Most American voters have already decided whether they will pull the lever for Barack Obama or Mitt Romney in November. Their decisions were largely predictable even before Romney emerged as the Republican standard-bearer. But there are still a few people out there who are truly undecided — and if the race remains as close as it is now, their votes will be crucial to the outcome. Who are these people, and why do they seem to be having such a hard time making up their minds?
The one fact everyone seems to agree on is that there aren’t many of them. Using its latest polling data, The Wall Street Journal writes that “American voters are growing more polarized and locked in their views.” The Washington Post describes the election as “a settled issue for nearly nine in 10 voters.” The race is “tight and stable,” according to the Post’s Ezra Klein, who adds that “Romney and Obama are realistically fighting over three or four percent of the electorate.” And Paul Begala says “there are about as many people in San Jose as there are swing voters who will decide this election” — 916,643 people in six swing states, to be much too precise.
Typical opinion surveys do not include nearly enough respondents to provide a statistically reliable portrait of this narrow undecided sliver of the electorate. However, the Cooperative Campaign Analysis Project has been surveying 1,000 people each week since January, providing a much larger pool of respondents than any single survey can offer. By putting together the responses from 10 of these surveys conducted from May through July, we have assembled a mega-sample of 10,000 respondents interviewed after it became clear that Mitt Romney would be the Republican nominee.
This mega-sample mirrors other recent polls in indicating a very close presidential race: among respondents who supported or leaned toward either major-party candidate, 51 percent chose Obama and 49 percent chose Romney. But crucially for our purposes, 592 of the 10,000 respondents (5 percent of the weighted sample) said they were not sure which presidential candidate they would vote for, then declined to express even a tentative leaning toward Obama or Romney in response to a follow-up question. These people seem to be truly undecided — and there are enough of them to provide an unusually detailed and reliable picture of undecided voters in the country as a whole.
These 592 undecided voters differ from those who have made up their minds (or are at least leaning one way or the other) in some unsurprising ways. For example, they are rather less knowledgeable about politics, and much more likely to say they follow news and public affairs “only now and then” or “hardly at all.” (Almost 40 percent are unsure which party currently has more members in the House of Representatives, and another 20 percent wrongly answered that it was the Democrats.) They are also considerably less likely to identify themselves as Republican or Democratic partisans, and less likely to call themselves liberals or conservatives (69 percent say they are moderates or not sure).
These differences may seem to confirm the common stereotype of undecided voters as “a group of people that have virtually no partisan or ideological attachments, pay very little attention to politics, and often create the crazy swings we see in the horse race polls,” as Jay Cost of The Weekly Standard put it recently. But that stereotype turns out to be quite wrong in some important ways.
For example, despite the marked overrepresentation of independents among undecided voters, most undecided voters are not independents. The accompanying figure shows the distribution of party identification among our 592 undecided voters, as recorded in a C.C.A.P. baseline survey conducted with the same people in late 2011. Only three in ten were “pure” independents (those who denied leaning toward either party), while another 7 percent said they were not sure about their party identification. Four in ten were Democratic identifiers or leaners, while the remaining 23 percent were Republican identifiers or leaners.
Larry M. BartelsThe fact that more than half of the voters who are currently undecided began the campaign season thinking of themselves as Democrats or Republicans naturally raises the question of why they have — at least so far — resisted supporting their party’s candidate.
Among Democrats, the most notable differences between the decided and the undecided are in their views about the president. For example, while 79 percent of decided and “leaning” Democrats approve of Obama’s job performance and 65 percent say they like him a lot, the corresponding percentages among undecided Democrats are only 17 percent and 8 percent.
Undecided Democrats’ dissatisfaction with Obama does not seem to be focused on any one issue. While 52 percent say they disapprove of his overall performance, the disapproval rates for specific questions about how he has dealt with the budget deficit, the economy, immigration, gay rights, health care, Social Security, Medicare, taxes and abortion all range from 61 percent to 51 percent.
Nor are their personal reservations about Obama focused on any one trait. The proportions who said they would “definitely not” describe him as “effective,” “inspiring,” “honest,” “strong,” “sincere,” or “patriotic” were each around 20 percent. Oddly, a larger fraction of undecided Democrats (30 percent) say they would definitely not describe Obama as “experienced,” despite the fact that he has already spent more than three years in the White House.
While undecided Democrats are clearly less than enthusiastic about Obama, there are indications that he could make some headway among them over the course of the next three months, especially if the economy shows some belated signs of further improvement. Only 18 percent of undecided Democrats strongly disapprove of the president’s performance; most say they “somewhat disapprove” (35 percent) or that they are not sure (30 percent). Presumably the third group, and perhaps the second as well, could be won over. Similarly, while only 3 percent of undecided Democrats currently say the national economy is getting better, 60 percent say it is “about the same” or that they are “not sure.”
If most undecided Democrats are taking a “wait and see” attitude toward the economy between now and Election Day, they seem relatively unlikely to be won over by Romney in the meantime. Three-fifths have an unfavorable opinion of the challenger, with only one-fifth still unsure about him.
If undecided Democrats are focused on Obama and the economy, undecided Republicans seem to have already made up their minds on those scores. Seven in ten disapprove of the president’s performance (45 percent strongly); only one-tenth say the economy is getting better, and even fewer say the country is “generally headed in the right direction.” Clearly, these are people who should be inclined to support the challenger. However, 35 percent say they “dislike” Romney personally, and remarkably — remember, these are Republicans — only 1 percent say they “like him a lot.” Moreover, 64 percent think Romney “says what he thinks people want to hear,” while only 8 percent think he “says what he believes.” (The corresponding percentages among other Republicans are 45 percent and 39 percent.)
Much of this personal antipathy seems to have emerged over the course of the primary season. In the baseline survey in late 2011, only 18 percent of these same now-undecided Republicans said they disliked Romney, while 9 percent said they liked him a lot. Some of this antipathy may stem from the fact that undecided Republicans are also distinctly more moderate than those who have made up their minds. Only 45 percent call themselves “conservative” or “very conservative,” while a similar number call themselves moderate or liberal. (The corresponding percentages among other Republicans are 73 percent and 23 percent.)
Undecided Republicans are twice as likely as other Republicans to say they favor gay marriage (40 percent), twice as likely to express positive or neutral attitudes toward African-Americans (31 percent), and only half as likely to deny the existence of global warming (23 percent). Only 42 percent favor repealing Obamacare (compared with 78 percent of other Republicans). These are the sorts of Republicans most likely to have been alienated by Romney’s dogged appeals to “the base” during the Republican primaries. Whether he can moderate his image enough to win them back without exacerbating the common complaint that he “says what he thinks people want to hear” remains to be seen.
Obviously, it is impossible to tell from our snapshot of the current views of these undecided voters where they will end up — with Obama, with Romney, or at home. Nevertheless, our portrait of who they are and what they are thinking at this point in the campaign may shed some light on the candidates’ prospects in November. In particular, the fact that four of every ten currently undecided voters said before the campaign began that they identified with (or leaned toward) the Democratic Party probably bodes well for the president’s chances. A Democratic skew also appears in undecided voters’ reports of how they voted in 2008 (43 percent for Obama to 33 percent for McCain, with only 23 percent saying they didn’t vote) and even in 2010 (23 percent for Democratic House candidates, 17 percent for Republicans). If they break similarly in 2012 — a big if — the result may be to swing a seemingly “tight and stable” race narrowly in favor of the incumbent.
Larry M. Bartels is a professor of public policy and social science at Vanderbilt University and the author of “Unequal Democracy: The Political Economy of the New Gilded Age.” Lynn Vavreck is an associate professor of political science and communication studies at UCLA, the author of “The Message Matters: The Economy and Presidential Campaigns,” and co-principal investigator of the 2012 Cooperative Campaign Analysis Project.