funding each year to states and localities. Data from the 2020 Census will guide the distribution of roughly $25 trillion in federal assistance to states and localities between 2021 and 2030. Countless decisions in the public and private sectors are based on Census data. Moreover, the impact of flaws in the census counts often last a decade because population estimates, projections, and survey weights, are derived from Census counts

Chapter 15 in this book outlines some of the difficulties that have hindered 2020 Census planning. But much of what will determine the accuracy of the 2020 Census will depend on what happens after this book is published.

The accuracy of the 2020 Census will depend on the balance between two broad set of forces. Many factors like lower response rates to surveys, growing distrust of the federal government, dependence on new methodologies, and underfunding of the 2020 Census would lead one to expect the accuracy of the 2020 Census will be worse than in 2010 and differential undercounts will increase. But the mobilization of groups outside the Census Bureau is more advanced than ever before. Growing numbers of organizations are recognizing the importance of the Census for their work and for the well-being of their community.

I believe the results of the 2020 Census will largely be determined by the balance between these two sets of positive and negative forces. To the extent that advocates, grass roots leaders, elected officials, and others mobilize their communities to be counted, the results of the 2020 Census can be better than 2010 with respect to the net undercounts and omissions in hard-to-count populations and communities. But convincing people to respond to the Census given the current political climate will not be easy. To the extent that the Administration and Congress soften the negative rhetoric and policies aimed at immigrant and minority communities it will be easier to convince people to respond to the Census.

16.5 What Can You Do?

This book has largely been a scholarly research effort. But I want to end with a few notes on how readers can easily get involved in making the 2020 Census a success.

Individuals or organizations can get involved as a Census Bureau Partner by signing up at www.census.gov/partners. Partners will get updated information from the Census Bureau and will receive advice about census-promotion activities. Through this mechanism one can get involved in national, state, and local activities to promote the 2020 Census. One example of the help provided by the Census Bureau, is the community outreach toolkit that is available from the Census Bureau at https://www.census.gov/partners/toolkit.pdf.

One can also join (or start) a state or local Census Complete Count Committee. Complete Count Committees are made up of volunteers and are established by tribal, state, and local governments and community leaders or organizations to increase awareness and motivate residents to respond to the 2020 Census. To learn more about Complete Count Committees or find one near you, contact the Complete Count Committee Program at the Census Bureau at https://www.census.gov/programs-surveys/decennial-census/2020-census/complete_count.html.

One of the best ways to stay up-to-date on Census-related developments in Washington, is to sign up for regular alerts at foundation-funded Census Project. One can also visit the Census Project website to find information on past census-related events https://thecensusproject.org/.

Another site with a lot of good information on hard-to-count populations is The Leadership Conference Education Fund. Their website is https://civilrights.org/census/. Material on the Leadership Conference Education Fund website includes topical briefs and data on a lot of hard-to-count populations.

For those interested in making sure all children are counted accurately in the 2020 Census visit the CountAllkids Complete Count Committee website (www.countallkids.org) or the KIDS COUNT website at the Annie, E. Casey Foundation (www.kidscount.org).