judg.me (blog)

judg.me – first impressions count

What makes one appear smarter and more sociable?

It has been a few months since judg.me launched, and thousands of photos have been uploaded and judged by users since.

We wanted to get to the bottom of how people make snap judgments about others, so we donned our Hardly Boys hats, and got down to following some raging clues (read: unsexy data crunching). What is it that makes Person A appear more extroverted but less smarter than Person B? Does ethnicity matter? Hairstyle? The environment?

A sample size of 1000 random photos, each with at least 100 ratings, was selected. As no descriptive information is provided when a photo is uploaded, this process involved going through each photo to analyze certain characteristics that the user possessed. To standardize comparisons across the photos, we only focused on the following 12 traits:

  •   Ethnicity: Asian / Black / Indian / Latino / Middle-Eastern / White
  •   Gender: Female / Male
  •   Glasses: No / Yes / Sunglasses
  •   Cap/Hat: No / Yes
  •   Hair Color: Bald / Black / Blond / Brown / Dyed / Grey / Red
  •   Hair Length: Bald / Short / Medium / Long
  •   Facial Hair: Mustache Only / Beard Only / Mustache + Beard / Stubble / None
  •   Environment: Indoors / Outdoors
  •   Visibility: Face Only / Body Only / Face + Body
  •   Facial Expression: Neutral / Smiling / “Blue Steel” or “Duck Face”*
  •   Body Type: Normal / Fat
  •   Alone in the photo?: Yes / No

*Note: The “Blue Steel” or “Duck Face” option was checked
wherever a user had an exaggerated facial expression.

Keep in mind that given the recent launch of judg.me, the sample size for each trait was not equal, meaning the insights gained from comparing across the traits could be skewed. For e.g., the male::female ratio was roughly 56::44, so this difference could have skewed the insights when comparing the results of both sets of users.

So far so good? Right. Let’s dive into what we’ve found.


Well, surprise, surprise. Asians are perceived to be the smartest of the different ethnicities taken into account, with an average “Smart” score of 5.850. Their social skills, however, clocks in at an average of 5.090, the second lowest among the different ethnic groups. Whites, Latinos and Blacks are all perceived to be more social than intellectual; this difference is especially pronounced in Latinos, who are judged to be the most social of the lot (average score: 5.793), while simultaneously being viewed most unfavorably on the Smart scale (average score: 4.848).

TL;DR: Be Asian if you want to appear smart; Latino if you want to appear extroverted.


The average lady is perceived to be both smarter (5.345 vs 5.123) and more sociable (5.314 vs 5.275) than the average guy. Let’s drill down further and see what happens.


Things are getting pretty interesting here. Indian women are perceived to be significantly smarter than their sistas (this, however, is probably due to the relatively small number of Indian females represented in the overall sample size), followed closely by Asian women. Black ladies are perceived to be the dimmest bulbs across all gender/ethnic groups, with a score of just 4.375. Asian guys come out top on the Smart scale with a score of 5.708; in comparison, white males have an average Smart score of 5.057.

TL;DR: Asian guys are perceived to be the smartest men around. Black females fared the worst on the Smart scale.


Are blondes really dumb? Blonds – male and female – are only beaten out on the “Dense” scale by redheads. Black-haired ladies and men were perceived to be smarter than they are sociable, while redheads are perceived to be significantly more sociable than smart. Perhaps not surprisingly, grey-haired men are judged to be the wisest guys around. Women with dyed hair are perceived to be the most extroverted group.

TL;DR: “Blonde” jokes should be modified to include redheads as well.


What’s interesting in this particular scenario is that there’s a significant difference between each group – hair length appears to play a massive role in affecting people’s perceptions. Females with short hair (boy-cut) are perceived to be the most sociable lot, while men with long hair are perceived to be, by far, intellectually and socially inept, relative to the other groups. Bald guys occupy the other end of the spectrum, being perceived as the smartest among men.

TL;DR: Bald guys are seen as the smartest guys around. Short-haired ladies are judged to be the most extroverted group.


How does wearing spectacles or sunglasses affect people’s perception of you? Those with glasses – both men and women – are assumed to be significantly smarter than those without, while simultaneously being perceived as less sociable. Wearing sunglasses appears to drastically improve how others perceive one’s social skills; however, sunglasses also have the effect of making one appear, well, more stupid.

TL;DR: Sunglasses make you appear relatively more outgoing – but more dense at the same time.


This section is, naturally, applicable to males only. Guys with beards (and no mustaches) have a Smart score of just 4.320, while they are perceived to be significantly more sociable. Those with only a mustache, on the other hand, are judged to be the most introverted of the lot, with a score of 4.096. Guys with a stubble are the best performers on the Smart and Social scales.

TL;DR: Go for the 5 o’clock shadow, fellas. It’ll do you good.


What can we learn here? We see that females with photos showing only their faces are perceived to be smarter than they are sociable. However, if a woman has a photo where she as much as shows her shoulders, she is perceived to be more sociable than she is smart. This difference is further reinforced by the setting of the photo – users who are outdoors are seen as significantly more sociable than those who are indoors. A single tree in the background in your photo could influence people’s perception of your social capacity!

TL;DR: Go outside if you want people to think you’re more sociable than you really are. Oh, and if you bare as much as your shoulders, you’d appear a heckuva lot more outgoing as well.


Once again, things look pretty obvious here. A “Blue Steel” or “Duck face” expression makes one appear much more dense than they would have seemed otherwise (4.954 for women, and 4.377 for men on the Smart scale). A neutral non-smiling expression makes one appear more smart than sociable, whereas users who smile are perceived to be both sociable and smart.

TL;DR: Life is short. Smile more often.

Pretty interesting stuff here. The differences in perceived intelligence and extroversion between fat and normal-sized women are much more pronounced than that between fat and normal-sized men. Fat women are perceived to have a Smart score of 4.946 (compared to a Smart score of 5.381 for average women), and a Social score of just 4.771, compared to a score 5.363 for average women. On the other hand, while fat men trail behind averaged-sized men on both scales, they are perceived to be more dense and outgoing than fat ladies.

TL;DR: Your weight has a much more significant bearing on others’ perception of you if you’re a woman. Tough luck.

This is somewhat related to the setting of the user’s photo. If the user is *not* alone – that is, if there are friends, or even strangers, in the background, the user is perceived to be significantly more sociable than if s/he were the only subject in the photo!

TL;DR: Even if you’re taking a self-portrait, try to take one with people in the background. It’ll make you look more outgoing.

This was a fun exercise in understanding how certain traits are mapped to one’s perception of a user’s intelligence and social capacity. As more photos and data points come in, we’ll be able to further refine our data analysis process to churn out even more interesting insights. Stay tuned.

PS: If you enjoyed the post, you should consider following me here on Twitter. Or learn more about me here.

29 thoughts on “What makes one appear smarter and more sociable?

  • droope says:

    Very interesting! I wonder what other things will impact other aspects of others’ perception of you.

    ^^ great job! thanks

  • Fiora says:

    (reposting from my tumblr so you can maybe see my suggestions…)

    This is a really interesting bit of statistics — some of the conclusions aren’t so surprising, but others are. The most surprising one to me was hair length — it had a huge effect, and not entirely the results I expected!

    The study really needs error bars though, so I’m not sure how much confidence to put in the results. There also might be some correlation that needs to be removed — like, for example, black hair is typically linked with Asian heritage, so the increase in perceived intelligence for black hair might be due to that confounding variable.

    I also wonder how this varies between cultures. Do hair length or beards, for example, have different connotations in different countries?

  • drew w says:

    it would be very interesting to see the distribution of ratings for each race, gender, and race/gender combination. additionally, the y axes are very much different from graph to graph, magnifying some differences while hiding others, and there’s not enough data to tell which if any of these variances are statistically significant.

  • Michael Chui says:

    It would be worth doing some statistical/geographical analyses on the people who did the rating, too.

  • James W Dunne says:

    Wow, I should really get a hair cut and lose some weight!

    Bizarrely enough, the Zoolander pose really does work. It’d be quite funny if you compared from the left or right and found yet another significant difference!

  • Scott says:

    Fascinating. Except the TL;DR kinda threw me off:
    “Be Asian if you want to appear smart; ”

    I don’t think you can choose your ethnicity can you?

  • That was interesting, keep up the good work!

    Minor nitpick: some of your histograms do not start at zero, while others do. That influences our perception even when we noticed. Plus, I felt that you tend to talk about large differences more when talking about a graphic whose bottom was cut of. I didn’t precisely checked, but maybe those difference weren’t so huge after all?

    Also, when you get more data, you may want to search correlations and check for confounding variables. Just in case all black women also happened to be fat, or scruffy looking men were also outdoors.

    • Al T. Rompson says:

      True. There’s a shitload of additional analysis that could be done (also, analysis of raters, where possible). Very interesting idea in general though.

  • Kyle says:

    You shouldn’t report extremely tiny effects so much (cf. race). Maybe report race after controlling for everything else, for instance. This is quantitative inference 101.

  • ct says:

    Wow I entertained a lot.
    Guys you should publish a paper about these data. Seriously

  • MB Dude says:

    You wrote “Guys with a stubble are the best performers on the Smart and Social scales.” However, it looks like dudes with beard+mustache are perceived smarter.

    PS – I have a beard and mustache.

  • Mike says:

    I think the charts would have been immensely more readable if you had clustered the “smart scale” (blue) together and the “social scale” (red) together.

  • Alex says:

    I noticed on the hair color graphs that you didn’t control for race. This worries me, since African-American and Asian people, among others, tend to be almost 100% black hair. Just sayin’.

  • strager says:

    Nice article, but it would be great if my eyes didn’t have to jump between the legend and the bars so much. The labels are short; just inline them into the graph and scrap the legend.

    Can you state the sample sizes, to at least a broad degree? (e.g. 50 Asians sampled, 300 males sampled, 30 red heads sampled)

  • Karen says:

    Hey, thanks for the info. Very interesting. I was also surprised by the hair. I noticed that there were NO grey haired women studied. I have predictions about that but interesting….

    More and more women, including, unsurprisingly myself, are letting their hair be natural as they age.

  • ia says:

    I noticed in the study that incluced “Middle Eastern” people only “Middle Eastern Males” were used, not “Middle Eastern Females” to discuss Ethnicity and Gender in Percieved Smartness and Extroversion. Were they just not included in the study at all? But their male counterparts were? They seems a bit wonky. As you have Latino, White, Black, Indian, Asian females — also, which “Middle Eastern” ethnicities did you include? I’d like to know that for the Latino, White, Black, and ‘Asia’ ethnicities, too. Such things can make a big impact on your study because there is such racial/ethnic diversities that exists under the umbrella ethnic terms you’ve used to label them as a whole.

  • Rajesh says:

    I just want to make sure if Indian means Native American rather than South Asian/APAC Indian.
    Is it right? Were they Native American?

  • I would also like to see a “misclassification rate” for each ethnic group. There’s probably a good number of people who look ambiguous between White/Latino/MiddleEastern.

    Gender misclassification would be less common, but no less interesting.

  • shauna says:

    For a thousand people of different races, I am guessing the sample size for redheads was too small

  • [...] A website polled Internet users on their perceptions of people’s intelligence based only on ph… It should be no surprise to learn that most people are racist. [...]

  • [...] morning… Large coffee in hand and reading What makes one appear smarter and more sociable? on judg.me blog, which I came across via Barking up the Wrong Tree blog (one of my [...]

  • Interesting data. But without errorbars, it can be very misleading.

  • Tony says:

    Have you tested your findings yet? It would be awesome to generate pictures that fall into all of the categories you tested and try to achieve the highest rating possible, or the lowest rating possible to prove your findings.

  • [...] Okay, so this is Take #2. Our earlier post on snap judgments was a little too “unrefined” and not as thorough as we would have liked, so [...]

  • Hammurabi says:

    Interesting Stats. My question is are the viewers picked from varying races and from different geographical setting or are they all say Americans? That will make a huge difference because if are showing an album to a group of Asians, you likely to get different data than if you show to a group of Latinos.

  • Nick Apuzzo says:

    Enjoyed this very much. Of course, the value of anything is in it’s usefulness, and I think it would be useful to find a way to ‘normalize’ the effects shown here. For example, affect of skill of persuasion or conversation (or maybe clothing style) versus both perceived social and intellectual score. Some way to show that for example, even though clean shaven is an advantage for a male interviewer, it might be a small influence when compared with another factor. Just a thought.

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>