Victory Lab

The 12 Kinds of Undecided Voters

Liars, haters, mavens, know-nothings, bandwagon riders, and other kinds of voters who just can’t make up their minds.

Iowa voters.
A roomful of Iowa voters in Ames come to hear GOP presidential hopeful Mitt Romney speak on Dec. 29.

Photograph by Jewel Samad/AFP/Getty Images.

Just over one-half of the Iowa Republicans surveyed by Time/CNN this week said they would “definitely support” the candidate they backed at the time of the poll. Among the remainder, opinions are still fluid—by any measure, a staggeringly high share of likely voters. A much higher percentage of voters appear to make late decisions in the caucuses than in presidential general elections: While 10 percent of voters nationwide in the fall of 2008 told exit pollsters they made up their minds in the race’s final week, 40 percent of Republican caucus-goers in Iowa that year did. Seventeen percent said they made up their mind on the day of the caucuses. 

This presents a serious challenge for the Republicans vying for votes in this year’s Iowa caucus. To make matters worse, many voters who tell pollsters they’re undecided are actually anything but—they’ve made up their mind, but for one reason or another, don’t care to share their feelings with pollsters. What’s more, studies have shown that many undecided voters don’t ever show up to vote in elections at all, making efforts to win them over doubly doomed. But the candidates do have some strategies for dealing with this squirrely segment of the electorate. Campaigns know, for instance, that undecided voters are not all the same. Here are a dozen different types of self-described undecided voters, and how the Republican presidential candidates are dealing with them in the final days before the Iowa caucus. N.B.: Some voters likely fit into more than one category.


WHO THEY ARE: A lot of Iowans want to vote strategically, but don’t know how. Since 1972, when changes in the nominating calendar placed the caucuses first, Iowans concerned with supporting a candidate who can capture the nomination and then be a strong general-election contender have had less evidence to go on than voters in all the other states. That is likely one significant reason why so many Iowa caucus-goers remain undecided for so long. The problem is particularly acute this year, when the polls have been fluid and have never pointed to a consistent national frontrunner.

In early December, the George Washington University political scientist John Sides and Republican consultant Alex Lundry (of the firm TargetPoint Consulting) collaborated on an online experiment to test whether new information on electability moved voters. They polled Republican primary voters on their preferences, and then showed them the latest predictions from Intrade, which indicated Romney was the most electable candidate, followed by Newt Gingrich. The Intrade data proved persuasive: Forty percent of voters changed their pick after seeing it, mostly by going to Romney or Gingrich from less viable opponents—or by moving from “no preference” to picking a candidate.

WHAT CAMPAIGNS ARE DOING WITH THEM: One pro-Romney super-PAC is already up with TV ads showing a luggage carousel and a voiceover that asks, “Know what makes Barack Obama happy? Newt Gingrich’s baggage.”

MEET ONE: “I’m still shopping. I’ve probably got it narrowed down to two right now: Romney and Paul. Electability is probably the big thing … I’m looking for someone who will challenge the status quo, but I think Romney is a safe choice—he’s a good campaigner, he’s made the fewest mistakes, seems like he’s not going to have any surprises down the road.” —Darren Kleis of Johnston, real estate portfolio manager


WHO THEY ARE: Much has been written about the role of wedge issues in general-election races, where an emotional issue can be used to divide a party coalition: Think of the Republican who defected to John Kerry over stem cells, or the Democrat who supported the Iraq surge and was pulled to McCain because of national-security concerns. But even in a primary, where the field largely agrees on many issues, a single subject can give a voter pause about backing a candidate he or she is otherwise drawn to. For instance, a voter attracted to Gingrich or Perry might be hung up by their relatively liberal attitudes toward legalizing immigrants.

WHAT CAMPAIGNS ARE DOING WITH THEM: One reason so many candidates, from Perry to Bachmann, have been intent on reminding voters of Ron Paul’s dovish posture toward Iran is to make undecided voters anxious about moving toward the Texas Congressman.

MEET ONE: “I like Ron Paul, and his ideas about finance—and he’s been talking about them for years, and he’s been right for years. But I’m still looking … If he gets elected, he’ll pull all our forces out and the military would be cut. It’s good to pull some of the reins back, but that seems a little extreme! It really is something I pray about. I just don’t think you can strip down to the levels that he would.” —Lisa D. of Des Moines, insurance salesperson


WHO THEY ARE: Just after the 2004 election, pollster Celinda Lake and mail consultant Hal Malchow convened a focus group with 13 voters who had cast ballots in a St. Louis-area state senate race but had been undecided until the last minute. Lake and Malchow wanted to know how each voter had made up his or her mind. In 12 of the cases, it was a single recommendation from a trusted friend or relative that pushed the voter toward a decision. (The 13th seemed to be an outlier: a voter whose door had been knocked by one candidate on the Thursday before the election, and was so impressed he invited the candidate to address his church group on Sunday night.) Their conclusion: late-deciding voters, often short on information and knowledge, can move because of word-of-mouth.

WHAT CAMPAIGNS ARE DOING WITH THEM: Campaigns with enthusiastic, committed supporters—Ron Paul seems to win the category—may expect their loyalists to spread the word even if their ads, mail, and phone can’t break through the din.


WHO THEY ARE: Too many friends, however, can be overwhelming. Dianna C. Mutz, now at the University of Pennsylvania, has found that voters whose social networks force them to encounter dissonant views—like, say, a businessman whose chamber of commerce buddies push him to Romney while his fellow churchgoers are talking up Perry—remained undecided later in the campaign. Such indecision can be paralyzing: Those facing what Mutz describes as “cross-cutting social networks” were less likely to actually vote at all.

WHAT CAMPAIGNS ARE DOING WITH THEM: The best indication of someone’s likelihood to vote remains their past voting history, so toward the end of a campaign field organizers start ignoring infrequent voters who are undecided—whether because of cross-cutting social networks or general laziness—to focus on those with a better record of turning out.

MEET ONE: “I’m still considering—I like a lot of them. I really haven’t made up my mind. I thought I had done that three times: I was with Newt before I realized who Newt was. I like Mitt Romney. Some days I even like Ron Paul … I have a lot of friends and family who have made up their minds, and they keep inviting me to come to the Westside Conservative Club.” —Jackie Fleming of Urbandale, retired AAA employee


WHO THEY ARE: In late 2007, Hillary Clinton’s data team noticed a peculiar trend coming out of Iowa: The numbers coming in from volunteer phone banks consistently overstated Clinton’s support when compared to the numbers coming in from the paid call centers the campaign also used to identify voters. One of Clinton’s analysts concluded that part of the problem might be exuberant volunteers overestimating a voter’s potential support—so you’re saying there’s a chance? But the bigger takeaway was that voters don’t always want to be honest with someone on the other end of the phone about their preference. The easiest way to let a canvasser down easy: claim you’re undecided when you’re not. And while the Democratic caucuses require attendees to declare their preference in public, Republicans vote by secret ballot—so it’s easy for a voter to keep a choice private throughout the process.

WHAT CAMPAIGNS ARE DOING WITH THEM: Some campaigns automatically adjust their predictive models to discount volunteer contact compared to paid calls. (In Iowa and elsewhere, Republicans have traditionally been far more reliant on paid calls than volunteers.) But if a voter is hiding behind “undecided,” there’s not much you can do to find that out—or change their mind.

MEET ONE: “They’ve all been calling me. One asked if I knew who I was supporting. I said no, and they ended the call … I’ve finally made up my mind, but I’m not willing to say.” —Dave Roederer of Johnston, director of Iowa Department of Management and John McCain’s 2008 Iowa chairman


WHO THEY ARE: A large chunk of people decide late because they just don’t care about politics and tune it out for as long as they can. Tulane’s Brian Brox and the University of Arkansas-Little Rock’s Joseph Giammo examined the attributes of late deciders in presidential general elections and found that a large chunk of them resembled the “stereotypical apathetic citizen.” They tended to be less partisan and politically active than early deciders, treated the candidates as interchangeable, and weren’t inclined to care who won. One surprising twist: Brox and Giammo found that these “low-interest late deciders” seemed to recoil at information sent their way. The more Republican ads they saw, the less likely they were to vote Republican; Democratic ads pushed them toward the GOP. “These voters may actually be likely to become irritated with the efforts of candidates to attract their votes,” Brox and Giammo write, “since we know that they have little interest in the campaign in the first place.”

WHAT CAMPAIGNS ARE DOING WITH THEM: That “low-interest late decider” doesn’t look much like the typical caucus-goer—among Republicans, at least, caucus-goers tend to be better educated and more partisan than the broader electorate—but a big turnout on Tuesday could pull in some atypical Iowans. The last-minute advertising barrage could backfire most with them.


WHO THEY ARE: Brox and Giammo found another subcategory, “high-interest late deciders,” who look entirely different: They are more partisan and make sharper distinctions among the candidates than even early deciders. “These high-interest late deciders more closely approximate the idealized view of voters that has taken such a beating in political science over the past 50 years, taking the time to consider all of the evidence before coming to a conclusion,” the authors write. These voters are more likely to pay attention to media coverage, but campaign ads and candidate visits to their states don’t move them.

WHAT CAMPAIGNS ARE DOING WITH THEM: Loyal, active Republicans are receiving the most regular contact from campaigns now in Iowa, and candidates will try to make a mark on them until caucus night.

MEET ONE: “I’m still not there yet. Isn’t it awful? I hate that … I don’t take this lightly at all. This is a serious time our country is in, so we have to make sure we have the right man in there. I have plenty of time.” —Nevie Roe of Waukee, housewife


WHO THEY ARE: Polls tend to treat being decided as a binary condition—like being single or being married—but what if people are dating the candidate they’ll end marrying but just don’t know it yet? University of Pennsylvania doctoral candidates Lauren Kogen and Jeffrey A. Gottfried invoke the dating metaphor in a new paper in Political Behavior that considers the National Annenberg Election Survey taken during the 2008 general election. The survey asked respondents throughout the campaign for their candidate preference, and then asked those who voted after the election when they made their final decision. By looking at the overlap between the two methods, Kogen and Gottfried found that nearly one-third of voters claimed to have been undecided later in the campaign than they actually were. They called this category “uncommitted early deciders,” and found they were more partisan and more knowledgeable about the campaign than those who really did decide at the end.

WHAT CAMPAIGNS ARE DOING WITH THEM: Romney has an advantage here. His data-intensive campaign has the names of voters who were identified as his supporters last time and can match it to a list of caucus-goers to see who actually turned out before. Internal research shows their profile has remained stable over the last four years, so the campaign has been able to infer the support of these voters even if they call themselves “undecided.”

MEET ONE: “I’m undecided. All of them are good, they’re all wonderful, but I’m interested in somebody who can run a business. It has to be a businessman … I haven’t decided for sure, but it will probably be Romney, and I voted for him last time.” —Ladonna Gratias of Clive, home-builder


WHO THEY ARE: Campaign strategists have long described late-deciding voters as “low-information” voters and microtargeters often spot higher rates of undecideds in areas the Census identifies as having low rates of high school or college-educated voters. Through a randomized experiment, the lefty Alliance for a Better Minnesota found a shortcut to identifying which voters were showing up as undecided because they lacked knowledge and not because they were conflicted or cross-pressured. The group asked “Can you name your member of Congress?” and found that those who didn’t know the correct answer were perfect targets: they could be moved late in a campaign with a piece of information-rich mail. To better profile these undecideds, the Voter Participation Center, formerly Women’s Voices Women’s Vote, recently added a battery of six questions—name the vice president, or the number of representatives in Congress—to measure political knowledge.

WHAT CAMPAIGNS ARE DOING WITH THEM: Some of Ron Paul’s handbills are text-heavy and dense with information about issues like right-to-work laws—they might not be addressing hot-button concerns but could give under-informed voters the confidence to think they have the information to make a decision.


WHO THEY ARE: Some voters will claim to be undecided even as the caucuses begin on Tuesday night. In a poll conducted for their book Why Iowa?, political scientists David Redlawsk, Caroline J. Tolbert, and Todd Donovan conducted a novel poll at caucus locations in 2008, randomly interviewing one Democrat and one Republican at each of the state’s 1,774 precincts. Even as they sat in the room minutes away from voting, just under 5 percent of attendees surveyed said, “I have not yet made my decision.” Their decision is likely an impulsive one.

WHAT CAMPAIGNS ARE DOING WITH THEM: Candidates are allowed to have a supporter make their case before voting begins, and having close to 1,784 of these so-called precinct captains is typically held up as the sign of a well-prepared campaign. The speaker doesn’t have to be an Iowan, and it’s likely that some of the surrogates crossing the state this week for Romney—like his sons and wife—will be dispatched to make a last-minute case for Mitt in precincts where the campaign has identified large concentrations of independents.

MEET ONE: “My decision is going to be at the last second at the caucus. Last time I wavered between three people at the caucus. You’ve got to have the sense that ‘this is the guy who’s ready to run America and watch out.’ Our leader needs to look good.” —Bruce Kalisek of Waukee, realtor


WHO THEY ARE: Are voters claiming to be undecided because they’re afraid to report what’s really driving their choice? In the 2008 general election, final polls closely predicted Obama’s actual popular-vote total, while John McCain performed nearly two points better, suggesting that late-deciding voters broke toward the Republican. Rutgers’s David Redlawsk found that among the 6 percent of voters who described themselves as undecided in the final days of the campaign, two-thirds indirectly acknowledged worry over the fact that Obama would be the first black president, according to a survey list experiment designed to discern hidden bias. Are Iowans now wary of acknowledging their bias against a Mormon or female candidate hiding their actual preference by claiming to be undecided?

WHAT CAMPAIGNS ARE DOING WITH THEM: It’s less likely that such hidden bias exists in a multi-candidate field than in a two-way general election, but the Romney and Bachmann campaigns can use survey list experiments or other forms of indirect questioning to profile out those likely to be problematic—and make sure they’re not turning out the wrong undecideds.


WHO THEY ARE: In a multi-candidate race, it can often be easier to decide whom to oppose than it is to settle on a favorite—and it turns out, despite their supposed niceness, there’s an august tradition of negativity propelling Iowans to vote. In a phone survey immediately after the 2008 caucuses, the Why Iowa? authors asked those who had turned out why they had done so. While 95 percent said “it was the right thing to do,” 27 percent identified a less uplifting motive for attending: to “oppose some other candidate.” In a year of so-called anti-Mitts, a bloc of voters may be intent on caucusing to block one candidate—but haven’t decided whom to back when they do. 

WHAT CAMPAIGNS ARE DOING WITH THEM: Media strategists often talk about replacing their negative ads in the final days with positive message, so expect this weekend to be filled with gauzy feel-good spots; Romney’s closing stump speeches are full of patriotic platitudes. In a multi-candidate field, where haters have many places to end up, this is even more important.


“I haven’t fully made up my mind, but if we’re down to two or three candidates there’s one I can’t vote for … Paul appeals to a segment of the Republican party, and some Democrats, who want no laws, no rules, no regulations.  I don’t like the fact that he wants to legalize drugs and gay marriage.  I have a real problem with libertarian policies.” —Bill Nichols of Ames, Christian book publisher