It pays to be hot (sometimes)

There are some research questions you have to answer simply because the data is there: the salaries of Ontario university professors are published on the public sector salary disclosure website, research output is listed on Econlit, university websites and calendars list a person's rank - there are even databases that list every doctoral thesis published, so we can find out when a person received his or her doctorate, and thus estimate their age.

But most tempting of all, on ratemyprofessors.com, students rank their professors for clarity, helpfulness, easiness and “appearance (just for fun): __Hot __Not.” If a professor receives more hots than nots, a chili pepper appears beside his or her name.

In a recent working paper, Marcel Voia, Anindya Sen, and I use this information to estimate the effect of hotness on the pay and productivity of economics professors in Ontario.

There’s lots of evidence that, in general, good looking people earn higher salaries. But why would hotness pay in academia? A professor’s salary is largely determined by what he or she teaches (Engineering pays better than English), rank (full professors earn more than associates or assistants), where he or she teaches (University of Toronto pays better than Trent), and his or her research productivity.

Yes, being good looking might make it easier to get an academic job in the first place. But surely, once a person has an academic job, intellectual ability – the ability to do original research – matters more than looks?

Indeed we found that female professors rated as “hot” by their students earn no more – and might even earn less – than the “not.” We included controls for age and a host of other factors, so our results are not due to the fact that that younger women are simultaneously more junior and more likely to be rated hot.

Other research has found that women who are more demanding in negotiations are rated as less attractive. Perhaps the personality traits that prompt a woman to negotiate a higher salary just aren’t seen as sexy in the classroom.

For men, however, we found a startlingly high pay-off to hotness. True, not many male professors are rated as hot, but those that are, earn much more. Controlling for age, rank, university, and research productivity, hot male economics professors are almost 20 per cent more likely to be earning more than $100,000 a year. Conditional upon earning more than $100,000 a year, being hot increases a male professor’s predicted salary by more than $3,000.

We do not know exactly why hot male professors earn more. One possibility is hot men are those who are more assertive, and this leads to successful salary negotiations.

Another possible explanation comes from the comments students write on ratemyprofessors.com. “I absolutely loved Professor ____. He’s one of the busiest people in the Economics department and yet he never sees it as an inconvenience to do what he can to help,” or he’s a “a semi-retired superman.”

A key source of variation in academic salaries is the ability to obtain an outside offer. Hiring a senior academic, however, is a bit like buying a used car - one that is on the market is much more likely to be a lemon. Someone who is charismatic, likeable and well-organized - the traits that our hot professors seem to possess - may be more likely to convince a hiring committee that he's not a lemon. An outside offers story would be an alternative explanation of why we do not see a hotness premium for women - women are more likely to be in situations where they are unable to move because of family or other commitments.

When we estimated the returns to hotness, we controlled for research productivity - number of publications, citations, and holding a Social Sciences and Humanities Council research grant. Yet there might be further indirect effects of hotness if, for example, people who are fit and healthy are more likely to be hot and also are more productive researchers. We explored this possibility, but didn't find much - the relationships between hotness and research productivity were generally positive, but almost always statistically insignificant.

Where there is a stong relationship is (not surprisingly) between students' evaluations of a professor's hotness and their evaluations of his or her teaching. Hot male professors received significantly higher scores for clarity (0.76 on a 1 to 5 scale) and helpfulness (0.68 on a 1 to 5 scale) than their not-hot counterparts. There was very little difference, however, between the hot and the not in terms of students' ratings of the professor's 'easiness'.

But the pattern for women was quite different. Hot female professors had somewhat higher clarity scores, but the difference wasn't statistically significant. Where hot female professors received significantly higher evaluations was in terms of helpfulness (1.0 higher on a 1 to 5 scale) and "easiness" (0.43 on the same 1 to 5 scale - how I have struggled and failed to avoid the inevitable innuendos).

As a feminist, I find the results in one sense profoundly depressing. Men can have it all - be attractive and also well-paid. Women have a choice - you can go for it, be tough, negotiate a high salary, have high expectations of your students. But don't expect to see a chili pepper beside your name any time soon.

Note: this is an expanded version of the column that appears in the Globe and Mail's Economy Lab today. Please link, but don't reprint.

 

Comments

[Andrew F]

Sorry for the metacomment, but this is a good idea. If you find it challenging to write a piece to the word count permitted for the Economy Lab, maybe it's a good idea to post the fully developed post here. It's also a good way to drive traffic to the blog.

[tomslee]

In general I'm sympathetic to your feminist concerns, but "Men can have it all - be attractive and also well-paid"? Not so inspiring for those of us less than Brad Pitt-like in our looks is it?

[Just visiting from Macleans]

Was there any apparent evidence of people gaming the ratings system - ie a chili pepper appearing beside the prof's name where an ice cream cone would have appeared more appropriate?

[Bob Smith]

I wonder if the "perceived" (as opposed to purely physical) hotness of male professors is correlated with actual teaching ability in a way that the it isn't for female professors as a result of systematic differences between what men and woman perceive as being "hot".

Let's make the following simplifying assummptions:

(i) universities reward teaching ability (a heroic assumption, I know, but bear with me);
(ii) a professor's "hotness" is generally determined by members of the opposite sex (probably a realistic assumption amongst the heterosexual population. Obviously, it doesn't hold for homosexuals, but I'm not sure that matters for our purposes given the relatively small portion of homosexuals in the population at large); and
(iii) woman are less superficial in assessing "hotness" than men, so that good teaching ability (or at least attributes linked to good teaching ability) is "hot" for woman in a way it isn't for men (a plausible assumption, in light of the large body of literature out there suggesting that woman are, for a variety of evolutionary reasons, less superficial in terms of what they consider to be sexy then men).

So, if these assumptions hold, the good male professors are more likely to be rated as being "hot" by their female students, then the bad ones. In contrast, the "hotness" of female professors shouldn't be affected by their teaching ability because, in determining "hotness", male students are only interested in superficial physical attributes. So, if teaching is rewarded by universities, you'd expect to see a corelation between perceived hotness and actual teaching ability (and therefore income) for male professors but not for female professors.

Now the question is, how do I square this hypothesis, with your evidence as to the apparent link between perceived hotness and perceived teaching ability (measured by clarity, helpfulness or easyness) at ratemyproffesor. My first observation is that, in terms of assessing teaching ability, universities are less likely to be impressed by "easiness", so maybe that isn't a great indicator of teaching quality. With respect to male professors, it seems that the "hot" male professors are considered to be better teachers than their uglier counterparts, which seems to jive with my theory. With, female professors, the clarity result is consistent with my hypothesis, while the helpfulness result isn't.

However, I wonder if there isn't a further problem with the dataset. While, for woman, teaching ability increases perceived hotness (or so I'm suggesting), what if it works the other way for men, what if men (particularly the shaved apes that are the 18-24 year old males of the species) are more likely to assess a "hot" female professor as being a good one (I realize I'm being awfully uncharitably to my fellow men, but come on guys, we all know its true). In both cases, there'd be a correlation between perceived "hotness" and perceived teaching ability, but in only one case (for men) would perceived hotness be correlated with actual teaching ability.

I wonder what would happen if you compared perceved hotness (rated by members of the opposite sex) against teaching ability as determined by members of the same sex (so, ignoring our homosexual friends, less likely to be influenced by perceived hotness). I don't recall if ratemyprofessor indicates the gender of the raters, but that might be one way of identifying actual teaching ability (assuming that hotness doesn't matter when evaluated members of the same sex).

Anyhow, interesting question.

[Rachel]

A prof in disability studies told me that disabled people who are physically attractive get substantially better care. So physical attractiveness has ramifications in many areas.

[Frances Woolley]

Tomslee - actually some of the male profs with chili peppers aren't, to my mind, particularly good looking. But most of them are the kind of people who you would like to be sat next to at a dinner party.

Bob Smith - "woman are less superficial in assessing "hotness" than men, so that good teaching ability (or at least attributes linked to good teaching ability) is "hot" for woman [i.e. female students] in a way it isn't for men [male students]"

Yes, I think this observation gets at the core of what's driving our results. But it isn't that for male profs good teaching = hotness = high pay. There are no financial rewards to good teaching - none, zip, nada. If anything, one might expect a negative relationship - good teaching skills = hotness = being hired by a teaching-oriented university, being assigned to large undergraduate classes.

I think what's happening is that for men, assertiveness/drive/ambition/charisma/likeability/je ne sais quo=hotness=ability to negotiate a higher salary.

And also for men assertiveness/drive/charisma/likeability=hotness=clarity and helpfulness.

[Nick Rowe]

And as a committed masculinist, I say it just confirms that men are treated as sex objects. Women are valued for who they are; men are valued for their hotness. Someone should write a book about this.

And I'm really pissed I don't get a chili!

This is good stuff Frances.

[Nick Rowe]

Actually, I think I agree mostly with Bob Smith.

[JeffJardine]

FW:"I think what's happening is that for men, assertiveness/drive/ambition/charisma/likeability/je ne sais quo=hotness=ability to negotiate a higher salary."

Might it be even simpler than this? Maybe when it comes to female students evaluating male profs, affluence=hotness.

[Just visiting from Macleans]

Say, for all you bearded profs who didn't make the chili list, an anecdote that will probably freak out MM even further:

I saw his latest twitter pic, and I immediately recognized the name. The individual in question taught me a couple of applied math courses in the early 80's (LaPlace transformations was one). And he was actually quite good. And I did pretty well in his course as a result.

He used to walk home after teaching along George St where he owned a rental property that my gf at the time rented with 4 others. When he saw that I had picked up a junked '69 VW, parked in the driveway, and was using it to scavenge parts for the same VW I was repairing, he immediately insisted that it be removed - my previous top standing in his class held no weighting (not that he would have remembered me, necessarily). So, I stalled for another three weeks until I had harvested all the good parts. When he returned even more irritated, I got out the cutting torch, and went at it.

I remember him vividly from the classroom. So, it's not at all surprising to me that he was acknowledged with a permanent memorial. In an undergrad class of mostly males, he would have scored low on the "chili" scale.

(You can delete this after MM sees it)

[Frances Woolley]

Jeff, "affluence=hotness" - do you realize that this equation destroys the entire basis of our argument as it suggests that salary causes hotness rather than hotness causing salary?

I am somewhat worried about this, but my concern isn't the one that you're probably thinking about - I know far too many academics with household incomes in the 1/4 million dollar range who dress in scruffy old jeans, brown bag it to work, and drive crummy old cars (remember an academic's overall affluence is largely determined by his or her spouse's income).

This is the story that I would tell if I was trying to destroy my own paper:

Finance/money/macro guys wear suits and ties - because that's how finance people dress. They also teach really cool stuff. Both of these things mean that they're more likely to be rated as hot. Coincidentally, finance people also are highly paid. So all we're picking up is some kind of field effect.

I don't think that's what's going on - in our data set we have hot labour economists, hot public finance economists, as well as hot money/macro guys. But it might explain part of it.

[David]

I think the comment I was going to make is embedded in your paper already: "Our results are not a pure measure of the effect of 'beauty' on either academic salaries or professional productivity. Hotness measures some combination of physical attractiveness and other personality traits. It is possible that some people’s hotness scores come more from their good looks, other people are hot because they are charismatic or likeable (and probably not bad looking)."

It could be that men are labelled hot more for different reasons than women, and that's why there are different effects of hotness. It is also interesting that females are roughly twice as likely to be labelled as hot than males. Perhaps students set the bar lower when judging female hotness, and if they were as discriminating in judging female hotness as they were in judging male hotness, you'd see a similar effect on earnings.

Finally, it's a little bit disturbing that you find there are no financial rewards for good teaching. I get that that's probably the reality at the more prestigious departments (e.g. Toronto, Queen's, Western), but your sample is 16 Ontario universities, so it's including the smaller schools where teaching is probably emphasized over research (e.g. Trent, Nipissing, Laurier). I wonder what the teaching-oriented universities are rewarding if they're not rewarding teaching…

[Frances Woolley]

David - "I wonder what the teaching-oriented universities are rewarding if they're not rewarding teaching…"

First, faculty at Trent, Nipissing and Laurier would probably point out that they do research too, and that their tenure/promotion etc depends largely on their research performance. Does the world actually need to have thousands and thousands of academic economists churning out research? I'd actually rather not ask that question, because I have a pretty good idea of what the answer is likely to be, and personally I really enjoy doing research.

But your point was about universities where teaching is relatively more important...

To some extent, salaries follow a model of "intercept negotiable, slope fixed." Once you're hired, pay increases are largely mechanically determined. Toronto has merit pay, so has Waterloo, I don't remember off-hand which other schools have it, but it's not that many (we control for the presence of merit pay in our regressions).

One exception to everybody gets the same % increase rule is if you are able to negotiate an adjustment, usually in recognition of an outside offer. The other exception is that sometimes one will get a pay adjustment upon promotion - at Carleton this has only been important for people hired in the early 90s when salaries were very low.

Nick, I think, may do a post on incentives in academia some time soon.

On why women are more likely to be labelled as hot - they're also younger, because there weren't women hired in substantial numbers until the late 80s/early 90s.

[Bob Smith]

"There are no financial rewards to good teaching - none, zip, nada. If anything, one might expect a negative relationship - good teaching skills = hotness = being hired by a teaching-oriented university, being assigned to large undergraduate classes. "

Frances, I'm just curious about the basis for this statement. Is that a general observation, or is that something that comes out in your dataset. If the later, I'd suggest that some of the measures of "teaching ability" on ratemyprofessor (such as easiness) are not particularly meaningful measures of teaching ability (I doubt universities are inclined to reward professors for easiness).

If the former, that I have to echo David's comments, what exactly are they rewarding at places like Trent and Nippising if not teaching? (And, perhaps that's a bigger question.

I'm not sure I buy the negative relationship argument though between teaching ability and pay, given that you're controlling for other aspects like research productivity. Also one might think that teaching abilities might correlate with other unobservable, but desirable traits (if you can explain Legrange multipliers to a gaggle of slack jawed 20-somethings, you might also be able to explain what your department does to donors, university administrators, etc. Similarly, a good teaching professor who can induce first-year students to major in economics rather than to become... shudder... sociologists, might be worth hanging onto if the retention of those students gives the department added heft in inter-departmental funding battles).

[happyjuggler0]

If, as Bob Smith suggests, that hotness is being quantified differently by male and female students, i.e. on more than physical attributes when women are doing the grading, then this seems like a great time to do an actual experiment.

Gather photos of rated professors, male and female, and then go to a shopping mall and ask random passersby of college student age for a few minutes of their time to rate photos of members of the opposite sex (or both sexes for that matter).

After gathering a statistically significant amount of data, tally up the responses and then compare them to the actual students rankings. If there is a divergence between student ranking and mall rankings in the hotness of men but not the women then Bob Smith is on to something....

[Frances Woolley]

Bob, o.k., well maybe I exaggerated a little bit, this is what we report in the paper:

"However being a good quality professor has few monetary returns. In the regression reported as case 7 in Table 1, the coefficients on ease, helpfulness and clarity are all positive but insignificant: easiness $2534 (standard error (s.e.) $1445), helpfulness $813 (s.e. $2043), clarity $814 (s.e. $1905)." That's compared to around $17,000 for being hot. And depending upon exactly how we do things, those coefficients move around, sometimes ease and helpfulness come out negative.

In 20 years of teaching, no one has ever come into my classroom to observe my teaching. No one ever looks at my final exams to see if they're not totally crazy. The only measure of teaching quality that is used at Carleton is students' in-class evaluations of professors. This is absolutely consistent with what is done in just about every university in Canada. And studies that have compared official teaching evaluations with ratemyprofessors evaluations find that they correlate pretty well.

There's another paper that I've done with Casey Warman and Chris Worswick that looks at academic salaries in Canada (to go scholar.google.com, type warman woolley worswick and click on the working paper version, it's ungated). That paper has a pretty good literature survey on most of the research on academic salaries in Canada. The 'intercept negotiable slope fixed' is something that has been found in other studies, but I can't recall the references off-hand.

happyjuggler0 - yes, we've talked about doing something like that a bunch of times, but we're lazy. And actually getting good photos of people is not as easy as you might think - e.g. the Brock economics department website is terrible, and has no photos of anyone.

[David]

We're getting off the main topic here, but it would be interesting to classify schools into "research-focused" and "teaching-focused" (yes, it's arbitrary, but we could do it). I'd be interested to see if the returns to teaching are relatively higher (as a percentage of mean salary) at teaching-oriented schools.

I realize research is valued pretty much everywhere, but the smaller schools generally don't attract top researchers. Given the minimal research output of many of those profs, I would have figured it would be hard to differentiate them mainly based on research alone.

I'm also surprised that the norm is a slope-fixed salary progression. I did my undergrad at UVic, where the government mandates that salaries of $75,000 be publicly available (http://www.vancouversun.com/business/public-sector-salaries/advanced.html). Looking at econ profs there, you can see that the wage increases between 2008-09 and 2009-10 as a percentage of salary are quite different from prof to prof (some people experience wage decreases, while some get increases as high as 10%), which doesn't lend itself to your intercept negotiable, slope fixed argument. I know at McMaster, the chair was saying that he had a fixed salary budget, so if he wanted to reward top performers with bigger wage increases, it meant a smaller wage increase for someone else. The fixed slope thing just doesn't make sense for professions like academia and pro sports, where performance is relatively easily observed.

[Blikktheterrible]

I'm interested in how this reflects on the different traits that are seen as attractive in men and women. Clarity seems to be viewed as attractive in males, while females are seen as attractive when they're "easy." The latter seems pretty self-explanatory to me. When it comes to the former, my comment would be that females are more likely to care what a potential mate says than males are. Males are more likely to ignore what a potential mate says and focus only on looks, pleasantness, and whether she seems interested. Females are more likely to care if a mate is dumb or boring.

"On why women are more likely to be labelled as hot - they're also younger, because there weren't women hired in substantial numbers until the late 80s/early 90s."

I feel like this trend could reverse. My own perception is that females are more likely to find professors attractive than males are. Basically, males tend to be age-phobic, whereas females are more likely to find intelligence and status (which professors could be seen as possessing) attractive.

[Blikktheterrible]

"As a feminist, I find the results in one sense profoundly depressing. Men can have it all - be attractive and also well-paid. Women have a choice - you can go for it, be tough, negotiate a high salary, have high expectations of your students. But don't expect to see a chili pepper beside your name any time soon."

That's an interesting reading of the evidence. Your theory is that this is largely about assertiveness, right? Wages are being driven by assertiveness, which is attractive in men but not women. That could be true, but I think it's plausible that other forces are just as powerful in all of this. For instance, it could be that intelligence is the driving factor. Intelligent people make more money. Intelligent men are seen as attractive, whereas intelligent women are not (necessarily). It's not really a pro-feminist interpretation either, is it? At least I wouldn't interpret it as penalizing women. Intelligent women just aren't reaping the same level of benefits as intelligent men, in terms of attractiveness. On the bright side, my interpretation makes it entirely possible that professors are being paid according to their abilities. That doesn't sound right, does it?

[Frances Woolley]

David, yes, "intercept negotiable, slope fixed" is a rough approximation. U Vic has merit pay - actually most of the larger universities out West (SFU, UBC, U Calgary) do. Also remember that any big year on year adjustment is almost certainly someone coming on or off leave (we have a variable "sabbatical" to control for that) or in and out of an administrative position (we drop the university level administrators from our sample).

What you say about McMaster is true of most universities with merit pay, which limits the scope for major salary adjustments through merit. An academic might be in the same department with the same colleagues for 10, 20, 30, years. In this context, keeping people happy and maintaining some degree of equity is important.

If you really want the comprehensive analysis with some totally ridiculous number of rainbow spagetti graphs, you can check out this other paper I wrote on the structure of academic salaries in canada (it's published in the Canadian Journal of Economics but I've linked to an ungated version).

What that paper shows is that there's pretty significant cohort effects in academia - so people (especially men) who were born in the late 1950s or early 1960s have a pretty different life-time earning trajectory from people born in the late 40s or early 1940s.

It takes a lot of merit increments to make up for getting a really lousy initial starting salary because you were hired in a year when it was really tough to get jobs.

Blikktheterrible - it's very hard to talk about this stuff without starting to sound really sexist, isn't it? But, yes, that men and women find different things attractive is, I think, what's driving our results. I don't think it's so much intelligence given that most academic economists are pretty smart by conventional measures anyways - as something like creativity, empathy, ...

[brian]

For men, you have the causation backwards. Male 'Hotness' is a function of status and your findings support this. Male status is attractive, the teachers who are more successful at teaching and research, and have the social skills to achieve that success are viewed as more attractive because they have higher status. It's not attractive (high status) leads to higher pay, it's higher pay (high status) leads to being attractive. Brad Pitt may be very physically attractive, but he is a sex symbol because he is a star (high status), if he was working for the local plumbing company he would not have the same allure (low status).

[Frances Woolley]

Brian, I disagree.

Can we agree that the broader social status of a professor depends upon his/her research and income?

Hotness and research output are pretty weakly correlated.

Hotness is related to a male academic's salary. But remember: salary is only one part of income. I don't have exact figures, but from my experience of working in academia for a number of years I would be willing to lay a good deal of money on the fact that consulting income, income from textbook royalties, inheritance, investment income and earnings of other household members explain much more of the variance in university professors' incomes than hotness does. So social status=income=hotness does not mean that we will necessarily find salary from main job=hotness.

Now if you're saying that there's something unmeasurable - ambition, drive, charisma, etc - that causes people to be both hot and well-paid then yes, I agree with you.

But that's not a matter of having the causation backwards, that's saying hotness is a proxy for some other unmeasured variables.

[Just visiting from Macleans]

FW - here's why I can't figure out why it would be advantageous to be "hot" in academia. Why would those that determine one's compensation necessarily pay more for a "hot" prof of either gender? If "hotness" was valued, wouldn't you want to not promote "hot" people - a threat to your status?

[Frances Woolley]

Just visiting - I suspect it's probably outside offers.

[Just visiting from Macleans]

Ahhh, well, you'll eventually be replaced with holograms for the larger general classes anyway. :)

[Manuel]

How do you know that male hotness drive salaries instead of high income driving perceptions of male hotness? This is another version of the causation backwards argument, but focused on the money itself, not on unmeasurable variables that facilitate high-income. I know it's a terribly un-PC hypothesis, but well backed by popular culture ("diamonds are a girl's best friends" and all that).

[Blikktheterrible]

I think the best argument against that interpretation is that the students probably don't know how much each professor is being paid.