Skip to content

No, no, nooooooooooooooooooo!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

I feel I have to respond to this item that people keep pointing me to:

John Antonakis and Olaf Dalgas presented photos of pairs of competing candidates in the 2002 French parliamentary elections to hundreds of Swiss undergrads, who had no idea who the politicians were. The students were asked to indicate which candidate in each pair was the most competent, and for about 70 per cent of the pairs, the candidate rated as looking most competent was the candidate who had actually won the election. The startling implication is that the real-life voters must also have based their choice of candidate on looks, at least in part. [emphasis added]


This came up a couple of years ago, when, in response to a similar study, I wrote:

It’s a funny result: at first it seems impressive–70% accuracy!–but then again it’s not so impressive given that you can predict something on the order of 90% of races just based on incumbency and the partisan preferences of the voters in the states and districts [at least in the U.S.; I don’t know about France]. If 90% of the races are essentially decided a year ahead of time, what does it mean to say that voters are choosing 70% correct based on the candidates’ looks.

I can’t be sure what’s happening here, but one possibility is that the more serious candidates (the ones we know are going to win anyway) are more attractive. Maybe you have some goofy-looking people who decide to run in districts where they don’t have a chance, whereas the politicians who really have a shot at being in congress take the time to get their hair cut, etc.

Anyway, the point of this note is just that some skepticism is in order. It’s fun to find some scientific finding that seems to show the shallowness of voters, but watch out! I guess it pleases the cognitive scientists to think that something as important and seemingly complicated as voting is just some simple first-impression process. Just as, at the next level, it pleases biologists to think that something as important and seemingly complicated as psychology is just some simple selfish-gene thing.

And see here for a discussion of some research by Atkinson, Enos, and HIll on this topic.

Just one more thing

From the news article:

“These findings suggest that voters are not appropriately weighting performance-based information on political candidates when undertaking one of democracy’s most important civic duties,” the researchers said.

No, no, no. Unless you want to take a very weak interpretation of “suggest.” Or, to put it another way, sure, I have no doubt that “voters are not appropriately weighting performance-based information on political candidates”–but I don’t see the personal appearance study as relevant to even close to definitive on this point.

I’m as cynical as the next guy, but this sort of thing is going a step too far, even for me.


  1. Christian Kleineidam says:

    If I understand it right then most US districts go to the person that was already in congress before. Maybe being a politician for a few years changes your appearance in a way that make yourself look like someone with experience in politics?

    In Germany you have a political system in which people who run in districts that they can't win directly are still able to go into the parliament because of our system with two votes.
    That might allow you to test the hypothesis of whether it's about people who are more serious.

  2. jrhs says:

    Let me examine the photos, I want to see if I have the power of picking the winner based on the looks. :)

    I wonder what the accuracy would be if those Swiss undergrads were given additional information about, say, the incumbency and the partisan preferences of the voters in the states. Worse than 70%? Better than 90%?

  3. Thomas J. Leeper says:

    A more interesting study was just published in POQ ( where the authors used photo editing software to morph candidate images with those of research subjects. Key finding: voters preferred candidates who looked like them, not realizing that the images has been altered.

  4. Seth Roberts says:

    Okay, how can the study be improved? Surely you don't think it's worthless and that nothing can possibly be learned from such studies.

    I think it would be a good idea to control for age. That is, make the students choose between people of roughly the same age.

  5. Anonymous Coward says:

    Not having read the paper, 70% doesn't seem very impressive. That sounds like the baseline 50%, plus picking the older-looking one, plus picking the ethnically European-looking one when the other one looks African, Middle Eastern, Algerian, etc.

  6. Radford Neal says:

    The linked article says…

    One possibility is that people's looks do actually correlate with their competence and it's that association that the participants in this study were tapping into. However, Antonakis and Dalgas note that past research shows there is no link between competence and appearance, at least not in terms of IQ.

    But if you follow the link to "past research" (active in the article), its abstract actually says that appearance and IQ are associated. This pretty much destroys the whole point of the article.

  7. alex says:

    You raise one substantive point – that the effect may be explained if "goofy-looking" people run in hopeless districts – but I really don't get your other criticisms.

    "it's not so impressive given that you can predict something on the order of 90% of races just based on incumbency"


    Suppose I produce convincing statistical evidence that I can predict elections with 80% percent certainty looking at tea leaves. Most people would probably be extremely surprised, but I presume your only reaction to this that you can already predict elections with 90% certainty using other data, so what does the result mean?

    I'm giving a fanciful example to try to make my point that these results are interesting because of the relationships they suggest may exist.

    I have no doubt that "voters are not appropriately weighting performance-based information on political candidates"–but I don't see the personal appearance study as relevant to this point.

    If the correct explanation of this phenomenon ends up being that people support more attractive candidates, then of course this study is relevant.

  8. son1 says:

    These same two Swiss researchers have a paper out in the most recent issue of Science, on exactly the same topic… except this time, they're testing children.

  9. Richard D. Morey says:

    Incumbency is simply about who has won previous elections. Appearance and incumbency should share a chunk of variance, if appearance accounts for any.

  10. Andrew Gelman says:

    Seth, Anonymous: I'm not sure about this study, but, when I commented on that earlier study, I noticed that the did adjust for age in their analyses. Regarding ethnicity, it's possible this is a factor here but I doubt it could explain most of what they're finding.

    Seth: I'm not saying the study is useless; what I'm saying is that it does not imply what the authors and the reporter seem to be thinking it is implying.

    Literally millions of research articles are published each year, and most of them get nothing short of this level of publicity.

    Alex: If you said that you could predict 80% of elections using tea leaves, I'd like to see what exactly are you doing. But, that aside, I agree with you that the study cited above is relevant to the issue. I don't see the study as proving anything or even as being particularly convincing to me, but it is certainly relevant, and I made an overstatement in my above post.

  11. tirta says:

    In some other studies, face effects persist after controlling for incumbency, showing that the relationship between facial characteristics and vote share, as Alex suggests, may exist.

    Specifically, this relationship is best interpreted in the context of the intuitive vs the rational brain; it shows that rapid, automatic trait inferences by the intuitive system can influence voting decision:

  12. James says:

    One possibility that my wife and I both thought of, as long as we're speculating:

    Incumbents usually win.

    Suppose incumbents are on average older than challengers. (Maybe because younger challengers keep having a first try at politics, with a large fraction quitting and doing something else when they lose elections, while older people are less likely to enter politics at all, so the most senior people left in politics are those who had their first success a long time ago, and who have been in office ever since.)

    Suppose most people translate "competent-looking" into "older-looking".

    In this case, asking people who looks the most "competent" is simply a muddling and confusing way of asking "who looks oldest and therefore most likely to be the incumbent"?

  13. noel says:

    There is a subset of psychologists that work within IDT (interpersonal deception theory) that put forth the idea that as young children those with desireable features (baby faced good looking kids) grow up facing few consequences for their lying behavior on account of their believable mugs. Therefore they learn to look believable when they lie creating good looking politicians as a process of habit and evolution. Not as attractive kids are punished when they are children for the first lies they tell and are less believable as time goes on just by virtue of looking suspicious (not cute). They try to tell the truth more (avoid punishment), find themselves being not believed by parents, grow large chip, try to remove it through politics…

    cute baby wins, not so cute baby grows bigger chip

  14. jeremy harris says:

    Gelman, before shouting at the heavens, why don't you say what your criteria would be for rejection of the null? I think you could repeat your same arguments even if you observed 100% accuracy at prediction. But I guess then for sure you would be right that there was something else going on. :)

  15. Anonymous Coward aga says:

    The other obvious criticism of work like this — again, having read exactly none of the relevant articles — is that the only way that their posited mechanism can work is if a sufficiently large number of people know what both the incumbent and the challengers look like. A person who does not know what both candidates look like cannot have cast her vote on that basis.

    I have no idea whether French campaign advertising and coverage features enough photos and videos of legislative candidates to make this necessary condition believable.

  16. Andrew Gelman says:


    What particularly irritated me was the flat-out no-qualifications statement that people were voting based on looks. I think if the researchers had forced themselves to state the descriptive formulation first, they might have quickly thought of alternative explanations.

    And, yes, I realize that my statement above is itself a causal conjecture not directly supported by any facts.

  17. Ned says:

    To Neal Radford:

    Why don't you read the abstract of the linked article carefully, where they note: "Judgments were more accurate than chance in childhood and puberty, marginally more accurate in middle adulthood, but *not more accurate than chance in adolescence or late adulthood*."

    In fact, Antonakis and Dalgas stated explicitly that intelligence cannot be inferred from appearance in *adults.*

    Ironically, this pretty much destroys the whole point of your posting.

    The results of Antonakis and Dalgas are very impressive, but also very disturbing.