Gender has been in the news a lot this week, since women seemed to be vital to President Obama’s reelection. Maybe they were, but the white ones? Not so much, John Cassidy noticed in a New Yorker blog post the Friday after the election. Challenging the much-hyped gender gap, Cassidy wondered “What’s up with White Women”? They've headed the “wrong way,” he claimed: in the Republican direction, causing what Cassidy called a “reverse gender gap" with white women preferring Romney to Obama by 14 points in 2012 (56% to 42%). What's more, wrong-way white women chose McCain (53%) over Obama (46%) last time around, while 55% of white women picked George W. Bush over John Kerry in 2004. So what's up?
A little bit of history would've taken the punch out of Cassidy's line of reasoning. What's up with white women? Absolutely nothing. White women have not reversed or changed course. In 2012 and the two elections before that, white women voted about the same way they've been voting since the 1970s: almost always for the Republican Presidential candidate.
Surprise that white women are doing what they’ve always done stems from distorting media spin on findings by polling organizations, such as Gallup, which suggests that women’s concern for women’s issues drives them towards the Democratic Party. Yet as political scientists have shown, the gender gap – that is the difference between women’s aggregate vote preferences compared to men’s – is explained best by differences in how the sexes see the size and role of government. Further, political scientists have shown that the gap opened up around 1980, when white men headed for the GOP, not when white women abandoned it for the Democratic Party. Way back in 1991, the venerable political scientist Warren Miller (gated article) cited work published even earlier, by his younger colleague Dan Wirls (gated article) in 1986, to wit: “the appearance of the gender gap in the Reagan years was not as much a function of a liberal, pro-Democratic growth in the partisan sentiment of women as a function of the sharply conservative pro-Republican move among men.”
What Wirls said in 1986 and Miller broadcast in political science’s flagship journal, the American Political Science Review, in 1991 should have become conventional wisdom, but it didn’t. So we repeat the (still not yet conventional) wisdom here. Using data from recent exit polls, NES data and other data, our colleague Adam Hughes created the graph pictured with this post to demonstrate the history of the gender gap. Like all polling data, these data are estimates of real votes (that of course are not recorded according to gender or any other personal attribute) to give us a sketch of the vote choices individuals make on Election Day.
The graph shows how, for a long time, white women have reliably voted more Republican than Democratic, though less dependably so than their white male counterparts. The graph also shows that the only year in which white women are more Democratic than Republican in their Presidential vote choice is 1996, which also happens to be the only year in which there is a large, unequivocal gender gap in the sense that more white women voted Democratic than Republican and more white men voted Republican than Democratic.
Another historical fact that does not get as much attention as it should is that over this time period women comprised an increasing proportion of the electorate. Women comprised 49 percent of the electorate in 1972 and gradually increased to 53 percent by 2008. Any stylization of the typical or modal "American voter" that implicitly suggests he is masculine (Joe Six-Pack, anyone?) after 1976 is misleading.
Alas, this new feminine American voter isn't necessarily Democratic, and pretty decidedly she is not if she is white. According to these data at least, if only white women voted, then the Republican Presidential candidate would have won in every year except 1996. Finally, the graph shows what political scientists know, that white male Republican voters are the “driver” of the gap.
In recent years political scientists have done a good job explaining why men and women differ in their partisan and vote preferences. In Unconventional Wisdom: Facts and Myths about American Voters, Karen M. Kaufman, John R. Petrocik and Daron Shaw argue that the divergence between men and women in the political views and behavior “that began more than forty years ago has grown into a significant and enduring political division.” Kaufman et al. demonstrate that this gender gap averaged about five points during the 1960s and 1970s, disappeared around the time of the Watergate scandal, increased slightly during the 1980s, and then increased to double digits around the time of Clinton’s reelection in 1996. In the last several elections, it has averaged about seven points. Perhaps most importantly for our purposes, Kaufman et al. show that the gender gap was caused by men becoming more Republican, rather than women becoming more Democratic.
Furthermore, new research on “The Macro Politics of the Gender Gap” (gated article) in public opinion toward the role of government by Paul M. Kellstedt, David A. M. Peterson and Mark Ramirez provides additional strong evidence that men are the major driver of change. Analyzing data between 1980-2005, Kellstedt et al. find that while both men and women respond to changes in public policy by shifting their policy preferences, men appear more responsive to policy changes than do women, causing a gender gap in policy preferences across time. When policy moves in a liberal direction, men move in a conservative direction at a faster rate than women. In contrast, when policy moves in a conservative direction, both men and women move to the left, but because men do so at a faster rate, the gender gap decreases and male preferences move closer to the preferences of women.
So, if white females are consistently Republican, how did President Obama carry the female vote by 11 points overall? First, white females make up a smaller proportion of the overall electorate, about 38 percent in 2012. Second, non-white women are growing in numbers – about one in six voters is a now a non-white woman. Obama won the majority among these voters, including 96 percent of black women, 76 percent of Hispanic women and 66 percent of women of other races, including Asians. Obama also did well among single women of all races, garnering 67 percent.
The bottom line: facts have been slow to sink into analysis about the precise nature of the gender gap in electoral analysis. To be fair, this year Gallup did note that the increase in the increase in the gap's size between 2008 and 2012 is attributable to men:
Notably, Obama's 12-percentage-point advantage among women is slightly less than the 14-point advantage he had over John McCain in 2008, while Romney improved on McCain's performance among men by eight points. Thus, the narrowing of Obama's winning margin between the two elections, from seven points to two points, can be ascribed mainly to men's shifting more Republican. (emphasis ours)
But as we’ve shown, the fact that men, rather than women, are the drivers of electoral change has a much longer and deeper history. Further, white women have consistently voted Republican, at least since the 1970s. So instead of asking “what’s up with white women,” maybe we should be asking “why the media spin?”. Journalists and pundits alike should pay closer attention to the range of economic, social, cultural and racial factors that contribute to the composition of voters in presidential coalitions.
Lynn Sanders is an Associate Professor in the Woodrow Wilson Department of Politics at the University of Virginia and the Director of Undergraduate Studies. Carah Ong is a PhD candidate in the Woodrow Wilson Department of Politics at UVa and a Research Assistant in Democracy and Governance Studies at the Miller Center. Adam Hughes is a PhD candidate in the Woodrow Wilson Department of Politics at UVa.