A DAMN CLOSE RUN THING

INSIDER INTERVIEW: Kathleen Hall Jamieson: This week, the University of Pennsylvania's Annenberg Public Policy Center released the analysis of its 2000 election survey, which involved more than 100,000 interviews conducted over 14 months. NationalJournal.com's Anne Wagner talked with Annenberg director Kathleen Hall Jamieson about what she learned from the 2000 polls, and what researchers have planned for 2004. (National Journal Group Inc., Aug. 6, 2003)

Q. Can you give a few examples of some of the interesting findings from the 2000 survey?

A. Because we are able to track daily, we are able to answer such questions as, "Did the Gore drop in vote intentions occur because of the first debate, or had it begun before the first debate?" The first advantage of this survey is that it lets you attribute findings to specific events. Rather than simply knowing that something occurred, we know approximately when it occurred and hence can draft some reasonable inferences about why it occurred.

We know, for example, that Gore started to drop before the first debate as a result of what appears to be a combination of factors. The petroleum reserve statement appeared to affect him adversely. The criticism of his going to Hollywood to raise money after his ticket had criticized Hollywood appeared to have an effect. And the traction of the news stories about the mother-in-law and the pharmaceutical price and the dog appeared to be gaining traction before the debate happened.

We also know -- and this is because we are in the field every day and we know who watched debates and who didn't -- that the effect of the first debate was felt on non-watchers as it affected Gore, not on watchers. Debate watchers continued to think that Gore had done a good job in the debate, and that persisted in the following week. The drop that Gore experienced as a result of the first debate was an effect on the non-watchers who were influenced by the media commentary about the debate, not among the watchers who would have been affected directly by the debate.

As a general point, the first set of interesting findings in the survey answered not the question, "Did Gore drop during that period?" but "When?" and "Why?" and "What was the relative impact of news and Republican advertising and the relative effect of debate watching as opposed to commentary about the debate?" [...]

Q. Were there any other interesting findings like that?

A. There are three sets of findings in the data set that are interesting in light of the communication and political science literature. There is a first clear effect produced by the conventions, largely produced by the second convention and largely produced in the form of Democrats recognizing that the economy was doing well and coming to accept that and giving Clinton-Gore some of the credit. So there was a priming effect as a result of the Democratic convention on a traditional, structural variable -- something that political scientists have believed makes an effect routinely turns out to make an effect only when you communicate about it. It was the communication at the convention about the economy that produced the effect. After that, because Gore didn't stress the economy, the effect dissipates. But there is one effect on people's perception of the economy. It's an effect on Democrats as a result of the Democratic convention. And it helps lead the Gore rebound out of the convention.

The second effect is... the perception that Gore's honesty collapses as a result of that set of variables we just discussed.

Then there's a third big effect in the data set, and it's an "issue effect." Coming out of the last debate, Gore begins to get traction on his claims about Social Security. That's a combination of an effect produced among debate watchers and among heavy broadcast news consumers. It appears to be a byproduct of the fact that [George W.] Bush did not appear on network news as much in the final week of the campaign as Gore did. Gore appeared more, in other words. And Gore hammered Bush on Social Security in those press interviews on national network evening news. The rebound effect occurs more strongly on Social Security -- that's rebound of Democrats coming back toward the traditional Democratic position on Social Security and independents gravitating toward the Gore position -- out of heavy news watchers and debate viewers in the final weeks.

Gives you some sense of how insanely bold it was for George W. Bush to make Social Security reform a centerpiece of his campaign and why he's not forcing the issue until he has a filibuster-proof majority in the Senate. Posted by Orrin Judd at August 7, 2003 8:16 AM