Psycasm is the exploration of the world psychological. Every day phenomenon explained and manipulated to one's own advantage. Written by a slightly overambitious undergrad, Psycasm aims at exploring a whole range of social and cognitive processes in order to best understand how our minds, and those mechanisms that drive them, work.
My posts are presented as opinion and commentary and do not represent the views of LabSpaces Productions, LLC, my employer, or my educational institution.
Please wait while my tweets load
[Wherein our Hero discusses a difficult conceptual problem, and explains why you're probably not as hot as you think]
So I’m dealing with a pretty big conceptual problem at the moment. It’s part of the study I’m currently conducting on behalf of another.
The study involves, in part, morphing a participant’s face with that of a more attractive target and with that of a less attractive target.
It follows the methodology of Epley and Whitchurch (2008) who found that, in a task where participants were asked to ‘estimate the likelihood that [a given] face is their own’ they were more likely to pick 20% more attractive than their actual face.
They conducted a number of other experiments within their paper, and generally found that 10% and 20% more attractive faces were routinely accepted as one’s own more readily than 20%+ and actual and ugly-morphed faces. For the astute, there was frequently observed a decline at 10% ugly that is not of the magnitude expected. In discussions we’ve concluded that this is likely due to the fact that morphing (to a small degree) stands to ‘average out’ a face. Thus features that might be anomalous (say, a big nose; a slight asymmetry) get worked out even with ugly faces, and create a strange illusions that begs the question is it me, and is it attractive? Generally there’s a particular features that makes the face less attractive, but there’s something about a ugly-morphed face that is also a little more average (and thus, slightly more attractive than the given anomaly).
Now I need to explain this x% ‘more attractive’ idea. Essentially we take the face of a participant, and one pre-rated attractive face and one pre-rate ugly face. Let’s say, for the sake of simplicity, that the attractive face is Brad Pitt and the ugly face is Steve Buscemi. Now when you morph a face you match up features and, using a sliding scale, create a 10% Brad Pitt morph on your own face, or a 90% Brad Pitt morph. In the later instance it’s essentially a picture of Pitt but with a 10% influence of your own face on his.
Can you see the problem?
Let’s assume I’m a completely average looking guy - A 5 out of 10.
On the following scale I am represented by the ‘ME’ and each ‘I’ represents a 10% increase in face-likeness to the target. If I am a hypothetical 5, Pitt is a hypothetical 10 and Buscemi is a hypothetical ‘0’ [11 point scale, myself at the midpoint] then each 10% increment represents an equal deviation from myself to the target.
[50% Buscemi] I—I—I—I—I—ME—I—I—I—I—I [50% Pitt]
However, let’s now assume I’m an 8 out of 10.
[50% Buscemi] I----I----I----I----I----ME-I-I-I-I-I [50% Pitt]
…or that I’m a 3 out of 10.
[50% Buscemi] I-I-I-I-I-ME----I----I----I---I---I [50% Pitt]
As you can see, a 10% morph does not create a change of equal magnitude depending on your starting point. It’s harder for a good looking guy to get better looking, and harder for a less attractive guy to get less attractive.
In fact, during the morphing process, making an less attractive face less attractive is really quite problematic. As I said before, the acting of morphing a face to a small degree (say, between 2 – 15%) actually makes the less attractive face better looking because it averages out most of the anomolies. This, I believe, explains the less-than-expected decrease at -10% in the Epley and Whitchurch (2008) paper.
The obvious solution is to try and estimate the ‘true magnitude change’ in faces between participant and target. Maybe, a participant (an 8 out of 10) morphed with 17% Brad Pitt is equal to the magnitude change of only 4% Steve Buscemi. In a sense, we need change measured in standard Pitt units and Standard Buscemi units. But this is problematic, subjective, and difficult to quantify or justify in publication.
I’ve personally come up with a few solutions – ball park stuff – which I believe address certain areas of the problem, but are not complete enough to use.
Research into the self-enhancement literature has revealed a few confounds as well. Asking someone to directly compare themselves to a target is more likely to lead to closer approximations to the target than if you manage to implicitly direct them to make a comparison (Stapel & Suls, 2004) - One caveat: it seems to magnify your dominant response. If you think you’re an 8 out of 10, and you’re asked to compare to Brad Pitt, you’re more likely to assimilate your score towards him. But if you think you’re a 3 out of 10 (say, depressives and folks with body disorders) they tend to push away in their scores. Implicit direction seem to elicit more conservative responses in estimates – but it’s unclear whether conservative = true self-perception. I’m inclined to believe it does, at least more so than explicit measures.
Furthermore, people’s estimates of their own abilities seem to be better predicted by levels of narcissism than actual scores (John & Robins, 1994). This is true at least in measures of performance and ability (like interpersonal skills and athletic prowess), it’s unclear if this is also true for things like body-image and measures of self-attractiveness. I’m inclined to believe it is, but I have no evidence for that statement.
So my question to you, and one I’ll continue to tackle – how can I get at true measures of own’s own estimation of their attractiveness which accounts for:
- True magnitude change (because the faces do need to be altered)
- Implicit direction and/or Priming instructions
…and of course, the underlying message is that you’re not as good-looking as you think. You’re probably not as smart, athletic, witty or generally competent. Additionally, you probably also fall prey to thinking you’re above average in intuiting how other people feel and your estimate of your ability to drive a car is above average, too.
These are all interesting topics worthy of a post of their own (once this research is complete), but the good news is it probably doesn’t matter you’re wrong. Unless you’re wrong by such a degree that people call you out, chances are you’re getting away with it ;)
Epley, N., & Whitchurch, E. (2008). Mirror, Mirror on the Wall: Enhancement in Self-Recognition Personality and Social Psychology Bulletin, 34 (9), 1159-1170 DOI: 10.1177/0146167208318601
John, O., & Robins, R. (1994). Accuracy and bias in self-perception: Individual differences in self-enhancement and the role of narcissism. Journal of Personality and Social Psychology, 66 (1), 206-219 DOI: 10.1037/0022-35126.96.36.199
Stapel, D., & Suls, J. (2004). Method Matters: Effects of Explicit Versus Implicit Social Comparisons on Activation, Behavior, and Self-Views. Journal of Personality and Social Psychology, 87 (6), 860-875 DOI: 10.1037/0022-35188.8.131.520
This post has been viewed: 1441 time(s)
Apparently face morphing is not necessary. Other measures of self-perception of attractiveness are acceptable.
For instance: If one were to rank order 20 faces from ugly to beautiful one could be asked to mark where their own face fits. Independent rankers would later identify if their self-rank was appropriate. Obviously this is confounded, but highlights a possible non-morphed solution.
I like Rift's idea.
For one thing, I'm a woman who thinks Brad Pitt is below average looking. I think Gwenyth Paltrow is prettier than Angelina Jolie, but I'm a dark-haired, dark-eyed, plump-lipped woman, and whether morphing my face with one of theirs looked like "me" to me or not, would depend more on which one you pick than how pretty I find them. So you've got 2 confounding problems there: one, that not everyone will measure a stranger's face attractive or not in the same way, and two, that morphing faces will have different results not only based on the attractiveness of the other face, but how alike or different they were to begin with.
Yeah, you've hit a few points for sure. What the hell is attractiveness anyway? In our study we intend to use pretested models who are consistently rated at a certain level of attractiveness. In psychology there's usually a healthy amount of error variance though, so although you find Paltrow better looking than Jolie, over enough participants it shouldn't really matter - but you're right, it's a geniune consideration.
In addition to the pretesting we're finding (and it's in the literature) that masculine features are consistently rated as more attractive (strong jaws in men, for example). This serves to reduce a bit of the error variance with regard to 'taste' or preference.
So again, you're right. Morphing might not be an optimal strategy to assess our hypothesis - it's just one possible avenue.
Really, Brad Pitt below average? Does that mean he's not attractive at all, or just not physically attractive to you?