In my cover story this week, "The Big Race," I mention that white women voters who have been backing Democrats should prove sympathetic to Barack Obama’s candidacy. I had a specific study of women voters in mind, but I didn’t want to exhaust readers’ patience with the description of another psychological experiment. I will describe the study here, because I think it has important implications for the 2008 election.
Many political scientists and historians have explained the gender gap between men and women voters as having really taken hold in 1980, when women went for Reagan by only 47 to 45 percent, while men favored him by a score of 55 to 36 percent. They attribute the disparity in that election to the Republicans opposition to equal rights for women and their embrace of the religious right. Saying the Democratic-leaning gender gap started in 1980, though, ignores quite a bit of history. Through 1960, a higher percentage of women than of men generally favored Republicans, but in 1964, a higher percentage of women than men backed Lyndon Johnson against Barry Goldwater. The gender gap reappeared in 1968, when more men than women backed George Wallace, and again in the 1972 election between Richard Nixon and George McGovern (women went with McGovern). In all of these elections, party differences on racial issues overshadowed any differences on gender issues.
After the 2000 election, four political scientists, Vincent L. Hutchings and Nicholas A. Valentino of the University of Michigan and Tasha S. Philpot and Ismail K. White of the University of Texas, devised a set of experiments to test whether the gender gap was indeed being driven by racial rather than gender issues. They reasoned that what mattered to women voters was how compassionate a candidate appeared, and that a candidates’ attitude toward African Americans was an important clue on that score.
The political scientists gathered together 145 white voters in Ann Arbor in July 2000. A control group read a set of non-political newspaper articles. A second group read articles, headlined “Gore, Bush Differ on Black Issues,” that, using facts from the campaign, set out dramatic differences between the candidates on such issues as civil rights, affirmative action, public education, and health care. A third group read stories titled, “Gore, Bush Similar on Black Issues,” that underlined the commonalities between the candidates. A fourth group read stories about Gore and Bush’s differences on women’s issues. The political scientists then compared the subjects’ preference for Gore and Bush after reading the racial and gender stories with the preferences of the control group.
What they found bore out their hypothesis. While the stories showing the candidates’ difference on racial issues increased women’s support for Gore and men’s support for Bush, those showing the candidates’ similarity on racial issues eliminated the gender gap. In addition, the stories about the differences on women’s issues failed to produce any significant change from the preferences of the control group. The political scientists got very similar results from a different group of subjects in Detroit when they had them view hypothetical ads from the campaign. The ads that most widened the gender gap were those that appeared to blame blacks for “wasteful government programs.”
The political scientists then turned their attention to George W. Bush’s campaign strategy in 2000. At the Republican convention, Bush featured prominent blacks including Colin Powell and Condoleezza Rice. By comparing the results of Gallup polls taken at the time, the political scientists established that this strategy won Bush support among women. At the same time, it didn’t lose him support among men. The reason, they speculated, was that Republicans at the convention also emphasized tax cuts, defense spending, and other issues important to white male voters. It was a win-win strategy for Bush that carried over to the election itself, where Bush narrowed the gender gap from 1996.
The political scientists’ experiments are not conclusive. By including affirmative action in the list of differences between Bush and Gore, they may have introduced an element of self-interest into women’s choices between the candidates. The experiments also can’t settle the question of why women disproportionately backed Jimmy Carter in 1980. Too much of the historical context is lost between November 1980 and a small group of Michigan voters in the summer of 2000. Nonetheless, the experiments are suggestive of how white women and men think about racial issues, and can provide clues to this year’s election.
The experiments suggest that Obama should be in a good position to win support from the white women who voted for Hillary Clinton in the primary. These voters will be amenable to supporting a black candidate and least susceptible if the Republicans attempt to play the race card. If McCain wants to cut into Obama’s advantage among these voters, and among white women in general, he will have to follow Bush’s example in 2000.
McCain, of course, has already begun doing that. Last month, the Arizona senator visited African American victims of Katrina in New Orleans and gave a speech in Selma, Alabama, commemorating the civil rights movement. Some pundits wondered whether McCain knew what he was doing--did he really expect to win black voters against Obama?--but they misunderstood McCain’s true motivations, and, in the process, wildly underestimated his campaign. McCain was not after the votes of blacks. He was trying to display his own version of compassionate conservative in order to win the support of white women voters.John B. Judis is a senior editor at The New Republic and a visiting scholar at the Carnegie Endowment for International Peace.