Princeton Election Consortium

A first draft of electoral history. Since 2004

The neuroscience of being “undecided”

October 7th, 2008, 10:30am by Sam Wang

At, Charles Franklin examines this year’s campaign trends. In his data, “undecided” voters make up a similar fraction to the current Obama-McCain polling margin. How many of them are really undecided? Based on psychology and neuroscience research, maybe not many.

First, here is his graph of decided-voter margins for 2008, 2004, and 2000.

Obama’s decided-voter popular margin is about 8%. Readers of this site won’t be surprised to see the swings, which are shown quite clearly (though sometimes in different proportions) in the 2008 EV estimator history:

(For another comparison, also see the Kerry-Bush race in 2004.)

Compare that with Franklin’s graph of undecided voters:

This second graph shows about 8% of voters saying that they are undecided. This figure suggests that movement in the race is still possible. For example, if all undecided voters decided to vote for McCain, the popular vote would be near-tied. But just how undecided are these voters, really?

In a recent study (news story), 129 residents of an Italian town were asked about their attitude toward a controversial expansion of a nearby U.S. military base. Researchers found that the opinions of 33 “undecideds” could be predicted a week in advance by a series of questions relevant to the issue. This result raises the possibility that decisions exist in an internal form before people can report them.

The work is reminiscent of neuroscience research by Antonio Damasio and colleagues, who found a way to measure the gap between hunch and recognition. People were asked to play a pretend gambling game in which they could choose cards from several decks. Without the participants’ knowledge, some of the decks were stacked against them. After losing repeatedly, they began to choose the more favorable decks but were unable to say why until after much further play.

So even when we claim to be undecided, we may have strong preferences that we cannot report to a questioner. This creates a major problem in interpreting the answers to opinion polls. There may not be many undecided voters out there at all.

Tags: 2008 Election

22 Comments so far ↓

  • Susan Weinschenk

    People will talk about why they are undecided, but the neuroscience research is very clear that they are sticking “rational” reasons on top of unconscious emotional decisions. You just can’t trust what people say when they are describing why they are deciding or not deciding. I wrote a blog on this just this week:

  • Gail Collins Nails the “Undecided” Voter « The Northstar Chronicle

    [...] October 16, 2008 · No Comments Gail Collins had a pretty hilarious article on the debate. I guess, in a way, she could be confirming Sam Wang’s posting on “The Neuroscience of the Undecided Voter“? [...]

  • MadamaAmbi

    I’m really into neuroscience, but right now I’m too busy calling voters in FL to read all of this interesting discussion. Based upon recent calling, I can tell you that undecideds don’t like either candidate & don’t think either candidate can solve this country’s problems. However, assuming the person will engage in conversation with me (many hang up on me, many state that who they will vote for is their private business, many are outright abusive when they hear I’m volunteering for Obama), when asked if they are better off now under a Republican administration, they will say no. If you follow this question up with “So, would it be fair to say that you are leaning Obama” you might get a yes.

  • Ed

    Wait a sec: We have a situation where the margin amongst decided voters changes (in this instance, moving in favor of Obama by 9 points), but the numbers of undecideds remain the same.

    To me the most likely scenario is that a number of undecideds have shifted to Obama, while an equal number of previously-decided McCain voters have now become unsure. Doesn’t this seem more intuitively right than a bunch of decided people changing their minds completely, while the undecideds in between fail to make a choice? Scold me if this has already been put forth.

    Imagine a sort of upside-down bell curve, with “candidate preference” as the x-axis and “decidedness” as the y-axis. The more strongly you prefer a certain candidate, the more likely you are to consider yourself “decided”, with a sort of valley of undecideds in the middle. In such a situation, if one pushes the preference of the nation as a whole in one direction, then some of the undecideds will now be pushed over the threshold into decided territory, and on the other end, some who previously considered themselves decided fall into the valley.

    If you take this as a model, it brings up all sorts of interesting stuff, because we don’t know how the actual population is distributed along the x-axis. For example, if there are a tremendous number of McCain supporters who identify themselves as “decided”, but are very close to the valley, and there are relatively small numbers of people in the undecided group, then a shift in mass public opinion in favor of Obama would create a big point shift, not because he gains a lot of newly decided voters, but because previously decided McCain voters suddenly become uncertain. You would be able to detect such a situation if the poll margin changed and the number of undecideds expanded. That hasn’t happened, which would suggest under this model that the distribution of the voting public along the “preference” axis is generally flat.

  • Sam Wang

    DanM, it might not be good. But on the positive side we’d be very distracted by all the flying pigs. Joking aside, look at post-debate reaction data on issues such as who’s better on the economy, foreign policy and so on. If we assume that undecideds have latent preferences they simply aren’t expressing, the split of undecideds might be around 60-40 in favor of Obama. However, I will continue to implicitly assume a 50-50 split. In any event, we’ll soon have data with far fewer undecideds.

    Josie, likely voter screens include questions such as whether the person voted in the last election, the degree of intent to vote, and so on. Evidently it is possible to be a near-certain voter yet be undecided.

  • Josie

    Sam: Two equally ranked executives are standing in an elevator deciding what to have for lunch. They could go to the steak place or an Italian place nearby that are equally good and equally priced. How do they decide which place to go? (In this actual event, it took them five minutes to decide on the Italian place) The executive indecision was born of apathy, they didn’t care which place, as long as they ate. I imagine some undecideds really are at a preference midpoint, for various reasons. Maybe they dislike both candidates equally, like the episode of South Park. If undecided voters do have a subliminal opinion, is it enough get them to the polls?

  • DanM

    So, if Senator Obama were to win all of the states in which he currently is polling over 50%, but no others, I guess that would be equivalent to the assumption that all of the ‘undecided’ voters end up choosing McCain. For those of us who are incorrigible worriers, tell us about this worst-case scenario for the democratic candidate: what would the electoral result be?

  • Sam Wang

    One simple idea is that “undecided” is a midpoint between preferring Obama or preferring McCain. However, as I have indicated in the main post, I think it is at least equally plausible that people have a stated preference, but are unable or unwilling to articulate it. This idea is assumed by several of you (Jeff, Nicholas). The data above are consistent with the idea, for the following reason.

    Look at the change in decided-voter margin over time. At a minimum, this swing represents either the number of undecideds forming a preference that they can state, half that number of decideds changing their mind, or a mixture.

    To pick the largest example, in the last 4 weeks the swing has been 9 points. At the same time, the number of undecideds has stayed the same. The simplest interpretation is that at least some voters had an expressed preference, but switched.

    One caveat has to do with likely voter screens. Obama and McCain supporters might have maintained their preference, but their likelihood of voting has changed. However, this should still cause fluctuations in the undecided fraction, which aren’t observed. So for now, I lean against this hypothesis.

    Adriel, your argument is appealing, but “common sense” is a speculation, not a conclusion. I am not aware of historical data that support the idea that undecideds will break massively in one direction.

    Josie – Do people say they are undecided because they don’t have access to their true opinion? Are they truly undecided? Are they mulishly stubborn about answering some stranger’s questions? I suspect the first two are dominant factors, but I don’t know.

    Nic, there appear to be fewer “undecideds” in swing states. For example, since Sept. 1 they are 5.5 +/- 0.6% in OH (N=30) and 5.0+/-0.5% in FL (N=28). In regard to the Bradley effect, based on the last 12 years of gubernatorial and Senate races, it is on average zero. Any other assumption would not be a suitable default assumption. I have posted on this topic before.

    Everyone else: individual pollsters can guess wrong in likely voter screens and other factors such as weighting. But on average, state polls are quite accurate in predicting the winner. Keep in mind that we are talking about effects here that are likely to add up to no more than a few percentage points. They could matter, though not if the election were held today.

  • Hans

    Hello Dr. Wang:

    Thank you for this marvellous website and its interesting analyses.

    Are you aware of any data which shows what proportion of “undecided” people eventually decide just to stay home on election day, and thus don’t end up affecting things one way or the other?

    On a more general note, I’m guessing that most of the polls you use are of “likely” voters. Are you aware of any data which shows how closely the polling firm’s designation corresponds to reality?

    I ask these questions because it strikes me that the difference in voter turnout from election to election (as shown in the table here: ) seems to exceed the current meta-margin and I do not understand completely how turn out interacts with and affects your analyses.

    Again, thanks for your work.

  • Jack Rems

    One piece of this I can’t see how to break down would be if one candidate had a block of voters that lies to pollsters. Conceivably, you could have a piece of candidate X’s base that gives consistently false data for some nefarious purpose. Farfetched, I admit; and a pollster who counted a voter with a lawn sign or wearing a button supporting X who said he supported Y should have a mechanism to note this. Still, it seemed to me at the time a lot of evangelical Bush voters in 2000 and 2004 did not trust pollsters, and their systematic refusal to participate on one side screwed with the polls.

    It doesn’t look like any of this will matter this year, but it is something that could come up again.

    Has anyone seen a worthwhile discussion of statistical and/or meta-analysis use in detecting fraud, say in those Diebold voting machines or other classic fixed voting scams? Will we have U.N. election monitors in the U.S.? Should we?

  • Nic Ross

    Hi there Sam,

    Have been enjoying your website, blog and analysis. Was just wondering if you (or anybody else) would want to comment on two points:

    1) Are the numbers of “Undecideds” generally higher or lower in the key, swing states (my gut feeling would be higher).

    2) Is there any evidence of the notorious “Bradley effect” going on here. My (completely anecdotal evidence) suggests this could still definitely be a ~few% point issue, which means this race is a lot tighter than people think right now.

  • Josie

    I don’t know about you, but I’ve been informally polled by friends and acquaintances for about a year now. Since there is so much buzz, and people have to say something to their friends, is it possible that you can become committed to being undecided? Being in New Jersey, I am more concerned with unaccountable voting machines than undecided voters but the psychology of indecision has always intrigued me. It’s even a question on the Myers-Briggs test.

  • Jeff

    What I find most interesting is that the number of undecideds remain relatively unchanged during the most recent spike in support for Obama (see day -60 to present).

    Sam, would you agree that the data indicate most of Obama’s recent increase in support is due to ‘decided’ voters changing which candidate they support?

    There are similar events seen in 2004 (days -100 to -60) and in 2000 (days -60 to -25). I hypothesize that, in all three instances, what is observed is a solidifying of support that occurs among ‘decideds’ (note that in all three instances the percent undecided remain relatively stable), possibly a result of political events or themes. In 2000 it was Al Gore sighing at the debate and being an ‘exaggerator’. In 2004 it was the Swift Boat and flip-flopping theme taking hold in the mainstream media, and in 2008 it was the economic crisis. However, there is no data to establish a causative link in all three cases.

    In 2000, about 50% of ‘undecideds’ made their decision from about week 4 to week 1 before the election, and the other half chose their candidate in the final week. In 2004, there were about half as many ‘undecideds’ as in 2000, and they chose their candidate in the final week.

    Unfortunately, there is not enough information to draw a causative link between the observations… the best we can do is draw correlations and try to identify possible causative factors or possible confounders.

    Based on the number of undecideds and the fact that we have no incumbent running for president this year, I anticipate that this election will more closely resemble 2000 in tracking the movement of undecideds. I expect that we will see about half of undecideds make their decisions from here until about Oct. 28, and the other half of undecideds make their decision in the final week (10/28-11/4).

    I hope that the economic crisis was the major event of this election that CHANGED and SOLIDIFIED the support of ‘decideds’ for the remainder of the election. If true, than we can expect at most a 7% decrease in support for Obama to Obama +1%, and more realistically expect NO MORE than a 2:1 break to McCain yeilding a 4.7% decrease for Obama with a final ‘realistic worst-case-scenario’ of Obama +3.3%. Would a final result of Obama +3.3% corresponds to a 95% CI of 260-330 Obama EV’s? Do you agree with these interpretations?

    Thanks for your stellar analysis this year Sam!


    Well, absent a numerical analysis, common sense would suggest that true undecideds have more questions about the most known candidate (M) rather than the barely new candidate (O). Looking for reasons to vote for the new guy (O) should turn into actual votes for the new guy.

  • Nicholas J. Alcock

    Dear Sam,
    Have you analysed the correlation of movement of “undecideds” with the October spikes in 2004 ? Also, you state that the number of “undecideds” are equivalent to Obama’s poll lead. However, if every “undecided” moved to McCain so would a proportion of committed Obama voters. So, the outcome isn’t a tie but a McCain victory. But, statistical analysis must put the likelihood of McCain of winning all “undecideds” at approx. 1%. You have the data, is my guesstimate close?

  • Sam Wang

    Greg (gprimos1) – this would require repeated survey of a cohort over time. I don’t know that it’s been done.

    GermanDem – this year, wild horses could not drag an assignment of undecided voters out of me.

    I plan to provide a graph like this one, which you can then use to make your own prediction. I will offer my own best guess, but more in the form of an uncertainty rather than a prediction.

    For example, based on Franklin’s graph, about 5% of undecideds wait until the last week. The split is unlikely to be more than 2:1 in either direction. That gives a range of +/-2% change to the decided-voter margins.

  • GermanDem

    Great post and analysis, as usual!

    But what I and many readers would be most interested in would be your take on how one could estimate how the “not so undecided undecideds” might break. Maybe along a bell-curve of probabilities with the maximum at the most likely outcome (say undecideds breaking O 60%, M 40%). Ultimately including its meaning for the EV map if this projection is factored in.

    Do you have any indications, ideas or plans on doing some analysis concerning this question?

    I am aware that this issue was the spoiler of your “spot-on” prediction (before undecideds) in 2004…

  • gprimos1

    It would be interesting to analyze historical data to get a better idea of how undecided voters tend to vote. For instance, if candidate A is ahead of B by 60/40, do they tend to vote in proportions close to the decided voters (60/40), do they end up going 50/50 (thus not affecting the lead of A), or do they tend “revert to the mean” and vote for whomever is behind (something like 40/60). Does anyone here know if this analysis has been done already?

  • Sam Wang

    Dave, a major problem with your hypothesis is that the fraction of the electorate polled is quite small.

  • Andrew Foland

    A somewhat related point related to this week’s McCain tactics. The crypto-racism in McCain’s strategy seems entirely wrong for undecided voters. Voters who might by swayed by that have, already, surely already made up their minds (even if, as this post points out, maybe they don’t know it yet.) So the marginal value of the strategy seems awfully close to zero.

  • Eddie

    A terrific post.

  • Dave

    Polling itself may also affect results. There is abundant psychological research showing that when one asserts an opinion publicly, he tends to subsequently maintain and reinforce that opinion. Answering a poll is asserting one’s opinion publicly, and therefore whatever opinion one voices will tend to be maintained. As a result, I think decisions will be solidified by intense polling. Research shows this is likely to be subconscious.

    Note that publicly asserting that one is ‘undecided’ will also reinforce that upon polling, so people are more likely to say they are undecided after they’ve claimed to be so, even though they really are not. This is another explanation for the phenomenon described here.