The other day I came across a blog post ranking academic fields by hotness. Important data for sure. But something about it was gnawing on me for a while, some connection I wasn’t quite making.
And then it hit me. The rankings looked an awful lot like another list I’d once seen of academic fields ranked by intelligence. Only, you know, upside-down.
Sure enough, when I ran the correlation among the fields that appear on both lists, it came out at r = -.45.
I don’t know what this means, but it seems important. Maybe a mathematician or computer scientist can help me understand it.
104 thoughts on “The hotness-IQ tradeoff in academia”
Looks like another instance of the compensatory relationship between perceived warmth and perceived competence
Click to access Kervyn%20et%20al.%20JPSP%202009.pdf
Are you implying that whatever we perceive may not be true? What exactly did you perceive to come to this conclusion?
perhaps just a reflection of the dumbing down of society?
you are exactly right…in comparison to MANY other countries, we are a very “dumbed down” society, and one political party in particular loves that it is this way, and is doing everything in their power to not only keep it that way, but to get us as dumb as possible, so we just follow like sheep, in blind faith. I lived in a few different European countries for a total of 14 years, and was shocked when i came back to America….sad, very sad for “the greatest country in the world”…
Astra I have heard this sentiment before and I don’t doubt much of its truth but I also have to wonder if some portion of this belief is due the comparatively low barrier to entry to American society. The ability to “hang out a shingle and make a million bucks” is not a quality typically attributed to many European societies, including to the present day.
What some perceive as numskull noise is instead the by-product of a much more inclusive society, where the opinions and works of many are heard and perhaps respected on their merits, not the perceived status of the individual.
Come on, there’s no need to single out the Democratic Party like that.
Like your blog a lot. One comment though — looks like you didn’t utilize all the potential data points in the scatterplot — or am i wrong?
Some fields didn’t appear in both lists — I only plotted fields that appeared in both places so there are some skipped ranks (I calculated the rankings before I deleted incomplete cases). Also, I consolidated some data, e.g., one list had a single entry for “Engineering” while the other broke them out by subfield, so I grouped together the largest cluster of engineering disciplines.
I am obviously making no claims about intelligence and gender, but a likely explanation is that “hotness” for a field correlates with gender ratio.
An intriguing idea – please explain?
That looks like a plausible explanation of the data. Unfortunately, as a philosopher I can tell you that our discipline is one in which women are, notoriously, greatly underrepresented. The most recent data indicate that only a bit more than 16% of post-secondary academic positions in philosophy were held by women.
even more dismal than the % of lady cavers…
You’re making the mighty big presumption that the “hotness” of a field was ranked by heterosexual men. Or perhaps homosexual women. Hotness isn’t automatically associated with feminine beauty, I’ll have you know! *shakes fist*
Yes, very good point. But surely the gender disparity in philosophy raises some questions about the hotness data. Do more women than men report hotness? What about the hotness of people in other fields with greater numbers of women (or more gender balance, even) amongst the raters?
Also, as flattering to me as the hotness value of philosophy is, I’d like to point out that as the Bluffer’s Guide to Philosophy puts it with some accuracy: in terms of appearance philosophers can be divided into two groups. First is the Nietzschean superman, fit and sartorially sublime. The other is the human wreck. So I suspect at least some serious sampling errors here.
Look at me, taking this silliness seriously.
I don’t think it’s a big presumption at all. I’m a female “hard-science” academic, I hear straight male individuals ranking the hotness of their colleagues infinitely more than female or LGBTQ individuals (i.e., they do it publicly at all).
We’re much too concerned with gender bias and being taken seriously as women in our fields to worry about the hotness of the males in our department.
Eyeballing this chart, it seems there is no clear relationship between IQ/hotness and gender ratio. If you cross-reference between that chart and the one provided here, it’s all over the place.
And the philosophers win!
I agree with Jennifer, philos seem to be the the smartest among the hottest.
Indeed, I would not care about the corr. line. I would just look at the two groups (above and below twenty on the hotness rank)
What’s the p value? The r value doesn’t count for anything if the p value isn’t persuasive. (It looks like it will be, but you should report it.)
Welcome to the blog! Have a look around.
A p-value is only an appropriate statistic in the context of sampling and inference. The disciplines here aren’t a random sample, and there’s no inference going on.
Steve, if we’re interested in this specific population of disciplines as characterized by these specific studies only, that’s correct, but since there’s a larger population of disciplines from which this sample is drawn, and subdisciplines with potential differences in intelligence from which we might wish to predict hotness, it would be good to have a p value.
I’ve reconstructed the chart and get p = .03 using Spearman’s rho. R’s a little different as the numbers are invented: Pearson’s r = .42. Still, it’s not clear to me what can be inferred from this because the methodologies of the original studies would need to be evaluated. Also, some data points seem to be missing.
in correlation, significance level isn’t particularly informative (not much of a difference of .07 and .04…just the way the formula works out). But, because r is an effect size in-and-of itself, it can be more informative than the p.
See point #3: http://www.methodspace.com/profiles/blogs/bonferroni-correcting-lots-of-correlations
As for what this “means,” it means that 20% of the variance between attractiveness and IQ can account for each other.
I would venture, that if you plot the rankings in this data with mean income of the field they would also inversely correlate! There likely would be some outliers though..
This is really confusing, who rates chicks on a scale from 0 to 40??????? Use the standard 1-10 pls. Anyway, I knew there’s a reason I never went for any college chicks
You must be gorgeous.
Very nearly spit all over my keyboard. Also, for the record, I majored in philosophy so…boo yah.
Isn’t it a statistical no-no to run Pearson correlations on rank-order data?
I’ve never heard that. Spearman rho is just a Pearson r on ranks, so technically I could have said “rho” but that seems like something an economist would do.
Right, but you should rank each variable for the available pairs (i.e., rerank after you dropped some observations). See http://en.wikipedia.org/wiki/Spearman_rank_correlation or http://www.economicsnetwork.ac.uk/statistics/stats_lec5.pdf for a nice explanation. Further, even if your graph was correct, you should say whether the correlation is significantly different from zero. Juts eyeballing your graph I would doubt that…
I was wondering about the same thing…i would love to see more statistical details such as significance or correlation coefficient on these data
Whoa, hits are spiking. Welcome new readers. Where’s everybody coming from?
Teh interwebz. But seriously, the article has created some conversational buzz in my news feed, flippant or not.
You’re going viral. Why not just use the numerical values instead of the ranks, btw?
Crazy. My wordpress stats page shows that a lot of hits are coming in from Facebook, but I can’t tell from where — like if it was linked from a major “page” or just people sharing in their personal feeds.
I used ranks because it was quick. I was too lazy to try to extract and enter numerical values (especially from the “hotness” graph). And because I wasn’t taking this seriously. I would like to believe that nobody else is either.
As to the hits, I think mostly from personal news feeds, mostly from grad students and other young researchers who have to be on facebook while their regressions and simulations are running, their code is compiling or their test tubes are, you know, doing whatever it is test tubes do…
Physics professor at Union just shared this via FB, FWIW…
And no, at least here, we’re not taking it *too* seriously… with jokes that Physics should be at the upper right :)
I blame Christopher Chabris, who brought me here. I can only assume the “legal” academics field is so far off the chart to the “northeast” that they just wouldn’t fit…
An economist from LvMI just shared this.
Since the ratings of hotness came from the highly reliable and construct-valid chilli pepper on “RateMyProfessors.com”, I am concerned about the impact of graduate student teaching. In many fields, it is not uncommon that graduate students and/or adjuncts are teaching the larger undergraduate courses…Perhaps all the hot Chemists are simply on course buyout as a result of large grants and these un-hot ratings represent the desperate and possibly semi-poverty life of a graduate student who is teaching the course for the first time…Another possibility is that since chemistry as a discipline is nearly incomprehensible, the students nearly all score a 13% on the exam and so view the (otherwise hot) chemistry professor through the hate-filled eyes of an undergrad that will not be getting into med school (where all the moderately hot and yet disappointingly slow-witted med school profs would be teaching them)…
In my grad school experience, the grad students dressed way better than the faculty when they taught. In part this was so our students would take us seriously rather than treat us like their buddies, but also because we were young enough to be just hanging on to the style trends. And now, since the thrift-store look is oh-so-hip, “semi-poverty” has never been hotter.
But that was just my philosophy department at a big state u. Might be very different elsewhere.
Got here from The Monkey Cage: http://themonkeycage.org/2013/09/19/political-scientists-are-pretty-smart-given-how-incredibly-hot-we-are/
But what if I am in Math Education? Now I am totally confused. (BTW, your post showed up in my twitter feed, and I am totally stealing this post for use in my AP Stats class.)
Reblogged this on and commented:
I studied BioChemistry.
Ya, found this from a professor’s facebook post. Someone had already reposted from his and I’m sure he was reposting. I’d guess you’re getting a lot of academic facebook traffic during office hours. That’s why I’m here, anyway.
Notice that it’s not or only close to significance.
sanjay – not surprisingly, this looks popular among my philosophy fb friends. i had nothing to do with it — i get enough upward social comparison at home as it is.
I’m sure the philosophers are applying their usual rigorous standards for empirical data.
We already knew the results of this chart a priori. The confirmation is nice though!
We certainly are. Extraordinary claims require extraordinary evidence. Conversely, this study only warrants a cursory “check and nod,” as previously mentioned.
*Philosopher walks in, checks results, nods, and leaves*
Hmm, so does the combination of my engineering, economics, and education degrees place me in the upper right, bottom left, or dead center? (I got here via a friend’s FB post)
depends what school of economic though :p
not sure why you (conversationally) imply in this post that you constructed the graph from two separate graphs you’d seen when that graph is simply taken from the second link you provide … an attempt to be smarter or hotter than you are? :)
I conversationally imply that I made the graph because I made the graph. The author of the linked blog appears to have updated his post to include my graph without attribution (but if you click on the graph in his blog post it will take you back to my site).
so then i guess that’s a question for him or her, … and a sign of my own laziness, or worse. sorry about that.
No worries, though a few more mistakes like that and you might need to change your nom de guerre.
What counts as medicine. Im biased but as a med student I find it hard to believe that all (most) of these groups, like business rank above in IQ.
I may just be a philosophy professor, but I have to admit that I find this data totally convincing!
My two cents: medicine should be way higher. And mathematics too.
Where is Physics? Off the chart?
your craizy if u think educationers r the most dumbest.
This Factor Analysis (looking for correlation between “intelligence” and “hotness” in academic disciplines) has a negative trend. The easiest explanation of which is that the majority of people prefer subjects they understand more easily, hence the negative slope correlation to IQ. However the analysis is probably faulty in that Intelligence appears skewed definitionally towards SQ (Systematizing Quotient) and not factoring (Empathising Quotient).
Click to access Billington.pdf
This is so awesome. I love it. Makes me feel good being a writer…. ;)
p-values? (run a quick t-test)
the hottness ranks came from rate my prof, maybe that just means the higher the students IQ, the less they bother to rate their teachers “hottness”. It also means that all of those subjects ranked for Hottness are in the education field…
Can you post the raw data so I can do my own tests?
Sorry, I cannot abet any attempt to take this seriously, much less conduct hypothesis testing on it. You’re on your own.
This shows intelligence in adolescence and later choice of subjects, which I think is relevant to your enquiry.
Can you clarify something since you’re catching some flack on the internet, specifically on Twitter? Did you regress the two rankings, or just find the correlation between them?
I’m defending you against someone who is say that if you were smarter you would know better than to regress two sets of rankings. I countered that you were calculating the correlation.
Side note: I’m an economist.
I was planning to convert both variables to z-scores and run a regression, but an economist told me to do a correlation instead.
You’re missing the bigger problem though, which is that this analysis does not meet the can opener assumption.
I was just about to suggest that a multilevel model was called for, but then you raised that dang can opener assumption and so I’m crawling back into my hole.
I can add that the political science crew have been quite pleased with both surveys and putting them together just made us giddy. A link is on Monkey Cage so expect more views…
If I recall correctly, it was either Socrates or Plato that claimed that smart people are ugly and that hot people are stupid. (Not seen any real data on it until now.) So, the negative correlation may just be an artifact of that general, individual tradeoff added to the fact that fields often sort by IQ to some degree (especially given that IQ tests one type of intelligence).
Though that does leave philosophy as a very interesting outlier…
I wonder – how much of this might be determined by the gender mix for those taking classes v those teaching the classes? Culturally, it seems acceptable for women to call men or women attractive, or for men to call women attractive. Men calling men attractive still seems to carry a social taboo (at least in the part of the US I’m in). So, if we’re in a field that has been male dominated, and continues to be male dominated, that would decrease the hotness rating… Feels like a believable story for the upper left corner.
(Full disclosure: Economist here.)
I’m with you, Lucas. Since the so-called (nod to Sanjay) hard sciences tend to be male dominated regarding the gender of both the professor and the student, I doubt that many males would even rate the dudes from the Twilight series as “hot”. Nor would I for that matter because vampires scare me.
For better or worse (worse, probably) I followed the links to the working paper using the hotness rankings. Hotness data are from ratemyprofessor.com. That website averages the number of “hot” and “not-hot” votes for an individual professor, then converts it into a number of chilis. People who get more “not-hot” votes, however, score 0 chilis rather than -numbers of chilis. I’m pretty sure these data don’t warrant more sophisticated statistical analyses. (Not even going to talk about the validity of that data.)
FYI: I am now planning to use this as an example in my (sociology! “Hot but Dumb”) research class, as an exercise in critically reading popular reports of Science.
The correlation I believe originates from the amount of time one is able to spend on studying based on their (lack of) social life, which is correlated to their hotness. I.e. less hot = more studied. This is also why glasses are correlated with intelligence and lack of social ability. Put on a pair of glasses and you gain 10 perceived IQ points and lose 10 hotness points.
I wonder why economics is so far away from religion.
I bet everyone debating statistics here is low on the hot scale.
Are your axes backwards? And if so, I’m surprised medicine’s IQ came out so low.
I am so confused by the numbers on the axes. Would someone bother to explain?
I am surprised so many people think Medicine should be higher IQ rank. As largely a practical discipline it does not involve deep contemplative thought or unique problem solving but memory and repetition. Both of which are low IQ activities. The average doctor is only supposed to have an IQ <110. This is still within one SD of average.
Amazing how you actually believe IQ means intelligence. Do you realize that the famous mathematician and physicist Henri Poincare scored quite low on an IQ test, but was able to revolutionize many fields within mathematics and philosophy – so greatly, in fact, that he is remembered as one of the ten greatest mathematicians of all time. My point here is that IQ does not measure intelligence to the degree that we believe it does. It generally tests the velocity that one reaches an answer, not the overall potential of the individual with regard to their mental talent. I can prove this quite easily. For instance, suppose I give two men ten questions each of the same difficulty. One man answers the questions right in a shorter period of time, while the other man takes longer but still reaches the same conclusion. But as I administer more questions, each with increasing difficulty, I notice that the man who answered the questions quickly is now stumbling over them, while the other man who answered the questions slowly is still able to answer these questions, regardless of difficulty. My point is that intelligence is not a “race” to see who can answer complex questions the fastest. It is the ability to solve complex questions regardless of their difficulty and regardless of the time it takes. Do you think Einstein was racing to answer the theory of relativity? True intelligence takes time and contemplation and intense meditation. It is not a race to calculate the probability of pulling a red ball instead of a blue ball. Intelligence is innovation. Intelligence is creativity.
You have no clue what IQ is.
I’m wondering about the GRE scores on which the IQ measure is based. Hsu’s source for V-R-A scores is a table on an Econ site (http://web.archive.org/web/20070104093613/http://www.econphd.net/guide.htm) which relies on “data reported in ETS Powerprep 2002 software” (see footnote). Would those data be subject test scores, general test scores reported to PhD programs, scores based on students’ reported/intended field of study, or based on or some other measure? It’s not clear how these were compiled.
I’ll call your tradeoff and raise you promiscuity.
Imagine the innuendos possible with Philosophy, Religion, Poli-Sci, English, and Biology.
In the film Men in Black II, Will Smith’s character uses the term “hotness” to describe the new, better, faster, more desirable automobile. It seems to be a non-gender specific term to me. That is unless you feel deeply that cars have gender.
I enjoy reading how everyone gives a damn about intelligence. Last I checked, none of you split the atom. And how you’re all so “grammatically correct,” as if the more “grammatically correct” you are, the more intelligent you’ll be perceived. Here’s a hint to everyone: no one cares. Human beings are all a bunch of egotistical fools, trying to obtain “admiration” and “bragging rites.” You’re lucky if you live 80 years on this planet. Quit caring so much as if it even matters you dumb twits.
So what you’re saying is, you think we’re all kinda hot.
In a really awkward way, yes. You’re making me blush hot stuff.
This shouldn’t be plotted by ranking, but rather by the actual values. Here is the plot by dimensionless intelligence or attractiveness.
As you can see, the trend remains, but the distances are exaggerated and the fit is much worse.
“the fit is much worse”
I’m just a lowly psychologist but even I know that if you are reporting R^2 and I am reporting r, we need to multiply your number times 2 in order to compare them.
This is b.s
Engineering at not more attractive than say a girl in communications sociology education and especially psychology
This hotness criteria is biased towards disciplines with an even number of men and women. If 90% of the students are heterosexual males then 90% of their male professors are very unlikely to receive a high “hotness score”, similarly if 90% are heterosexual females…
You should ask a philosopher, look at the chart man!
Looks like this dataset suffers from heteroskedasticity. That is, your errors aren’t constantly distributed. Note the four clusters of datapoints in (roughly) each corner of the chart. The grouping of “hard skills” majors (math, econ, chemistry, etc.) and the grouping of “soft skills” (history, religion, philosophy) are both fairly close to your line.
However the two groupings below the line (I’ll say “professions” on the left and “other soft skills” on the right) are all roughly at the same IQ, but different hotness. So your model gets less good at predicting uglier people’s IQs.
What does this mean? There are ways to filter out it, but a simple interpretation is that the “uglier=smarter” effect works for some college majors but not others.
There is no way political science people are smarter than historians. They may be hotter, in a polished CNN anchor kind of way.
R -0.45. This essentially means a one point increase in intelligence causes a twofold drop in hotness. Lol. I beg to differ. Some of the hottest girls i ever met were on my math degree. AND it was tough!
Comments are closed.