As 2020 Census Nears, Focus Turns to Reaching the Hard-to-Count

New reports shed some light on the places and the people the government has trouble counting.
by | October 30, 2017
Low-income neighborhoods, rural regions and those with large concentrations of minority groups often fail to respond to Census questionnaires. (Flickr/U.S. Department of Agriculture)

Every 10 years, Americans are asked to fill out and return their Census questionnaires. It's an important decennial event, given that population counts guide billions in federal spending, determine congressional apportionment and play a key role in shaping future policies.

Yet vast segments of the population often fail to respond. Greater racial and ethnic diversity, more nontraditional living arrangements, elevated poverty rates and a litany of other factors are also putting more people at risk of not being counted in 2020.

As state and local governments start to prepare for the next Census, making sure these communities participate is a priority. Newly released reports shed light on the places and the people least likely to be counted.

The places with low response rates aren’t confined to just a few regions of the country. The City University of New York’s Center for Urban Research recently published an online map identifying "hard-to-count" Census tracts and found them in every state in the country.

 

(Red and orange Census tracts had lower 2010 Census mail return rates.)

 

Census participation is measured several ways. The most precise is what’s known as the mail return rate, which compares completed survey forms to numbers of occupied housing units that received questionnaires.

The national mail return rate for the 2010 Census was 79.3 percent. States in the upper Midwest returned their forms at the highest rates. Minnesota’s mail return rate of 85.6 percent was the nation’s highest. By comparison, less than 75 percent of households responded in Alaska, Louisiana and New Mexico.

 

 
2010 Mail Return Rate (%)
SOURCE: U.S. Census Bureau
 

Smaller rural and urban areas, particularly some counties and individual neighborhoods, completed their forms at much lower rates. About half of households returned their forms in a few predominately rural counties. (View county data.)

Much of the disparities in Census participation are tied to demographics.

African-American and Hispanic neighborhoods, especially those with limited English proficiency, have historically been more difficult to count, as have low-income communities. Areas with more renters and transient households similarly tend to participate at lower rates.

Young children are another demographic that’s more difficult to count. In many large cities, most live in low-responding neighborhoods. According to calculations from the Leadership Conference Education Fund, about 96 percent of Newark, N.J., children under age 5 reside in hard-to-count Census tracts. The rate similarly exceeds 70 percent in Buffalo, N.Y., Cleveland and New Orleans.

Enumerators follow up with nonresponding households, but aren't always successful. Visiting homes in-person comes at a much higher cost for taxpayers and is one reason why the Government Accountability Office reported the 2010 Census cost approximately $98 per housing unit. For the next decennial count, the Census Bureau has set a goal of not exceeding this cost, without adjusting for inflation.

One major change that could help the agency curb costs is that, for the first time, responses will be submitted online. Respondents will be able to enter their information using a unique ID code mailed to their homes or by providing identifying information that matches administrative records. The agency expects the majority of Americans to file electronically rather than mail in their forms.

But this added option could still fail to reach many low-income households because research suggests that nearly half of the poorest households -- those with annual incomes below $20,000 -- still lack internet access. An analysis by the Center for Urban Research found congressional districts with lower mail return rates were correlated with higher percentages of households lacking internet subscriptions or having poor connections.

Low-income households are more likely to use mobile phones to get online, which presents another potential hurdle for the 2020 Census. Mobile devices have longer load times, and respondents are more likely to not complete surveys than those using desktop computers.

All of these issues are compounded by limited federal appropriations that haven’t mirrored increases in the years leading up to prior Censuses. As we’ve reported previously, some have expressed concerns that underfunding the Census could potentially hinder its accuracy and lead to undercounting of disadvantaged communities.

A report last week by the Leadership Conference Education Fund and Georgetown Center on Poverty and Inequality outlined recommendations for the Census Bureau and local governments to prepare for the 2020 Census. They included setting aside computers in public facilities for completing the Census and emphasizing participation using social media, email and text messaging. In a call with reporters, Indivar Dutta-Gupta, co-executive director of the Georgetown think tank, also said localities should partner with the Census Bureau, share data with the agency and participate in what are known as complete count committees.

Officials are further encouraged to assist the Census Bureau in updating its list of residential addresses for inclusion in the upcoming count. Census officials on Friday reported that more than 6,000 local governments have signed up to help update the list. Localities have until Dec. 15 to register.

“There’s a lot that can happen now,” Dutta-Gupta says. “In many cases, we’re at a pretty critical point where people shouldn’t be waiting much longer.”

 

 Census Mail Return Rates for U.S. Counties

Mail return rates represent final rates after all initial and replacement survey questionnaires were received. Initial mail response rates recorded in April (not shown) are typically lower.

SOURCE: U.S. Census Bureau; Download data