By Carol Morello
Washington Post Staff Writer
Wednesday, March 31, 2010; A15
He was happy heading the University of Michigan's prestigious Survey Research Center, which documents the attitudes and behavior of Americans from childhood through retirement. Ann Arbor had been his home base since graduate school four decades earlier, shaping him into a world-renowned statistician. If he became director of the Census Bureau, he would have to abandon four students he was advising. And he wouldn't have time to ride his beloved red BMW sports touring bike.
A scientist at heart, Groves understood the political perils of the job: He would be second-guessed and lambasted for the cost, questions and methodology used to count 309 million Americans. He had been in the line of fire before, as an associate census director in the 1990s.
But he also knew that the census is a foundation of American democracy, a once-every-10-years effort used to determine congressional seats and how hundreds of billions of dollars in federal aid is distributed.
Colleagues and family members said he told them, somewhat mournfully, that this was his moment to contribute. "It was payback time for me," Groves said of his decision to accept the job.
Since the nation's first census in 1790, the count has been directed by politicians, academics, businessmen, civil servants and journalists. But never has it been run by someone like Groves, who worked one summer as a prison guard and studied the Portuguese revolution of 1910 before becoming steeped in the minutiae of survey-taking.
On Capitol Hill, his appointment was greeted with suspicion by some Republicans, including Rep. Patrick T. McHenry (N.C.), who said Groves will usher in "the political manipulation of census data for partisan gain."
But in his 10 months on the job, Groves, 61, has managed to disarm congressional detractors, largely by listening to their criticisms and, in some cases, making changes. Now, McHenry said, he is impressed by Groves's performance: "I believe he's shown significant success in his first year, getting the bureau on the right track."Complaints about cost
Groves has deflected several controversies. He severed a partnership with the liberal community organizing group ACORN soon after a video surfaced in which ACORN advisers gave tax tips to conservative activists posing as a prostitute and pimp. He offered an apology for "Negro" being included as an alternative to black or African American on the census question about race and said it probably won't be in the next census.
But complaints about the $14.7 billion cost of the census persist. Several congressmen have questioned a $2.5 million ad that aired during the Super Bowl as part of a $340 million marketing campaign. And Groves was bombarded with e-mails asking why the census spent so much money to mail notices that the forms were coming and, later, postcards that the forms had been mailed. Groves said it will save money in the long run if more people respond because of it.
Although many decisions on the census were cemented long before Groves arrived, his success will be measured in part by how many Americans return the forms. Seventy-two percent mailed in the 2000 Census. As of Tuesday, 50 percent had returned the 2010 Census.
Groves has crisscrossed the country, encouraging people to return their census questionnaires by census day, Thursday. He has ridden an Alaskan dog sled, taken a spin in a NASCAR pace car and posed with Dora the Explorer. On most mornings, he takes the Metro from the downtown apartment he shares with his wife, Cynthia, to the bureau's Suitland headquarters, arriving in time to preside over an 8 a.m. breakfast meeting with senior staff members.
Groves's path to Washington began in Kansas City, Mo., where he was born in 1948. His father was a metallurgical engineer who worked for defense contractors and NASA. The family hopscotched around the country before Groves headed to Dartmouth on a scholarship.
A term paper on recidivism led him to a Vermont penitentiary. When the warden told Groves that he couldn't understand what life inside a prison was really like, Groves said, "I don't know why, I took up the challenge." He grew a mustache to look older and got hired as a guard. He considers the summer of 1968 "the most intellectually rewarding summer of my life."
He struck up conversations with a lifer who had killed his wife. The prisoner lent Groves books he kept in his cell, and they discussed the philosophies of Marcuse and Kierkegaard.Influences
Groves also spent eight months in Portugal as a research assistant for a Dartmouth professor, emerging with an honors thesis on Portugal's 1910 revolution and the realization that he didn't want to be a historian.
When he enrolled in a class on how to analyze surveys, "it changed everything," Groves said. He graduated with a sociology degree in 1970 and went to the University of Michigan, where he would receive two master's degrees in 1973 and a doctorate in 1975, stepping stones to a life in academia.
It was a heady era, when social scientists believed statistical surveys helped policymakers understand and improve people's lives. "By providing good data for good policy, you could contribute to society," said Peter V. Miller of Northwestern University, a classmate of Groves's who is president of the American Association of Public Opinion Research.
Michigan's Institute for Social Research was the epicenter of that ethos. It was founded in the 1940s by men who had done pioneering surveys at the Agriculture Department of farmers. Its first chief was Rensis Likert, a psychologist who developed a five-point public opinion scale still used today, measuring whether people strongly agree or strongly disagree with a statement.
Likert is one of Groves's heroes. "He thought if government got too big in a democracy, it would be controlled by bureaucrats," Groves said. "Surveys were a way to get the voice of the people accurately read. They are a tool of democracy."
Groves has changed the way survey methodologists think about their work. Miller considers Groves "the most important survey researcher in the world." He made his first big splash with a book that compared surveys conducted in person with those conducted over the phone. He has examined how to get meaningful results from surveys when less than half the people contacted answer any questions and how the order in which questions are asked produces different results.
He has co-written eight books and dozens of academic journal articles, with titles such as "The Impact of Nonresponse Rates on Nonresponse Bias."
He is considered an expert in sampling, a technique in which results are adjusted to compensate for people -- often minorities, immigrants and the poor -- who are undercounted. Many Republicans suspect sampling helps Democrats amass political power.
As an associate census director on leave from Michigan, Groves supported using sampling to adjust an undercount in 1990, but the Supreme Court ruled that methodology couldn't be used to apportion House seats. It can be used to draw district lines and to distribute tax dollars, though.
"Statistical readjustment of the census is a complicated matter, and only about 20 people on Earth understand it," Groves said, noting there is no controversy in using sampling to calculate unemployment rates and the consumer price index. "It's like asking if you believe in surgery. It depends what the disease is and how good the surgeon is. That's how I feel. It depends how good the census is, and how good the adjustment is."
One reason he became Census Bureau director is to defend the integrity of government data.
Groves was offended by what happened to Lawrence Greenfeld, who headed the Bureau of Justice Statistics. In 2005, it produced a report that motorists from different races and ethnicities were stopped by police at the same rate but that blacks and Hispanics were more likely to be searched or arrested. His superiors ordered him not to mention the disparities, but Greenfeld posted the report online intact. The White House demoted him.
"Larry was fulfilling the ethical and professional norms of the statistical side of our job, to report findings to the public without commentary," Groves said. Once you fail to do that, "it's game over. Your credibility is destroyed. The public won't believe your unemployment numbers or GDP."
He recalls something that Janet Norwood, a former commissioner of the U.S. Bureau of Labor Statistics, advised him: "Don't let yourself fall in love with the job. Be willing to walk away, or you'll get sucked in to compromising."
He hasn't had to do that and is trying to position the Census Bureau for the 2020 count.
Does the race question need to be re-engineered? Is the household unit outmoded as a way to categorize data, when less than half the households are intact families? What's the cheapest way to reach more than a quarter of all households that simply don't participate, no matter how often they're contacted?
Ultimately, he will be judged by the next census more than this one. "This census will be over, and the politics will drift away," said Ken Prewitt, director of the 2000 Census. "Then he will roll up his sleeves and get to work."
Staff researcher Madonna Lebling contributed to this report.