Princeton Election Consortium

A first draft of electoral history. Since 2004

Jul 07: Biden 375 EV (D+6.8% from toss-up), Senate 53 D, 47 R (D+5.7%), House control D+6.0%
Moneyball states: President AR IA AZ, Senate MT KS ME, Legislatures KS TX NC

The undecided voter’s secret ballot

September 14th, 2012, 1:00pm by Sam Wang

As I wrote yesterday, persuadable and undecided voters in 2012 seem quite scarce. What is actually going on in their brains? In 2008 my colleague Josh Gold and I wrote for the NYT on this subject. I also write about hidden biases in how we make decisions (in all aspects of life) in Welcome To Your Brain (see the left sidebar).

Despite your likely incredulity that they even exist, undecided voters may be smarter than you think. They’re not indifferent or unable to make clear comparisons between the candidates. They may be more willing than others to take their time. It is even possible that they have essentially already made a choice, but are unaware of the fact.

In the natural world, self-awareness in decisionmaking is unnecessary. The true readout of a decision is action. Action gets us what we want right away: escape from a predator, access to food or a mate, and so on. In many cases, knowing what it is we want is only useful on longer time scales, for instance to allow explicit learning. So in some cases, undecided voters may have actually committed, but don’t know it yet.

This is why pollsters push “uncommitted” voters to state a preference: the answer gives a fairly accurate reading of a candidate’s support. In psychological studies, people who describe themselves as undecided often reveal a pronounced preference when they are forced to choose. My colleague (and Nobel laureate) Danny Kahneman’s collaborator, the late Amos Tversky, used to ask his colleagues who were offered a new job whether they were inclined to take it. He found that when they were even only “moderately sure” about their decision, their eventual choice was in fact all but certain.

Hidden commitments can be queried in creative ways that go beyond pushing “leaners” to commit. In one study, 33 residents of an Italian town initially told interviewers that they were undecided about their attitude toward a controversial expansion of a nearby American military base. But researchers found that those people’s opinions could be predicted by measuring how quickly they made automatic associations between photographs of the military base with positive or negative words.

Pollsters implement this idea by using questions they already ask that are likely to correlate with support for either candidate: Who do you think understands your problems better? Are you more concerned about the economy or terrorism? Which candidate has the better temperament? Patterns in the answers of decided voters could be used to predict the final choice of undecideds. One could even track poll respondents repeatedly over time, as is being done in the RAND study I mentioned yesterday.

Mitt Romney has to find a way not to measure undecided voters, but to change their opinion. Partisan appeals like the Benghazi business are unlikely to do the trick, since undecideds who are moved by such a thing are already committed, even though they don’t know it.

A possible lesson in the hidden steps behind shifting candidate preference comes from the Democratic convention. As I pointed out the other day, President Obama’s approval/disapproval underwent a 16-point swing within a day of a powerful speech by Michelle Obama on Sept. 4.  Yet the Obama/Romney margin seems to have taken at least one day longer to even begin to appear, as evidenced by a close look at Pollster’s post-DNC tracking poll wrap-up. Job approval is just one piece of information that goes into establishing candidate preference – and may take a while to make its impact.

Yesterday, the Votemaster pointed out an anecdote that might have helped Romney with some of the voters that Michelle Obama reached:

While Ann Romney got generally good marks for her speech at the Republican National Convention, she may have blown her opportunity to do what she was supposed to do: humanize her husband. Judith Grey has written an excellent piece on what she should really have said. She should have focused on her multiple sclerosis and how her husband stood by with unwavering support through good times and bad times. Off stage he once said to her: “I don’t care how sick you are. I don’t care if you’re in a wheelchair. I don’t care if I never eat another dinner in my life. I can eat cereal and toast and be just fine. As long as we’re together, everything will be OK.”

I have to say that this is the most moving thing I have ever heard about Mitt Romney. Too bad for him it was not used.

I’m no Jonah Lehrer. I freely admit that several paragraphs were lifted from the NYT piece. I thank Josh Gold for co-writing the original with me.

Tags: 2012 Election · President · Uncategorized

17 Comments so far ↓

  • Amitabh Lath

    I was wondering about undecided voters, polling bumps etc. Sorry to nerd out, but here is my model: Ferromagnetic domains.

    Most spin domains line up pretty quickly. Some domains are hard to budge, pointing neither up nor down. Then there is an extra jolt in one direction (Ryan! Michelle Obama and Bill Clinton!) and these undecided domains start to line up. But the jolt is momentary, and there is a relaxation time constant, and back they go to unaligned.

    But spin-spin interactions are pretty strong. If you have a really strong jolt (or a long lasting one) large enough majority of the undecided domains to line up. And once aligned will start to pull other undecided domains in their direction. Even after the original jolt is gone.

    In ferromagnets this happens quite suddenly. One moment you have a lump of iron, and the next you’ve got a permanent magnet. A huge jolt sets a magnet’s polarity.

    Flipping polarity once things are set is really difficult, you need an opposing jolt of immense magnitude.

    • Sam Wang

      That is indeed very nerdy.

      One way in which opinion has a certain amount of “stickiness” is the fact that once we develop a worldview or opinion, we start to evaluate new information in the context of that framework. Data that are consistent are more easily assmiliated, a phenomenon called “biased assimilation.” We also ask tougher questions about information that doesn’t fit. This is called “motivated reasoning.” I wrote about these phenomena in the NYT a few years ago.

      These phenomena make sense from a Bayesian standpoint: information that fits our prior views is more likely to be correct. But in the political domain, it leads to opinions becoming hard to dislodge.

    • wheelers cat

      Because of how learning works, it can take up to 10,000 reps to “unlearn” a falsehood.
      This is widely exploited by trial lawyers.
      Its called catalepsis, meaning to be seized. The lawyer introduces a salacious untrue statement, knowing it will not go on the record, but also knowing it will persist in the jurors brains.
      I like the ferromagnetic immense magnitude jolt analogy. People with organic conservative tendency (more grey matter in the amygdala) have been shown to exhibit “backfire effect” where correction actually increases the salience of the falsehood.
      It would indeed take an opposing jolt of immense magnitude to change their minds.
      Heres the U Mich paper, if anyone is interested.

  • hankleberry

    This is psychology, not neuroscience. I sure wish neuroscientists would stop taking behavioral research in psychology, strapping “neuro” to the front of it, and claiming it as their own. As if we didn’t know that these processes occur in the brain. Wow!

    Love your site!

    • Sam Wang

      Hankleberry, I write on neurobiology, psychology, and human development on a regular basis for general audiences. I pick out work that I feel is relevant and well done. Because understanding comes from different levels of analysis, I do not draw the disciplinary distinction that you do. But yes, my home field is neuroscience.

      I left out the neurophysiology for purposes of addressing the subject at hand. That gets into current research, which is not really the subject of this website. Psychology has established interesting groundwork, the neural mechanisms of which can be very illuminating. Eventually concepts such as affective and cognitive decisionmaking will be sharpened by understanding those mechanisms.

      I understand how a psychologist might be apprehensive about the changing scientific landscape. In my view, you should be pleased that the questions pioneered in your field are now being studied at multiple levels.

  • Amitabh Lath

    Thank you for the link to the NYTimes article. I remember reading about the basic idea: repeating a false rumor could actually reinforce it.

    So it makes sense that people reject evidence that creates tension with their pre-held beliefs. And they would associate with other people who share these beliefs, where the tension-making evidence can be mitigated by peers who are also rejecting it.

  • G. Camp III

    ‘Mitt Romney has to find a way not to measure undecided voters, but to change their opinion.”

    He’ll have to get around this: (or here:

  • Mike B.

    Dear Prof. Wang,

    Could you please explain to a layman what criteria is used at the end of September that allows a different level of prediction? Will the numbers change dramatically from the tight range in the 80s that you have had for months in your opinion?

    Any reply would be appreciated, and a heartfelt thank you for the good work you do.

    • Sam Wang

      Mike B., on a short enough time scale, opinion moves slowly enough to be more predictive than the general range. For a brief discussion see this.

  • Olav Grinde

    @Amitabh Lath: “Thank you for the link to the NYTimes article. I remember reading about the basic idea: repeating a false rumor could actually reinforce it.”

    Isn’t that the whole premise of Fox News? Moreover, the faux news network gives many readers immense emotional satisfaction by giving them readily identifiable targets for their anger, blaming, hatred, etc.

  • Renee

    Sorry but there is nothing Mitt can say to this former strict conservative, now independent women voter (due to McCain picking Palin) that would change my mind about him. Everything he says has a motive. He may really be that man you say in the statement about his wife, but sadly he comes off as the man that would toss her under boss also, if the crazy right wing of the GOP were against her. He has just done to much damage by a$$ kissing the crazy.

  • Andy

    @Renee: “Toss her under boss”…Charming! …and, one assumes, unintentional.

    In my experience undecideds are either apolitical or believe they are not represented by either major party…but they all have a default to one party or another, usually based on their family’s longtime political affiliation. This seems to be consistent with the large group of so-called “independents” I know who will agree with me on virtually any issue in the campaign, but then say their minds cannot be changed: They’ve always voted Republican.

  • ICDogg

    I know people who consider politics to be a toxic subject. They hate talking about it, they hate thinking about it.

    They probably would not be considered “likely voters” but they vote once in a while.

    They are not unintelligent people. They just hate negativity in their lives, and politics is almost all negative.

  • The undecided voter's secret ballot | Neuro Physiology Blog

    […] Read the article: The undecided voter's secret ballot […]

Leave a Comment