Behind The Scenes: Surveying Americans' Attitudes and Beliefs

This Behind the Scenes article was provided to LiveScience in partnership with the National Science Foundation.

For nearly four decades, the General Social Survey has been asking Americans how they feel about critical societal issues such as civil liberties, crime and violence, tolerance, morality, race, stress and happiness. It may be the most important social science survey you've never heard of.

It is second only to the U.S. Census in being the most frequently analyzed source of social science information for researchers, students and agencies of the federal government. But unlike the Census, which most citizens know as the entity that counts the number of U.S. residents every ten years, the GSS, with a much lower public profile, digs beneath the surface of society to gain a richer understanding of Americans' attitudes and beliefs.

"Society is more than its dollars and head counts," said Tom Smith, director and principal investigator of the GSS, the flagship poll of the National Opinion Research Center of the University of Chicago."We are trying to cover a wide range of attitudes, beliefs and attributes of the American population. The study emerged out of the social indicators movement of the 1960s, a desire to go beyond basic Census demographics. We want to know what changes—and why. The core basic goal is to understand longtime societal change."

The most recently released data, from 2010, show that support for gay marriage continues to increase—a question that wasn't even asked in 1972—and tolerance and support of civil liberties is also rising. "In the American population, there is a growth in the idea of acceptance of groups where respondents may not share their positions," Smith said. In race relations, "it's exactly the story you would expect, given that we have our first African American president," he said.

Support for tougher gun control is high, a reading "which hasn't changed in 20 years," Smith said. Also, individuals remain happy with themselves, and with their jobs. "This surprises people because of the economic downturn, but there has been little fluctuation," Smith said.  

"We ask people if they are 'very happy,' 'pretty happy,' or 'not very happy.' A little over 30 percent say they are 'very happy,' and about 60 percent say they are 'pretty happy.'

"Job satisfaction also has been very stable over time, so all this talk about the impact of burnout and downsizing, in the aggregate, doesn't affect peoples' satisfaction with their jobs," he adds. "The picture of psychological well-being has been very stable."  

Until the 1990s, the study was conducted annually. Now it's done every other year. It's also gone global. In the 1980s, the GSS began working with similar organizations in other countries to compare American data to information gathered abroad. Today, GSS scientists collaborate with 47 other countries.

Most of its funding comes from the National Science Foundation, which provides about $3 million a year.     

Nearly 5,000 citizens are randomly selected to participate. In 2008, the survey began to include a group of prior respondents, re-interviewing a cross section of them to examine whether these individuals have changed their behavior or attitudes. The goal is to interview each group of earlier respondents a total of three times, each time two years apart.

Many of the survey questions remain basically the same, although some recently emerging issues—gay marriage, for example—have prompted the addition of new questions to reflect the changing times. Also, for the same reason, several of the questions have been updated.

For example, a battery of questions about tolerance which seeks to elicit feelings about "suspicious" groups now asks about "Muslim clerics who preach hatred against America," whereas the earliest such questions asked about communists, socialists and individuals who oppose all religions.

Also, the language in some questions has changed, again in recognition of an evolving social climate. The term "Negroes," for example, used in the 1972 questionnaire, has been replaced with "blacks," or "African Americans."

A 1972 question, for example, asked: "Do you think Negroes should have as good a chance as white people to get any kind of job, or do you think white people should have the first chance at any kind of job?"

By contrast, a 2008 question reads: "Some people say that because of past discrimination, blacks should be given preference in hiring and promotion. Others say that such preference in hiring and promotion of blacks is wrong because it discriminates against whites. What about your opinion—are you for or against preferential hiring and promotion of blacks?"

"The vast majority of questions are unchanged over the years following the principle that 'the way to measure change is not to change the measure," Smith said, adding: "We have retired items that have become dated, skewed or irrelevant. We add new items to cover topics of emerging importance, such as climate change, gay marriage and genes."

The survey field work takes about five months, starting in March. Before the interviewers begin visiting homes, every household in the random sample receives a letter. The envelope feels a little hefty. This is intentional. The goal is that people will feel the weight, and be less inclined to toss it out unopened. Inside, a letter explains that someone will show up within a few weeks to conduct an important national survey. There's a brochure explaining the survey, and a refrigerator magnet in the shape of the United States to help folks remember.  

The hope, of course, is that everyone will welcome the visitors, and cooperate. Most of the time, that's exactly what happens. But not always. Occasionally, for a variety of reasons, people refuse to open their doors, prompting interviewers to become very creative in order to get the information they need.

"We've had cases where people do the interviews from behind a screen door, while the interviewer sits on the front step," Smith said. "Once, an interview was conducted on a tractor, while a farmer plowed his field. Interviewers sometimes need to hitch rides on snowmobiles to reach isolated mountain cabins, but that's unusual. We've had cases where we have knocked on doors ten times and never gotten a response—and we know someone is there."

The interview takes about an hour or more to complete, which, for many people, is a significant amount of time. For that reason, some resist.

"Some people are willing, but sometimes you have to win them over," Smith said. "For example, if a person said, "I'm not interested in politics, I'm too old.' The interviewer might say: 'That's why you should do this. We need people just like you who aren't interested in politics.' Or if someone said, 'I have health problems, I don't have time,' the interviewer can say: 'But that's exactly why you should do this, so that the views of people with health issues can be represented.' You can take whatever excuse they give you and turn it around."

About 200 interviewers fanned out to conduct the 2010 survey. Most of the interviewers are college graduates; a large percentage are women. They vary in age, and race/ethnicity. They are paid by the hour, and also receive travel expenses.

Sometimes they incur additional costs to purchase certain incentives used to encourage the reluctant to cooperate. These must be approved by the central survey office.

For example, interviewers who are not successful on their first try may return the next time carrying a bouquet of flowers. This approach has proved especially effective with elderly women, Smith said. "It's a strong signal that the interviewer is thoughtful and friendly—and that you can be friendly back by doing the interview," Smith said.

Then there are the distracted mothers with young children. They want to cooperate, but can't. The solution: by prearrangement, the interviewer arrives with a pizza and an age-appropriate DVD. "That keeps the kids busy and happy, while the mom does the interview," Smith said.     

Interviewers question only one adult in each household. For households with two adults, half of the interviewers are randomly pre-assigned to poll the older person, while the other half is randomly pre-assigned to interview the younger one. This provides a balance.

Imagine the plight of the flustered interviewer who encountered elderly identical twin sisters—and forgot to ask which one was born first. She had no idea which sister she needed to interview.

Taking pity on her, one of the sisters confided: "Dear, it doesn't make any difference. We both think exactly alike."

Editor's Note:This research was supported by the National Science Foundation (NSF), the federal agency charged with funding basic research and education across all fields of science and engineering. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation. See the Behind the Scenes Archive.