FRIDAY, SEP 27, 2019
Although Americans only fill out a decennial Census form once every 10 years, the Census Bureau gets information about the nation’s population from a variety of surveys — including a massive effort that asks hundreds of thousands of Americans every month about their age, race, gender, housing, origin, kitchen and plumbing status.
Despite its prevalence, the vast majority of Americans are oblivious to the existence of the American Community Survey (ACS), a flagship data collection effort aimed at almost 300,000 U.S. households every month. The ACS, which replaced the decennial “long form” sent to 1 in 6 households until 2000, has been used since 2005 to collect a broad range of economic, housing, transportation, education, ancestral — and citizenship — data for the entire country.
(The entire range of ACS data, from the initial responses received during the 2006 survey to 2018 estimates released earlier this month, is available for customizable analysis and mapping from Social Explorer.)
Most Americans know that the decennial Census is used to reapportion seats in the U.S. House of Representatives and to award Electoral College votes to individual states, based on population shifts during the preceding decade. The 2010 Census was a simple questionnaire, with 10 questions on one page for a single person, covering housing status, relationship status, age, gender, race, and ethnicity.
More complicated and less well-understood, however, has been the ACS, an annual survey that’s used to help the government distribute almost $900 billion in federal funds every year. Instead of a single page with 10 questions for each person in a household, the ACS adds an extra seven pages with roughly 70 additional questions. The answers are tallied and key demographics are used to calculate amounts that the government will spend on programs such as Medicaid, the health insurance program for the poor, disabled, and young; Title I funding, money used to educate children from low-income families; and the food stamp program, which provides nutrition assistance for more than one in nine Americans.
The ACS was created primarily to support federal programs, but it’s become known as “the flagship survey,” according to Census Director Steven Dillingham, and an invaluable tool for constituencies that include major corporations, trade industry organizations, social science researchers, and journalists. It’s the sole reliable source for figures that include educational attainment, commute lengths, primary home heating fuel, ancestry, marital status, health insurance coverage, economic mobility, and the presence of complete plumbing facilities, among other data. The nation’s most powerful trade organization, the U.S. Chamber of Commerce, called its data “vital for monitoring trends in the economic, social, and demographic landscape at a local level.”
The decennial Census traditionally was viewed as a vital but limited product for researchers, covering only age, race, ethnicity, gender, relationship, and housing characteristics. To supplement it, the Census Bureau began issuing a more detailed “long form” questionnaire in 1940, eventually reaching approximately 1 in 6 U.S. households.
Although the long-form Census questionnaire had offered an extraordinarily accurate picture of the nation, it suffered from a fatal flaw: Near-immediate obsolescence. The long-form results were usually released 18 months after the actual Census. Depending upon circumstances, important data points such as median income, educational attainment, commuting patterns, and even housing characteristics were outdated even before they were released. When the 2000 Census was taken, the economy was growing at a healthy 4.1 percent annual clip; by the time the first long-form data was released in late 2001, growth had slowed to a one percent annual rate in the wake of the collapse of a tech bubble and the 9/11 terror attacks in New York and Washington, D.C.
Because the long-form Census was too expensive to conduct annually, the Census Bureau settled upon a rolling survey as an affordable and accurate solution. A single-year survey based on more than three million households would provide adequate data for places with 65,000 people or more (such as the one-year 2017 ACS); a three-year survey with roughly 10 million households would include coverage for places with more than 20,000 people (a three-year ACS); and a five-year survey that included more than 15 million households would provide coverage for all places in the United States, down to a block group-level (a five-year 2013-17 ACS). The first single-year survey was released in 2005. The first three-year series followed in 2007 (before being eliminated to save money after the 2011-13 ACS), and the initial five-year ACS was released in 2009.
The ACS also remedied another important problem with the long form. Since it was an annual survey, questions could be changed from year to year. The annual nature of the ACS made it far more flexible than the long form, which could only be changed once every decade. Since its first release, the ACS has added questions to reflect changing priorities, such as the availability of high-speed Internet. For the 2019 survey alone, the Census Bureau added questions to clarify same-sex relationships, cell phone service, health insurance premiums, and the ability of people to work at home.
While the ACS has added greatly to the Census Bureau’s ability to offer more current and detailed data about the United States, the survey has also been criticized and came close to being eliminated in 2012, when the Republican-led House of Representatives voted to abolish the survey. Conservatives argued that the ACS was intrusive and unconstitutional. After a series of protests by business organizations, social science researchers, and economic development agencies, the Senate declined to take up the House proposal.
Even with the outpouring of support, Congress has continued to chip away at the ACS. A 2014 House proposal to make the ACS voluntary was thwarted after business and research organizations argued it would effectively destroy the accuracy of the survey. Federal law requires people to respond to the ACS, but no one has ever been prosecuted for refusing to answer the ACS questionnaire. Congress backed down on the idea of making the responses voluntary, but it still slashed funding, forcing the Census Bureau to cancel the three-year ACS product in 2015.
The Census Bureau has also responded to the tighter resources by combining the ACS responses with administrative records from other federal agencies, such as the Internal Revenue Service, Health and Human Services Department, and Social Security Administration. The secondary records have helped the Census Bureau reduce its reliance upon in-person follow-up visits for households that don’t respond to mailings or telephone calls. In addition, the Census Bureau began offering households the option to respond via the Internet.
The future of the ACS is far from guaranteed. The political attacks helped drop the response rate to 93.7 percent in 2017, its lowest level for a full year since the ACS became operational in 2005. A record 2.7 percent of households simply refused to participate in the survey. Yet, it remains one of the few government programs with a lengthy track record of support from conservative and liberal organizations.
“Without good data, policymakers are essentially flying blind, lacking solid knowledge of the Americans they are seeking to assist,” Andrew Biggs, a scholar at the conservative American Enterprise Institute, told a House oversight committee in 2012. “We don’t learn any of this information by osmosis,” said Wade Henderson, president and chief executive of the liberal Leadership Conference on Civil and Human Rights. “We know it because of the Census Bureau, through the ACS.”
Author: Frank Bass