James Davis


National Opinion Research Center

University of Chicago

“Did growing up in the 1960s leave a permanent effect on attitudes and values?” James Davis asked last year in Public Opinion Quarterly, drawing on a body of research that dates back to the 70s. The answer, he found, is yes–but not to the degree that aging boomers might think.

Harold Henderson: You used data from the General Social Survey. What’s that?

James Davis: It’s a long-term project started in 1972. We survey a random sample of the general population every other year (it used to be every year), about 3,000 cases. Trained interviewers ask the questions face-to-face. It’s unique in a couple of ways. One, the questions are general, not topical. Two, about half the questions are repeated every time so we can get a time series.

HH: This isn’t like an election poll, where there’s a reality check. How do you know you’re measuring something real? Or that people aren’t lying?

JD: That’s the most important question this kind of social survey faces. Not many of the questions we ask can be checked objectively, either because it would cost too much or because there is no way to check the answer to “How happy are you?” But there are a lot of little things. We check for internal consistency, and we take extraordinary care in collecting and processing the data. The average respondent is obtained after 11 attempts–not because people are rude, but because so many are working or traveling.

HH: So once someone has been randomly selected to be in your sample, you try really hard to talk to them to preserve the randomness.

JD: And we reinterview a fraction to check on our own interviewers. It is possible that in a given year we might by chance get too many factory workers or Chicagoans or Catholics in our sample of 3,000, just as you might sometimes flip a coin and get a few more heads than tails. But our sample is so large the chances of a serious sampling error are small. Another check is fairly consistent results from one survey to another. That adds credibility.

HH: Doesn’t the wording influence people’s answers to some questions?

JD: That’s a real problem. Some claim this shows surveys are no good. But it also shows that people are listening–otherwise small changes would make no difference. Unfortunately we don’t always know what people think the questions mean. It’s like that question asked after the election, about whether you voted on moral grounds. Professionals should never have written that question.

HH: I gather that one way you deal with that is not to rely too much on any single question.

JD: Right. We usually group them in clusters, which helps compensate for this problem.

HH: So you find some noticeable long-term liberalizing trends in American public opinion?

JD: There are four or five clusters of items that have been studied over and over. One is race relations. In the middle 1940s sociologists asked, “If good jobs are scarce, should white people get first chance at them?” About 50 percent said yes. By the late 1960s it was down to 3 percent. If you asked that question today, you’d get thrown out! So there’s a trend to color blindness there. Another cluster has to do with whether women should work outside the home. Back in the 1930s Gallup was asking this question, and there has been a tremendous liberal shift toward gender equality in these answers. Another cluster goes back to 1954, when there was a study on civil liberties during the McCarthy period with questions like “Should a communist be allowed to speak or to have books in the library?” Here too there’s been a definite shift in the direction of–well, maybe the word is more permissiveness.

HH: So there’s a long-term trend, whatever we call it, on these five clusters–sex, authority, race, free speech, family.

JD: Yes, all these things change, but slowly. And the change seems to be cohort-driven; in other words it happens as one generation gradually replaces another. That’s what led to my article. We know there are steady changes. Are there any lumps and bumps, and if so, are they because of when you reached adolescence?

HH: People who were around in the 1960s tend to think it was different.

JD: Well, we’ve got to be careful of that. I attended a dinner party recently with maybe a dozen people, and nobody there knew anybody who’d voted for George W. Bush. If your own experience was always trustworthy, we sociologists wouldn’t be in business.

HH: This long-term trend would imply that the generation coming of age during the 1960s would be somewhat more liberal or permissive than the generation before it. But you also found that they added an extra bump over and above that?

JD: Yes. And you can take that either way. You can say there was an impact, yes. Or you can say, given all the hullabaloo about the 1960s there’s only 3.9 percent extra?

HH: You also found, as part of this long-term trend, that the generation coming of age in the 1950s was a bit more liberal than the one before it.

JD: Yes. And for the 1970s generation and later, we don’t find a general reversal. On some items–adultery, for instance–there is a less permissive attitude. But on most clusters there’s what my colleagues and I call a liberal plateau. It’s important to study change, but there hasn’t been as much as most people think.

Art accompanying story in printed newspaper (not available in this archive): photo/Lloyd DeGrane.