There's something about the shrubbery of paper in Jon Krosnick's office that looks neatly gardened. He certainly knows where it's all sprouting from. "At my last annual review," says the professor of communication and political science, "I had to list how many projects I'm working on. And I listed 52."
Some are more conspicuous than others, such as his collaboration with the Associated Press, which combines academic scholarship with topical news. The result has been surveys addressing environmental questions, health care and politics, including an analysis of antiblack racism in the 2008 presidential election.
Krosnick is no stranger to time-consuming projects. He has 10 five-drawer file cabinets for a long-percolating book on questionnaire design. But he also rebels against any culture "where, oh, you know, you just take however long you have to take in order to pump your research out." He has seized on the joint efforts with AP to make himself and his students "part of the public debate today."
The link with AP has solid methodological roots: When Krosnick was a professor at Ohio State, one of his graduate students was Trevor Tompson, now global director of polling for the wire service. Tompson says the association with his former mentor reflects AP's desire to "go beyond horse-race polling" and take on "more in-depth social questions."
The 2008 election study made the biggest headlines. The research indicated that racist attitudes cost Obama about 6 percentage points in votes, including those of voters who swung to McCain or who abstained.
This kind of electoral influence, Tompson notes, was assumed by many to be "an unknowable question." Krosnick's team went about the task by enlisting the participation of University of North Carolina professor B. Keith Payne, who contributed expertise in measuring subliminal racism. Two of Krosnick's former graduate students, Josh Pasek, PhD '11, and Alex Tahk, PhD '10, created a new tool for tackling the challenge.
"In past studies of elections," Krosnick explains, "scholars have always treated two different decisions as if they're completely separate: one, whether to vote or not, and two, which candidate to vote for. Josh pointed out that these two decisions are likely to be closely related. Specifically, racism could convert someone who would otherwise not have voted into a person who voted for McCain in 2008. Or racism could convert someone who would otherwise vote for Obama into a nonvoter. And racism could cause people to shift between other pairs of categories as well. So he and Alex developed a sophisticated statistical technique to model all this movement."
Not surprisingly, Krosnick's work can generate controversy. Sometimes it's sparked by the way he formulates questions compared to other pollsters. Among his fundamental concerns is the basis of people's answers.
"So often," he says, "political scientists accuse Americans of knowing little or nothing about the facts of important political issues. And when people lack such information, the opinions they express are of limited value. We explored this issue with regard to President Obama's health care reform bill. We designed a new type of test to gauge each survey respondent's level of knowledge about what exact provisions were in the bill that Congress passed.
"Most Americans knew the right answer to most of our quiz questions, but people varied in how much correct information they had. More importantly, we found among Democrats and independents, the more they knew about the bill, the more they liked it. But among Republicans, people who had lots of correct knowledge and people with little knowledge were identical: Fewer than 5 percent of them favored the bill.
"This indicates," Krosnick concludes, "that if President Obama, and the news media, had been more effective at educating Americans about the components of the bill, a large majority of Americans would have supported it—many more than opinion surveys over the last year have suggested."
