How con-men make their faces look trustworthy

200808201050.jpg

Drake Bennett of the Boston Globe wrote an article on the various ways con men gain their marks' trust, including body language, verbal language, and facial expressions.

When deciding who to trust, the research suggests, people use shortcuts. For example, they look at faces. According to recent work by Nikolaas Oosterhof and Alexander Todorov of Princeton's psychology department, we form our first opinions of someone's trustworthiness through a quick physiognomic snapshot. By studying people's reactions to a range of artificially-generated faces, Oosterhof and Todorov were able to identify a set of features that seemed to engender trust. Working from those findings, they were able to create a continuum: faces with high inner eyebrows and pronounced cheekbones struck people as trustworthy, faces with low inner eyebrows and shallow cheekbones untrustworthy.

In a paper published in June, they suggested that our unconscious bias is a byproduct of more adaptive instincts: the features that make a face strike us as trustworthy, if exaggerated, make a face look happy - with arching inner eyebrows and upturned mouths - and an exaggerated "untrustworthy" face looks angry - with a furrowed brow and frown. In this argument, people with "trustworthy" faces simply have, by the luck of the genetic draw, faces that look a little more cheerful to us.

Just as in other cognitive shorthands, we make these judgments quickly and unconsciously - and as a result, Oosterhof and Todorov point out, we can severely and immediately misjudge people. In reality, of course, cheekbone shape and eyebrow arc have no relationship with honesty.

Judging trustworthiness in the face (via Mind Hacks)

44

  1. Does anyone else see the Masculine/Untrustworthy – Feminine/Trustworthy pattern?

    It may be, that there is a deep rooted masculine, aggressive, robust vs feminine, gracile, passive dichotomy inbuilt in the reptilian part of our primate brains.

    The trick then is to mimic these characteristics in certain social situations.

  2. #3: I see the “trustworthy” end as more “juvenile/childlike” then feminine, but considering that a lot of cultures (*cough*scrawnyhollywoodteenagers*cough*) tend to get feminine and juvenile muddled up together, I’m not surprised.

  3. The case studies in the article are really interesting, but:

    “That is not to say Gerhartsreiter was harmless – he is, after all, a person of interest in an unsolved disappearance in California – but fooling people seems to have been not merely a means but an end.

    Con men have a term, “taking off the touch,” for the point in the con when they take the mark’s money. Gerhartsreiter doesn’t seem to have had much plan for taking off the touch. When he finally did steal something, it was his daughter, and it’s hard to imagine that was for financial reasons. His divorce settlement had given him enough to live on. But that, apparently, was not all he needed.”

    Oh my god ew. First of all, a child is not “something” that you steal. Second of all, child abduction is not something done out of love.

    So in all: “Although he was a manipulative person and perhaps a sociopath and/or murderer, he didn’t steal money, he just abducted his daughter–so he’s a nice dude after all! who just wanted love!”

  4. So how did they come up with this? Show some test subjects a bunch of faces and ask them to pick the most trustworthy looking? I call baloney on this whole study.

  5. Hmm. this is interesting.
    Of course it’s not a good thing to not be trusted if you’re trying to sell something legit, so learning how to look more trustworthy (or at least not untrustworthy) isn’t a bad thing.

  6. “In reality, of course, cheekbone shape and eyebrow arc have no relationship with honesty.”.

    Really? Are we positive about this? I mean, there’s probably not a correlation, but I’m wary of any truth that’s so obvious that nobody bothers to test it.

  7. Um, guess the fact that the untrustworthy face has a big ol’ frown on it while the trustworthy face is smiling is completely irrelevant, as is the fact that the mean frowny face’s gaze is confronting the viewer, while the smiley trustworthy face is looking upward.

  8. #10: That’s what I was about to say.

    The quote

    “In reality, of course, cheekbone shape and eyebrow arc have no relationship with honesty.”.

    actually makes two implications,

    1: That the differences between the faces are due to eyebrow arc and cheekbone shape, and that

    2: The differences in the faces have no relationship to honesty.

    1, of course, is nonsense. The face on the left has his eyebrows deeply furrowed and his mouth is scowling. This is a face that was made, not a case of bad bone structure.

    Given 1, 2 is also not obvious. When I see a crazy person on the street, I can often tell they are crazy — and not the kind of person I’d want to buy a used car from — by their facial expression. I would say it’s very likely that our evolution in social groups has equipped us with ways to tell the friendly from the non-friendly.

    This is not to say that our intuition can’t be duped by con-men, but it’s perfectly plausible that it is actually accurate much of the time.

    So, first, their “of course” statement isn’t “of course” at all, and second, it would be an interesting study to see how often we can pick out the untrustworthy types by face/video/voice/whatever alone.

  9. I’m a three-card monte dealer who’s been widening his chin and pronouncing his cheekbones for years now. People really fall for it!

  10. By this analysis I think susceptibility to gravity could make us look untrustworthy. I wonder if we’d look friendlier in space. Also, Robbaily, you’re right!

  11. Also, re: cheekbones et al “of course” not being related to trustworthiness.

    As the Wikipedia article on testosterone notes, an increase in testosterone leads to, among other things, “Growth of jaw, brow, chin, nose, and remodeling of facial bone contours.”

    So, while the author says “of course, cheekbone shape and eyebrow arc have no relationship with honesty,” would he be so quick to say that testosterone has no relationship to honesty.

    I have no idea if it does, of course, but I doubt that the statement would have been made so casually.

  12. Here’s a theory: 1) People deep in thought tend to squint or furrow their brow, lowering the inside of the eyebrows. 2) In movies and television (from which modern people seem form their views of reality) villians are often depicted as evil geniuses–thus the facial expression of a person who is often deep in thought (i.e. thinking about one’s master plan to take over the world) comes to be associated with evil. 3) The result: Thoughtful people are perceived as untrustworthy while those who are not so thoughtful are perceived as innocent. This would explain why no one trusts Al Gore but everyone wants to have a beer with George W.

  13. I guess people with severe Microcephaly (pinheads) are more trustworthy than normal folk. No one noticed the HEAD SHAPE CHANGE FROM LEFT TO RIGHT?

    Forget you suck at Photoshop. You suck at (Adobe) Illustrator.

  14. Using the Facial Action Coding System (used by Pixar, other animators, and researchers) it looks to me like this is…

    UNTRUSTWORTHY:
    AU4: inner eyebrow lowering (like you’re solving a puzzle)
    AU2: outer eyebrow raising (hard to do, especially with AU4!)
    AU7: mostly like eye squinting (think Clint Eastwood)
    AU9: very mild nose wrinkling? or
    AU10: very mild upper lip raising (although AU10 usually results in a distinctive flat looking middle upper lip)
    AU15: mild amount of pulling the mouth corners down and out diagonally (not hard to do voluntarily)
    AU17: moderate chin raising (easy to do)
    AU??: nostril dialator (most people can’t do it)

    TRUSTWORTHY:
    AU1: middle forehead raising
    AU5: sort of like eye opening
    AU6: outer cheek raiser (very hard to do voluntarily without adding other junk in with it)
    AU11: mild amount of one of the sets of smile muscles (not the big easy one to do, which is AU12)
    AU13: perhaps some of this, which is a nearly impossibly involuntary smile muscle group which pulls the corners of the mouth nearly straight up
    AU??: nostril constrictor (most people can’t do it)

    With enough training it’s possible for someone to mimic the most of the trustworthy face on cue.

  15. > In reality, of course, cheekbone shape and eyebrow arc have no relationship with honesty.

    On the other hand, if you have facial characteristics which don’t engender trust, you may become less trustworthy over time as a reaction to the societal expectation.

  16. Aloisius – “I like how they made the broad flat nose (typical of Africans) untrustworthy.”

    By “they” do you mean the researchers, the study participants, Princeton, the journalists, or the Gray elders?

  17. So…the point of the article is that con men get plastic surgery all the time?

    #6 MsAnon:

    “Oh my god ew. First of all, a child is not “something” that you steal.”

    Those were the article’s words, not the guy’s.

    “Second of all, child abduction is not something done out of love.”

    It’s “abduction” because, presumably, he didn’t get custody. Otherwise we would call it “a parent bringing his child home.” Maybe a court rightly decided he wasn’t a good parent, but even if he was a bad parent and an outlaw, you don’t have enough information to say whether he was acting out of love.

    “Although he was a manipulative person and perhaps a sociopath and/or murderer…”

    Neither “sociopath” nor “murderer” appears in the article.

  18. is it just me, or has anyone else noticed how the facial features typically associated with people of color are on the far end of this supposed “trustworthiness” scale? i can’t help but think that a big explanation for the outcome is simple social conditioning. through white patriarchal hegemony, we learn to view modern criminals through lenses that were created by systems of oppression and racial domination.

    i see merit in such studies, but i also can’t help but think about where that time and effort could have been redirected into the community.

  19. #27: At least unconsciously, the article would suggest that yes, they do. In fact, there’s some slightly more disturbing research by the same people on how faces affects election results.

    #7: Please read a little about experimental design in psychology before making sweeping, ignorant statements like that.

    #20: So the article talks about more characters than just smiling, but you pick on that to make a sarcastic criticism of it? Shocking, I say.

    #21: I haven’t read the actual research paper behind this, but I rather suspect that yes, overall width and shape of the face is a feature here. And I really, really doubt the researchers used any commercial graphics tool to generate these images.

    #23: Thanks :)

  20. Surely trustworthiness must be only one of the traits that people form biases based on facial features.

    I’m sure similar studies could be done regarding biases about intelligence, leadership ability, kindness etc.

    I’m willing to bet that sharp, strong features would be generally thought of to indicate intelligence, decision-making, assertiveness and rounder more open features indicating generosity, kindness, trustworthiness etc.

  21. Normally I can count on Boing Boing to cull interesting articles from the web AND offer valuable insight. This reposting offers no extra insight. In fact, it includes the same editorial error that the original article includes:

    “In reality, of course, cheekbone shape and eyebrow arc have no relationship with honesty.”

    The “of course” aside should have been a major red flag to the Boing Boing editors that the statement was untested.

    A number of explanations for this inclusion exist. Bennett did not cite the various journal articles he used as sources for the article, so it is hard to tell if the idea originates with him or Oosterhof and Todorov.

    No matter why the original article included this error, Boing Boing should have picked up on it and should not have included it in the excerpt.

    Please be more attentive. I value Boing Boing too much to see this junk get through.

    Thank you to the other readers that also caught this (10, 15, 24 — cheers).

  22. Looks like one of those corny Before and After ads. “Hey you, on the left, look angry…and you, on the right, look like a simpering idiot”.

    Another question could be, is trust relative? Which of the above would you trust to protect you from Tony Soprano?

  23. @StellarsJay – I’m not sure you read the article (or comment 10) right.

    The correlation on perception WAS tested.

    The correlation between perception and reality WAS NOT tested

    Thus the ‘of course’ statement when he went on to say that facial structure does not have a proven correlation to actual trustworthiness, but does have an correlation to perceived trustworthiness.

  24. @37: “Thus the ‘of course’ statement when he went on to say that facial structure does not have a proven correlation to actual trustworthiness….

    He isn’t just saying there’s no proven relationship, he’s saying that “of course” there IS NO relationship. He says this without explanation, as if it’s obvious.

    However, it is far from obvious, because (1) the implication that the study is only changing bone structure is incorrect, as the facial expression also changes, and (2) testosterone can change bone structure (particularly chin and cheekbone structure), as I cited above, and I don’t see any evidence of an “of course” lack of relationship between testosterone and bone structure.

  25. …sorry, that should read “and I don’t see any evidence of an “of course” lack of relationship between testosterone and trustworthiness.

  26. …Ok, so the more egg-shaped or pinheaded a head is, the more trustworthy the person. I get it.

    [shakes egg in utter dismay]

  27. @ #24

    On the other hand, if you have facial characteristics which don’t engender trust, you may become less trustworthy over time as a reaction to the societal expectation.

    I believe the effect can be profound on developing personalities. A nice kid with an untrustworthy face can find himself accused unjustly, disliked, picked at, snapped at and suspected by others (particularly teachers and other authorities) for no reason he can see or understand. He can find himself treated a lot less favorably and punished a lot more often than children who have more childlike, trustworthy faces.

    So yeah, people’s faces may not be an indicator of who they intrinsically are, but it can be a big factor in how experience shapes their feelings.

  28. A friend of mine was talking about a character on the new Battlestar Gallactica, of which I’d only seen snippets.

    “Which guy?” I asked. “You mean the one who looks like a thug?”

    “Yeah, him,” he agreed.

    And I bet you know exactly which guy I’m talking about, too.

Comments are closed.