Census Bureau plan to cut questions from survey puzzles experts
A Census Bureau proposal to cut seven questions from the American Community Survey would make the United States the only industrialized country without annual information on marriage and divorce rates, a demographics expert said.
“That’s just crazy,” said Steven Ruggles, professor of history and population studies at the University of Minnesota.
Five of the seven questions up for elimination deal with marital status, data that can be important in determining government benefits or identifying societal trends. A sixth question asks college graduates in what field they earned a degree.
Eliminating the college degree question could cost taxpayers millions of dollars, said Norman M. Bradburn, a senior fellow at NORC, an independent research organization at the University of Chicago. That’s because the National Science Foundation uses the information to meet a congressional mandate.
“It would cost the taxpayer a lot of money to take it out,” he said. “It wouldn’t save the Census Bureau any money.”
Neither expert could think of a reason to keep the seventh question, which deals with home-based businesses.
No one from the bureau was available for comment, and the agency responded to the Tribune-Review’s emailed questions by providing links to the Federal Register Notice and other documents describing the process it used to propose eliminating certain questions.
The bureau conducted a review to determine how it could reduce the amount of time and effort people spend answering the questionnaire, yet still provide data required by federal law. Though the review determined the seven questions were “low cost” in terms of the burden they placed on people answering them, it found that they were of “low benefit” in terms of using the answers.
Before it can drop the questions, the bureau needs approval from the Office of Management and Budget.
The bureau sends the 72-question survey to more than 250,000 households a month, or about 3.5 million annually. Twenty-four of the questions are about housing; the others address the people living in that housing.
Federal agencies use the responses to determine how the government distributes more than $400 billion. Academics and private companies use the data for everything from tracking societal trends to deciding where to build facilities.
The bureau’s proposal is drawing little notice among city or county agencies. Officials in Pittsburgh and Allegheny County were hard-pressed to find an entity that would miss the data.
Although Chris Briem, regional economist at the University of Pittsburgh’s Center for Social and Urban Research, said he probably could do some “neat stuff” with the numbers, he can’t recall ever using them.
The lack of local government use was a main factor in the bureau’s analysis of which questions to cut, the bureau said in its review.
Ruggles and Bradburn said the focus on small geographies such as neighborhoods and cities misses the big picture on the marriage and college degree questions.
Neither is useful on a local level because the number of people marrying or divorcing in a given year, or the number with degrees in physics, is too small to capture below the state level.
But the data are important on a national level.
The National Science Foundation uses responses to the college degree question to meet a mandate that it track the nation’s supply of scientists and engineers, according to John Gawalt, division director for the foundation’s National Center for Science and Engineering Statistics.
Before the Census Bureau added the question to the American Community Survey, the foundation did its tracking by using a couple of other surveys. One that is released after each decennial census cost about $17 million. The other, which was conducted every two years, cost about $4.5 million.
In addition to saving that money, using the American Community Survey question allows the foundation to account for people who received science and engineering degrees in other countries, Gawalt said.
“It has been very effective in helping us to have higher-quality estimates and in reducing our costs,” he said.
Like any federal statistical agency, the Census Bureau is under pressure to justify the time and effort it requires of citizens responding to its surveys, so Gawalt does not fault the overall intent of streamlining the survey.
“They’re not the bad guys here,” Gawalt said. “They’re doing, really, what they’re supposed to do.”
Yet the foundation hopes to persuade the bureau to retain the college degree question, he said.
As for the marriage data, it is a necessary component for projecting demand for Social Security benefits, Ruggles said.
“It’s almost as important as knowing how many people were born or how many people died,” he said.
By combining the marital status answers with answers on other subjects such as income, education or race, researchers can look in-depth at important trends.
“We are in a period where marriage is changing more rapidly than ever in our history,” he said.
For example, Ruggles projects that about a third of 20- to 24-year-olds will never marry. The maximum for previous generations has been about 8 percent.
At the same time, baby boomers are divorcing at record high rates as they hit retirement age.
“The American Community Survey gives us a tool to understand why that is,” Ruggles said.
Bradburn and Ruggles said there are other questions that could be pruned from the survey.
“One of the most annoying questions I know is: ‘What is your average gas bill?’ ” Ruggles said.
People don’t like answering questions about their finances, and in this case, the government could get more accurate information from utilities.
“That would probably be a lot cheaper than asking individual people,” he said. “It’s important information for energy policy, but it doesn’t have to be in this survey.”
Though he’s not sure what other questions he would delete, Ruggles said, he believes the agency could cut seven questions “without doing serious damage to the statistical system.”
Bradburn nominated the question on indoor plumbing. Though it was an important measure of infrastructure needs several decades ago, it has become obsolete — and annoys respondents.
“There are three or four that everybody would be happy to (have) out,” he said.
Brian Bowling is a Trib Total Media staff writer. Contact him at 412-325-4301 at [email protected].