U.S. college majors: Median yearly earnings vs. gender ratio

Last year, I looked at the gender ratios across college majors and discovered an interesting-yet-spurious correlation: College majors with higher male:female ratios (i.e., with more men than women) tend to have students with higher estimated IQs. After much debate, the correlation seemed to be explained by the fact that male-dominated majors tend to be more quantitative in nature, and the IQ estimation procedure relied heavily on students’ quantitative SAT score. Thus, we were only observing a gender preference for quantitative vs. non-quantitative majors.

Recently, I ran across an interesting FiveThirtyEight article that analyzed the estimated median earnings for recent college graduates broken down by major. The data that they used from the American Community Survey was publicly available on their GitHub, so I decided to take a closer look at the correlation between estimated median earnings and the gender ratio for each majora.

Below, each square is a college majorb scaled by the number of recent college graduatesc. The squares are colored by the general category of each major, and I annotated a handful of the most popular majors. If you’d like to look up specific majors, use FiveThirtyEight’s lookup tables.

us-college-majors-income-vs-gender-ratio-ann

The trend that’s immediately apparent from this chart is that female-dominated majors make less on average than male-dominated majors. Some interesting exceptions to the trend are Nursing (90% women; $48k median earnings) and Transportation Science (12% women; $35k median earnings), where Nursing especially stands out as a relatively lucrative major despite being primarily women.

Despite the outliers, we’re still left to wonder: Is it really true that women earn less than men — even for college graduates? I decided to investigate this issue a little further. Of course, the first step is to make sure that our eyes aren’t fooling us, so I fit a linear regression onto the data and weighted the majors by their number of recent graduates.

us-college-majors-income-vs-gender-ratio-regression

Sure enough, there’s a clear and significant negative correlation between a college major’s median year earnings and gender ratio. Let’s try to find out why.

Why do female-dominated majors earn less?

If you look through the list of female-dominated majors, you’ll notice that many of them can be quite difficult to find a job with because they’re so competitive: There are far more graduates than jobs in many of the fields. Could female-dominated majors be earning less because they couldn’t find a job after college? The unemployment rate for each major was included in the data set, so let’s take a look at that correlation below.

us-college-majors-unemployment-vs-gender-ratio-regression

For the most part, this idea seems to be busted: There is no significant correlation between the gender ratios and unemployment rates of college majors, and most majors are sitting around the national unemployment rate of 5.5%.

What about underemployment? In his FiveThirtyEight article, Ben Casselman wrote that drama majors (which tend to be female more often than male) are more likely to end up waiting tables than using their degree. Underemployment rate was also available for each major, so let’s take a look.

us-college-majors-underemployment-vs-gender-ratio-regression

Yet again, our idea has been busted: There is only a weak correlation between the gender ratios and underemployment rates of college majors, so the typical narrative of useless college degrees doesn’t seem to explain why male-dominated majors earn more than female-dominated ones.

I was a bit lost at this point until I remembered my controversial post from last year.

Back to an old study…

If you read my blog last year (or read the introduction to this post!), you may remember the correlation between quantitative SAT scores (and IQs) and gender ratios of college majors. What if we matched up the quantitative SAT scores — which can be used as a proxy for how quantitatively-focused a major is — with the median earnings for each majord?

(Again, the majors are scaled by the number of recent graduates to provide a sense of relative popularity.)

us-college-majors-income-vs-quant-sat-score-regression

(For the stats nerds: R^2 = 0.624)

Perhaps this all makes sense now: It seems possible that male-dominated majors — such as Engineering, Physics, and Computer Science — earn more than female-dominated majors because male-dominated majors are often more quantitative in nature. These quantitative majors are often employed by large companies to design products, perform data analysis, manage the company, etc., and their salaries are higher to match the responsibilities of the job.

It’s another question of whether businesses and governments should value the services provided by quantitative majors more than, say, Education majors, but I’ll leave that discussion for another day.

Disclaimer

Of course, correlation != causation, and I’m not trying to make any causative inferences here. We would need to perform controlled experiments if we wanted to establish a causal relationship. I’m simply building a plausible narrative around these correlations in case anyone wanted to explore these correlations in more depth in the future.

Summary

  • Female-dominated majors tend to earn less than male-dominated majors
  • This correlation isn’t explained by the employability of the majors
  • It seems plausible that male-dominated majors are usually paid more because they are more quantitative in nature, which large companies tend to value highly

Alternative explanations

So, what do you think? Do you think this explanation holds, or do you think there’s another explanation at work here? I’ll do my best to describe plausible alternative explanations in this section.

  • One of the commenters mentioned that we should view the correlation in the opposite direction: “Men are more driven by income when choosing majors than women. Women value attributes other than income more than men.”

    • Technical notes

      a There are plenty of caveats that come with this data set, and I advise reading up about them in the FiveThirtyEight article.

      b I excluded Petroleum Engineering from this analysis for clarity, which is a small outlier major (only 2,339 recent graduates) with a median salary of $110k and only 12% women.

      c “Recent” graduates are graduates under age 28, which is roughly within five years of graduation on a normal schedule.

      d I performed this merger manually and did my best to match the majors when their name did not match exactly. Several majors were dropped since quantitative SAT scores were only available for a subset of majors. You can find that data set here.

      Dr. Randy Olson is a postdoctoral researcher at the University of Pennsylvania. As a member of Prof. Jason H. Moore's research lab, he studies biologically-inspired AI and its applications to biomedical problems.

Posted in analysis, data visualization Tagged with: , ,
  • Henk Doorlag

    So there’s 2 data points:
    1) higher IQ correlates to higher income
    2) Higher income correlates to higher male:female ratio’s

    I think the explanation of point 1 is simple and undisputed. More intelligent people are more valuable to society. and I’ll wager it works whatever measure of intelligence you use.

    I think the explanation for second data point becomes less disputed if you look at it in reverse: apparently men are more driven by income when choosing majors than women. Women value attributes other than income more than men.

    In a society where the baseline assumption for the last 25 years (formative for those recent graduates) is still a male provider and female care-giver, is it beyond expectation that men will be more focused on income and women multiple attributes?

    • apparently men are more driven by income when choosing majors than women. Women value attributes other than income more than men.

      That’s a very good way of putting it.

      • Kit

        apparently men are paid more for the same job.. so why would it matter? In all my working experience- women are smarter, much much more efficient.

        • D.M.S.

          NO, males and females are Equal.

          • Kit

            LOL D.M.S. you Disqus stalking me? 🙂
            LOLOL “males and females are Equal!” if by 70 cents on the dollar you mean ‘equal’.

            • D.M.S.

              I don’t know about you but I try to have a equal relationship with my wife. Don’t you try to have an equal relationship with your spouse? Or are you better than him/her.

              • Henk Doorlag

                I am better than her in many ways, almost as many ways as she is better than me.

            • Henk Doorlag

              do you define yourself by your income alone… I don’t

          • Henk Doorlag

            but not the same… equal opportunity, Yes; Equal outcome, NO!

        • conlawartist

          Compensation is equal when performing the same work. I can’t believe this trope exists today.

          The reason for the wage disparity is career choice – literally the entire thesis of this article.

          If you take the average salary earned by the entire population of working women, it will be lower than the average salary earned by the population of men.

          This is not the same as saying that women earn less than men for the same work. It is saying that the average man makes more than the average woman. Why? Because men choose higher paying jobs and rarely take years out of their careers to raise children.

          This trivially easy distinction has escaped the public conscience for decades.

          • Kit

            you are wrong. women earn less for the same work.

            • conlawartist

              I’m eagerly awaiting the robust study to back that up.

    • Greg Hammack

      “apparently men are more driven by income when choosing majors than women.”

      …And that may be explained because men are typically the provider and have the expectation to provide for their families.

    • Tegan Maharaj

      You’re calling this a ‘choice’ and a ‘focus’, but I think that that is misleading. It is ‘choice’ the same way someone with no money ‘chooses’ to live in low-income housing. Portraying it as a voluntary decision implies there is not a problem that needs to (and can be!) solved in the gender imbalances here.

      • Henk Doorlag

        I disagree,
        First, there is no problem to be rectified (let alone if it can be, please suggest something).
        Second, The choice is in what to major. Women choose majors that lead to lower income careers (on average). There is no institutional pressure against women choosing higher income education. There maybe a societal pressure, but if you think social engineering is a good idea… again, I disagree.
        I do notice that om my side of the pond, it seems eastern European countries show higher more equal male:female ratio’s in engineering. maybe because communism had no patience for such social constructs as male providers, female caregivers, ar maybe because in poorer area’s women will look more to the income potential of their education and the inequality is down to luxury in the west.

        • Tegan Maharaj

          When a group of people is excluded from or not participating in an activity for some reason other than their ability or desire to do that activity, I think there is a problem. Women are not less good at quantitative-related jobs (e.g. physics, engineering, etc.). Therefore it is a problem for the women who might have been kept from these fields, for the girls in the future who might still be able to go into those fields, and for the fields themselves, in loss of diversity.

          This is a complex problem and there is no one-liner solution. There is a long list here: http://www.theguardian.com/science/occams-corner/2015/jun/19/just-one-action-for-women-in-science complied from tweets, about things you can do to encourage/promote women in science.

          You seem to be defensive of past humanities’ faults and the status quo. I completely disagree with your statement that there is ‘no institutional pressure against women choosing higher income education’. The best hard example I can think of is that study where they switched female names for male names, and they were more likely to get hired/get in to schools, and switched male names for female names, and men were correspondingly less likely to be hired/get in. http://blogs.scientificamerican.com/unofficial-prognosis/study-shows-gender-bias-in-science-is-real-heres-why-it-matters/

          I also kind of disagree that social engineering is a bad idea. We should all be acting in a way that, if everyone acted that way, we would have the society we want. If you want egalitarian representation in sciences, you should act that way, and encourage women in science to achieve that reality.

          • Henk Doorlag

            Nobody is excluded, women chose different paths in their life. If you have evid er nce that wonen are excluded, taken it to court.
            Look at medicine, an historically male dominated field; arguably more complex and higher paid than STEM, completely taken over by we omen because it apparently beter matches their average interest.

            Why do you want to force women into fields they don’t want to go? Why must those girls go into math? Let the see m chose for themselves and be happy.

            And women are less good at quantative stuff… on average.

            • Tegan Maharaj

              Women are excluded from many science-related and high-income work by laws (e.g. no paternity leave that would allow the father instead of the mother to stay home with kids), social pressures, and negative work environments.

              It’s not illegal to make someone uncomfortable, and I’m not saying that a specific person or company has a policy that intentionally excludes women from equal status with men in the workplace. This would obviously be something to take up with the courts, but the problem today is more of lingering social ideas that still support and subtly uphold old ideas of sexism. Addressing this kind of problem is not for the courts; it has to happen with individual people in a society understanding and accepting points of view that are not their own.

              I don’t want women to go into fields they don’t want to go into. You are looking at data about which fields women are currently in and assigning them preferences for those fields based on no information whatsoever. I want them to be able to choose for themselves with the same confidence and security that men have, and this is not the current reality.

              Women are not less good at quantitative stuff. It has been shown by many, many studies (here is a very readable summary: http://thesocietypages.org/socimages/2013/03/07/the-truth-about-gender-and-math/) that this is simply not true, and is mostly an issue of confidence – if you tell male people their whole lives that they are bad at something (e.g. showing emotion), that will end up being true, but that’s a societal pressure, not anything inherent to the gender. And more importantly, good science requires all kinds of skills, not just good quantitative analysis.

              I’m curious what your aim is here. Do you think there should be fewer women in science? Do you really believe that there is no lingering historical effect of sexism that individual members of society need to address in their everyday lives? Do you have any data to support your repeated claim that women are going in to non-quantitative fields out of pure preference? You seem very defensive. I apologize for the antagonistic use of ‘you’ in my previous comments. Maybe you are not the kind of person who would discriminate against women in the workplace, so you can’t imagine it happening and are upset with me for suggesting that it does happen.

              But believe me, a lack of role models, lack of experience, and antiquated laws and social customs all make it more very likely that some people will accidentally do things that make it harder for women to get into science, to enjoy their experience working in science, and to stay in science, and that is a problem that needs to be addressed.

              • Henk Doorlag

                We’ll have to agree to disagree, however:

                “Women are excluded … by laws (e.g. no paternity leave that would allow the father instead of the mother to stay home with kids)” – so the lack of opportunity for men to take time off from work to take care for children is oppressive to women?

                “more of lingering social ideas that still support and subtly uphold old ideas of sexism” I’m against positive discrimination and quota’s specifically to prevent these… outdated world views from finding confirmation. the whole “she’s just here to make up the numbers”
                Most high achieving women agree on that.

                Women have been free to chose for decades now. They broke the glass ceiling in medicine and 33 other STEM fields (https://www.aei.org/publication/women-earned-a-majority-of-2012-doctoral-degrees-in-33-stem-fields-can-we-stop-calling-it-a-national-crisis/), the majority care related… they have chosen, stop trying to insist they do pure mathematics or astrodynamics.

                You look around and see few women and blam everyone except yourself… maybe you are an exceptional individual?

                Are men stronger than women? Any individual women can certainly be stronger than many individual men, and ronda rousey will beat most males… in fact, any women will, with a little training beat most untrained men, but on average – men are stronger.
                The same is true of certain cognitive faculties. men and women are different, with different strengths and weaknesses. That does not make one less than the other – it just makes it more interesting
                (and your link results in “Sorry, we couldn’t find what you’re looking for.”)

                I didn’t claim women should be avoided in science. My claim was that women are on average less strong in quantitative cognitive tasks.

                My Goal? Less sex-discrimination. fewer women only support groups, women only scholarships. Scholarships should be earned, support should be universal and opportunities for men and women should be equal. The best men/woman should be selected for the job, not whoever fills a quota. The most competent scholar should be supported, not based on what goes on between their legs.

                I want a world where I don’t have to tell my daughter that her having vagina is such a problem for her professional future, we need billions in special interest money and effert just to get her on an equal playing field with the boys…

                Firstly, because it is bullshit. Her mother shows her so. And secondly, because you only real arguments: “social pressures, and negative {} environments” and , “lack of role models, lack of experience, and {} and social customs” are my responsibility for the next decade and a half. If she believes that crap, it’ll be my fault!

                “some people will accidentally do things that make it harder for women to get into science, to enjoy their experience working in science, and to stay in science”

                Replace women with men and it is still true. Life is shot, but when life hands you lemons, you don’t run to your local SJW and set up a trigger warning. You kick the lemons between their legs (whatever sex they are) and go on with your day!

          • Hector_St_Clare

            What’s your evidence that women are ‘not less good’ at physics, engineering etc?

            Women have equal average iq to men, but smaller variance, so you will have fewer women at the top and bottom of the iq spectrum.

          • Hector_St_Clare

            Having said that, I don’t think high iq jobs *should* pay more than other equally important jobs- the work of a farmer or factory worker is as important to society as a biology professor- but I think ensuring equal numbers of women engineers, etc. is a fool’s errand.

        • Hector_St_Clare

          Actually, the old Soviet Union (of which I’m something of a fan) had a similar gender-related wage gap as modern America.

    • jimmyt

      on IQ’s

      According to the 1994 report “Intelligence: Knowns and Unknowns” by the American Psychological Association,
      “Most standard tests of intelligence have been constructed so that
      there are no overall score differences between females and males.”

      testing standards changed to make people feel better, rip society

      GamerGate

      • Henk Doorlag

        do you ave any reason to suggest there should be a significant mental capacity different along biological sex lines?

        The main question is, Why were the testing standards amended?
        RIP skepticism.

    • drklassen

      Actually, that wasn’t IQ and income, it was SAT Quant score and income. So 1 and 2 are related in that society places a higher value on “men’s work” than “women’s work”.

      • Henk Doorlag

        at best, I should correct my statement to
        higher (Quantative SAT scores) correlate to higher income

        how do you jump from that to men’s/women’s work?

  • Jonas

    Two comments:

    1. The increase in women Computer Science majors from 2011-2015 seems next to impossible. Looking at your earlier blog, such an increase would be historically unprecented. I’m not saying the data are faulty, but perhaps it’s because of some category changes?

    2. Your blog didn’t touch upon the feminist narrative, which might be that “men’s work” is seen as *inherently* more valuable than “women’s work” and therefore paid less. That would mean a breakdown of the “rational” market, which is odd but not entirely unheard of.

    Some reflection of this can be seen in the fact that even after you compare for profession, experience, etc., women earn less than men (http://blog.dol.gov/2012/06/07/myth-busting-the-pay-gap/). Part of this might be attributed to men’s negotiation strategies, which are more forceful, but it’s there nonetheless.

    I’m not sure I buy this explantion entirely, but the fact that un- and underemployment do not at all correlate with income strengthens the argument. After all, aren’t those measures an expression of the true demand for these majors?

    • 1. The increase in women Computer Science majors from 2011-2015 seems next to impossible. Looking at your earlier blog, such an increase would be historically unprecented. I’m not saying the data are faulty, but perhaps it’s because of some category changes?

      I agree. The 58% women in CS figure comes from the American Community Survey sample. There may be sampling error there, since figures in 2011 showed that CS is sitting at about 20% women. That would make CS less an outlier, but it would still fit the overall trend.

      2. Your blog didn’t touch upon the feminist narrative, which might be that “men’s work” is seen as *inherently* more valuable than “women’s work” and therefore paid less. That would mean a breakdown of the “rational” market, which is odd but not entirely unheard of.

      I’m purposely trying to avoid extrapolating too far beyond what the data states, since I think we are too quick to do that in research relating to the gender wage gap. I think we would be better off sticking to making statements that the data supports, finding holes in the data (as you pointed out in question 1), and mending those holes where possible.

      • Ironica

        But the data, as it doesn’t imply causation in any direction, equally supports the surmises that “majors with a higher ratio of men earn more money” as “majors that make more money attract more men.” The first statement is consistent with modern feminist theory; the second continues to explain gender wage differences normatively through male behavior as default “desirable” behavior.

        If this isn’t a spurious correlation (and the fact that there’s a consistent gap between what men and women earn, even when you control for a huge number of other factors, implies that it probably isn’t), we still have the conundrum of whether “men select for higher-paying majors” or “pay selects for more male majors”.

        And then there’s the *further* possibility that “Majors that pay more select for men.” You note the factor of quantitative majors being more male-dominated and higher-compensated, but one explanation for this that is well-supported by other data (much of which is necessarily qualitative, ironically) is that women are steered *away* from highly quantitative disciplines. This would be the “women are just bad at math” narrative, which is highly evidenced in popular media, opinion surveys, and so forth. If we *believe* that is true, we do not encourage women to pursue heavily mathematical disciplines, and we may even actively discourage women who express interest in them.

        I agree it’s important to not make any conjectures that aren’t in the data, but I think you’re still including a particular narrative bias in how you’re stating the conclusions.

        • brainy37

          The first statement is consistent with modern feminist theory; the second continues to explain gender wage differences normatively through male
          behavior as default “desirable” behavior.

          First, I’m not sure why you cite feminist theory to state that higher paid jobs have more men. That has nothing to do with feminism and everything to do with work type.

          Second, the data does not show that male behavior is more desirable. Where did you even come up with that?

          You note the factor of quantitative majors being more male-dominated and higher-compensated, but one explanation for this that is well-supported
          by other data (much of which is necessarily qualitative, ironically) is that women are steered *away* from highly quantitative disciplines.

          Actually there hasn’t. The data point(s) you’re describing deal in middle wage research science and not engineering. Research being qualitative more than quantitative unsurprisingly. Unlike CSS or research science, there has been no massive gains in women studying engineering fields with women only amounting to 12-19% of all engineering graduates. With that low of a graduation rate you’re going to see a disparity regardless.

          I agree it’s important to not make any conjectures that aren’t in the data, but I think you’re still including a particular narrative bias in how you’re stating the conclusions.

          Translation: I don’t want to make a conjecture so let me make some conjecture. I don’t like your data presentation so please use this feminist narrative even though the data does not support it.

          Sir, this is why many have a hard time taking feminist theory seriously. It relies on begging the question or circular logic way too often.

    • notaname “my big fat kot” nota

      That link provides a 404, and this article here seems to disprove your hypothesis. Women don’t get paid less because vagina, they get paid less because they choose lower paying jobs.

  • Jesse Hunter

    while you did not state it outright, it seems this would imply that the average male graduate has a greater iq than the average female graduate.

    while there is some debate about whether men and women have differences in iq variance, I think this can be explained away by the total number of men and women who enter university.

    Because more women enter, it seems likely that women of average and below average intelligence still pursue education. Men of below average intelligence have other outlets to pursue success in (eg military, trades).

    • I explored IQ a bit in my previous post related to the gender ratio in college majors issue. I certainly wouldn’t say that the average male college graduate has a greater IQ than the average female graduate. From my research, the apparent higher average IQ of students in male-dominated majors seems to be more about preference for quantitative majors rather than the average intelligence of any particular demographic.

    • Henk Doorlag

      Great point. “Because more women enter, it seems likely that women of average and below average intelligence still pursue education”

      Add to that that the lower intelligence males have more options into the job market outside the colleges than women (unschooled work is also either prominently male (construction) or 50/50 (service industry)), there is a stronger pre-selection of males into college.

      thus the IQ distribution at the onset of college is further distorted. Lower en males stay away from college more than lower end IQ females (just talking IQ, not placing a value label here)

    • jimmyt

      iq tests are constructed to make it so there are no differences in gender

      they are now doing this for race differences too

      #feefees

    • Anon

      Male and female IQ distributions are different. Female IQ tends to cluster around the mean more than men, so there are both more male retards and more male geniuses. It’s possible that majors that correlate with a high IQ are male dominated because there are simply more men with very high IQs on average.

  • Would the sizes of the various cohorts change or inform the stats somehow, i.e., supply & demand coming in as a function of the number (vs. gender) of people available trained in field X?

  • buddyglass23

    Interesting analysis.

    Can you produce a version of the chart comparing earnings to SAT scores that has labels on the dots? I’m curious which are the outliers. That is, professions that pay more (or less) than avg. SAT score would predict.

    • David Wilson

      That would require individual data lines that tie specific individuals/ultimate majors to college admissions data. I would be VERY surprised if data of that specificity is published.

      • buddyglass23

        I was just asking for labels for the data points linking avg. SAT (Quant) score and avg. earnings for different majors.

  • Miguel Dias

    Excellent study, congratulations.

    In my country,
    it is extensively repeated that women earn 2/3 of men, on average, and this is
    due to sexism within the society, period. I always asked myself why my greedy
    employers were not just firing men and hiring women, and saving 1/3 of the
    payroll.

    I
    understand the limits of the conclusions. And certainly some aspects on
    earnings differences may be, or could have been, influenced by some kind of
    prejudice, but thinking about the future… the fresh students can choose
    differently, be more oriented to their future incomes, and just change the
    table. I can´t believe the income of some major, let´s say, Chemical
    Engineering, would just drop if its male-female ratio changes.

    Best
    regards

  • John Chilton

    Regarding accounting there appears to be an inconsistency. You describe it as male dominated, but in your figure the orange square labeled accounting exceeds 50%. When I look at the get-hub data it says the share of women in accounting is 25%.

    • You’re right – that’s outdated text from a previous version of the article. I’ve updated the text to reflect the latest data. Thank you for pointing this out, John.

  • darl

    Maybe I missed something? But to me any analysis of PUMS-based data without consideration of standard error/margin of error is incomplete.

  • CMay

    In this dataset, a correlation can be seen between the higher IQ segment and the quantitative majors. I’ve recently been reading a lot about giftedness. Giftedness can also be correlated with the higher IQ segment… provided it’s a giftedness in math, spatial relations, verbal reasoning – things which are easy to quantify. Giftedness in general social skills (i.e. leadership, complex interpersonal and emotional situations) or creativity exist, but aren’t easy to quantify and aren’t well represented in standard IQ tests.

    It’s challenging to measure creativity or emotional intelligence in a test, but we all know these people when we meet them. They do significantly better when thinking outside-the-box, or when put into a complex political situation, but this may or may not directly translate into pay, especially if the profession as a whole is considered to be low- or mid- level.

    This makes me wonder if there is still a measurement problem when discussing gender, IQ and profession. How many women are actually gifted in the social sense, which is not easily measured? How many professions which contain more women require these poorly measured skills? Western societies certainly place emphasis on measurement in order to calculate value, so salaries appear to reflect that. If we could measure creativity or leadership or empathy in a quantitative way, would that change our value measure and therefore salary in a profession such as nursing or teaching? I’m not sure, but it’s interesting to ponder…

  • Violet

    I would be curious about the careers where there is an equal male-to-female ratio and what the median pay is for each gender. Its one thing to say that more “male” centric jobs get paid more and vice versa, but what about within those fields. is the pay scale the same or different?

    • winstongator

      I am likewise curious, and would love to see that addressed.

      fwiw, my wife makes > 4x what I make. We do not work in the same field, but both our fields are > 50% male.

      • Violet

        I work with men who do the same work as I do, have less experience, but make about $5K or more than I do. Its pretty frustrating as a single parent to get paid less for doing the same work.

        • Femghazi

          Why don’t you take your company to court then?

          • Violet

            How could I even provide proof? There are too many factors involved. However, generally men are seen as more valuable to the team than women are.

            • Femghazi

              Exactly, they are more valuable and will take their skills elsewhere were the business to pay you the same. It’s pure economics.

              • Violet

                Not sure if there is a typo somewhere, but are you saying men are more valuable in the workforce?

                • Femghazi

                  Yes. Most certainly they are.

                  • Violet

                    Please explain further?

                    • Femghazi

                      Men, on average, work longer hours, choose more demanding majors, don’t need maternity leave, call in sick less, work more dangerous jobs, are willing to move, don’t have children, are not emotional and are just more logical and rational and get the job done.

                    • Violet

                      I work the same long hours as my male counterparts. My company grants the same family medical leave to both men and women. In the 4 years that I have worked for my employer I have never called out sick, not for myself or kids. Besides who says that if the kids are sick that only the mother has to call out of work? Who says women don’t work dangerous jobs too. We have women as cops, in the military, and even first responders.

                      I am completely floored that you make these assumptions about women. Do you treat women in your family and office this way?

                    • Femghazi

                      Blah blah blah, you answer quite clearly shows you have no logic. I specifically said ‘on average’ for a reason. Maternity leave is a cost the company should not have to bear, it is both unproductive and costly. Men don’t bring this cost at all and thus bring greater productivity to their employer.

                      I treat women as they are, women. Men and women both have a place, both have strengths and weaknesses. In order for women to ‘earn’ the same as men, women would have to be men, which they can’t be. It’s simply both a combination of biology and economics. It’s not discrimination. That is the end of the lesson, I care not one wit if you feel offended.

  • hacksoncode

    While I can’t quite tell just by looking at the data, the qualitative appearance of that first graph seems to support an idea I’ve had for a while: I’ve often thought that much of this issue is a simple matter of supply and demand.

    Historically, there have be a small number of careers that women have gone into, and to a degree this has continued over the decades that women have entered the workforce in large numbers. Perhaps the supply of candidates in those careers is simply higher than the demand.

    The careers with larger blocks are mostly on the right side of your graph, and there are a large number of small blocks on the extreme left.

    • Joseph Zsombor

      That’s a pretty good point. Take psychology, for example. America has a huge supply of psychology graduates (more and more of whom are female), and the wages psychologists are earning, not to mention the ability to get a job, are representative of that high supply.

    • Joseph Zsombor

      This is a really good point. I think psychology is a good example. It’s a field that’s growing rapidly, with an increasing ratio of female graduates, where the supply is beginning to outweigh the demand. Supply & demand seems like a very reasonable interpretation.

  • John Hancock

    And, yet again, the “gender wage gap” claim is proven false. There is NO GENDER WAGE GAP.

    • BROSEPH

      You completely misinterpreted this blog post. Control for career, and gaps do exist, there is a study published by researchers at Harvard about it.

      Edit: Found it. http://scholar.harvard.edu/files/goldin/files/goldin_aeapress_2014_1.pdf

      • John Hancock

        NO. THEY. DON’T.

    • Please don’t take this personally, but it drives me crazy when I see claims like this.

      Whether or not men’s choice of major leads to higher salaries (it probably does) doesn’t say anything about whether women are discriminated against at work.

      In fact, there have been a ton of experiments that have tested this question. And every one I have ever seen suggests that discrimination does affect women’s salaries.

      For example: http://blogs.scientificamerican.com/unofficial-prognosis/study-shows-gender-bias-in-science-is-real-heres-why-it-matters/

      The analysis here is great. But you are grossly over-extrapolating.

      • John Hancock

        The actual empirical data and analysis demonstrate, unequivocally that there is NO gender wage gap. You have no idea WTFZ you’re talking about and here is the actual data:

        http://www.americanthinker.com/2010/09/the_wage_gap_myth.html

        http://www.bloomberg.com/news/2012-08-13/don-t-blame-discrimination-for-gender-wage-gap.html

        http://www.businessinsider.com/women-in-tech-make-more-money-and-land-better-jobs-than-men-2010-9

        http://www.reuters.com/article/domesticNews/idUSN0752118220071107?feedType=R

        http://www.nytimes.com/2007/08/03/nyregion/03women.html?_r=1&oref=slogin

        http://www.thedailybeast.com/articles/2014/02/01/no-women-don-t-make-less-money-than-men.html

        http://www.washingtonpost.com/blogs/fact-checker/post/the-white-houses-use-of-data-on-the-gender-wage-gap/2012/06/04/gJQAYH6nEV_blog.html

        http://www.telegraph.co.uk/comment/columnists/vickiwoods/7957186/Sorry-ladies-Im-not-worried-about-wage-gaps.html

        http://www.cbsnews.com/8301-505125_162-28246928/the-gender-pay-gap-is-a-complete-myth/

        http://wwww.examiner.com/x-22884-Canada-Politics-Examiner~y2010m2d22-Gender-pay-gap-is-not-what-activists-claim

        http://www.slate.com/articles/double_x/doublex/2013/08/gender_pay_gap_the_familiar_line_that_women_make_77_cents_to_every_man_s.html

        http://www.aauw.org/research/graduating-to-a-pay-gap/

        http://www.nationalreview.com/corner/345106/equal-pay-day-debunked-three-minutes-carrie-lukas

        There is no glass ceiling. Just ask Meg Whitman, Carly Fiorina,
        Indra Nooyi, Andrea Jung, Anne Mulcahy and many others. You get what you
        NEGOTIATE for compensation, NOT what you perceive you deserve. Men are better
        negotiators when it comes to compensation.

        Factually speaking, that is not our men’s fault that this occurs, IF it does
        occur. Do you have a scientific peer reviewed study which can be cited for this
        data? You get what you negotiate and what you are willing to work hard for to
        achieve. There is no glass ceiling. If men in similar jobs aren’t paid equally
        to men they compete with — and they’re not
        — how is it
        possible to mandate equal pay between genders? It isn’t. Everyone is paid what
        he or she is worth, and that compensation is determined by job skills, job
        requirements, personality, geography, industry, company size, AND negotiating
        skills. When a company offers a woman a salary, and she accepts it — without
        negotiating a better deal —
        she is agreeing to the amount. She can’t later complain that some man out earns
        her. People often cite, and blindly accept, a mysterious “statistic” indicating
        that women earn 70 cents for every dollar a man earns. Surprisingly, few ask
        how that figure was derived. It’s meaningless. Moreover, it does not factor in
        child support and alimony, which women receive in at least 90% of cases. And,
        let’s not forget the imputed income to women when men pay for their
        entertainment and travel. All told, according to Allianz, the
        financial-services behemoth, women control 60% of American wealth. Take a man
        earning $2M per year and another living under a bridge, earning nothing. Lump
        them together: each has an average annual salary of $1M. Makes sense on a
        calculator but not in reality. Men and women, in general, have different jobs.
        In addition, more women than men are in low-paying, part-time jobs. Yes, there
        are men and women with equal jobs in the executive ranks. But, rarely do you
        find women welding steel atop bridges or men changing diapers in daycare
        centers. So, if you take a group of women, ranging from CEO to secretary, and a
        group of men, ranging from CEO to Navy SEAL, the arithmetic averages of their
        wages will not be equal. But, because most men feel guilty for being alive, let
        alone walking upright, they easily succumb to this glass-ceiling wage-gap
        crap. If you take maternity leave you
        are not working ergo why should your salary level remain consistent with those
        who remained and are advancing? You are making a conscious decision of family
        over career. Live with your choices and the consequences of your decisions.
        That’s the real world.

        Unless the data trotted out and called “factual” is
        peer reviewed legitimate studies following scientifically accepted statistical
        formulas and processes which can be confirmed and checked by 3rd parties for
        accuracy, it is no more relevant or real than anecdotal experience.

        • Many of those are just opinion pieces. And none of them are randomized experiments.

          Observational Study vs. Experiment
          “IMPORTANT: An observational study may reveal
          correlation between two variables, but only a
          randomized experiment can prove cause‐and‐effect”

          http://www.csus.edu/indiv/n/norrisa/stat50/Observational%20Studies%20and%20Experiments.pdf

          You’re probably not a bad guy, but the science says that gender discrimination is real. It may not affect you, but it’s a real bummer for a lot of people. So unless you can point to some experimental evidence showing otherwise, you should not go around claiming it does not exist.

          The link I sent before was published in PNAS. Here it is again, along with a few others.

          http://www.pnas.org/content/109/41/16474.abstract#aff-1
          http://www.utstat.utoronto.ca/reid/sta2201s/gender-teaching.pdf
          http://facultyhiring.uoregon.edu/files/2011/05/ssrn-id2063742-1pop41q.pdf
          http://www.nber.org/papers/w5024.pdf

          • brainy37

            The only study that actually shows a lower wage offered for equal work is the Scientific American blog. Even the restaurant study showed hiring, not wages.

            Furthermore, many of the biases in those studies do have grounds which is why even women fostered the same for other women. Yes, #4 from the SA study is a legitimate complaint especially when a majority of those women will not return to work as they want to raise their children. Thats a loss in training time, resources, and requires a team to make up the slack.

          • John Hancock

            The cite DIRECT EMPIRICAL DATA. The facts don’t change regardless of whether your agenda and “message” is in conflict with them. There is NOT GENDER WAGE GAP and hasn’t been for a very long time. Grow up.

            • Aaron Blenkush

              John, you’re not really making a good case here.

              • John Hancock

                Actually, I am. You’re delusional and making fallacious, erroneous statements in a failed attempt to lose your virginity. Good luck with that.

                • The Dark Knight

                  So which hooker did you have to pay to lose your virginity a$$wipe…Or did that catholic priest take it away from you back in the day, which is why you’re so angry with religious people now…………..FOOK OFF MORON

          • Anon

            The point is that the wage gap is not as large as many would lead you to believe. The figure is closer to 97 cents to the dollar rather than 77. The 77 cent figure is determined by comparing the average earnings of all women vs the average earnings of all men. Your data is showing that these difference are mostly driven by women’s career and major choice which is focused less on income and more on interest, quality of life, work life balance, etc.

            In our society, a woman who focuses on a high salary career will not benefit in the same way as a man. Likewise, a man who focuses on a balanced lifestyle will not benefit in the same way as a woman. The reasons for this should be clear, and have to do with the fact that men and women on average look for different traits in an ideal partner.

            I realize that as a university professor your hand are tied here and I enjoy your research. Thanks for the article.

            • “Your data is showing that these difference are mostly driven by women’s career and major choice”

              This is your speculative interpretation of the data. All the data shows is correlation, which could have any number of explanations.

              Assuming it is a factor in the gender pay gap (I think it probably is), how do you conclude that it accounts for most of it? Your 97% number is nothing more than a guess.

              I agree that the 77% number is not correct, but there are way too many factors involved to dismiss the pay gap because of a single correlation.

              Here is another correlation that implies women should be earning more.

              Median salary of a college grad (2013): $45.5k
              Median salary of a non-college grad: $28k

              % of bachelors degrees held by women: 60%
              % of bachelors degrees held by men: 40%

          • Bill Angove

            Max, John did not claim that gender discrimination was not real, he claimed that the wage gap (a macro-economic phenomenon) was not real. You’re arguing about different things.

      • Nicolas Versteher

        my daughter a chem E. paid the same as the men.
        So, BS.

        • SumAnon

          Well *my* daughter works in engineering, and is paid less. So BS to you.

          • brainy37

            If this is true then your daughter needs to ask for a raise and you as the father ought to be pushing her to do the same.

            • SumAnon

              MY GOD that fixes everything! If only we could get the word out, there would be no workplace inequality anywhere!
              I’m nominating you for the Nobel Peace Prize.

          • lawgone

            And you eliminated all other variables and came to the conclusion it’s because she’s a woman?

      • SumAnon

        So I’m not going to reply to the blatant cut/paste essay from John Hancock, but I have to say: Wow. Just …wow. Someone’s got some issues a lurkin’ if they have that prepped and ready to go.
        If he’s going to go on a huge Men’s Rights rant, at least he could be sure that all the links work and are related to the discussion. But I guess no one take him seriously and actually clicks on those articles? Though it WAS amusing to see him demand ‘scientific, peer reviewed’ articles when half the articles he posted (because yes I did check) are opinion pieces that cant even bother to site their own data.

      • Henk Doorlag

        1) wage is only one part of the total compensation for the work you do
        2) calling it a gender wage gap implies it is due to the bits between your legs. discrimination and such. There may be a difference in wage between men and women, and the difference can be statistically significant without any discrimination.

        and if there is no discrimination, and any wage difference is down to free choice (not choosing less money, but choosing different work/careers), what is there to be done. We are spending a lot of money and effort on getting women into STEM and such – what about using some of that money and effort to get men into nursing and primary education??? That should solve that problem as effectively as the other way around…

        • Completely agree. A difference in pay is not a problem per se. What matters is whether or not there is discrimination.

          You can try to test this by decomposing that 77% pay gap, and seeing if you can explain it by other factors. But the results will never be conclusive because the number of possible explanations is infinite and the different factors are not independent.

          Instead, you can test the question directly by running an experiment that holds everything constant except for gender. This has been done many times, and every case I’ve seen suggests there is gender discrimination in the workplace.

          I don’t have the solution. I just think it’s important to be honest about what the problem is.

          • Henk Doorlag

            One step back… the one making the positive claim has the burden of proof.
            there is a statistically significant difference in income divided along sex lines. That is a fact. the rest is opinion and claims.

            • Peer reviewed science is not opinion

              • Henk Doorlag

                data is not opinion – but no-one was charged with a crime so no actual discrimination is found.
                Data is interpreted and discrimination is one possible explanation, but the discrimination claim itself is not factual.

  • ChadEnglish

    One hypothesis is that men and women are genetically (or at least hormonally) motivated toward different choices including things vs people, providing for (earning) vs caring for (family), max income vs life balance, etc. If true, some of these (things vs people) predict the sexually dimorphic choices of fields, e.g., engineering & hard sciences vs social sciences and people-based sciences. Others (providing/income vs caring/balance) predict that men will aim for higher paying fields regardless of what they are.

    Looking at the data, it appears hard to separate. In the first plot, the types of majors ranked by income level appears to more or less correspond to the order you’d put them in for “things vs people”, though that isn’t a strict definition. There is insufficient data here to test that hypothesis. However, the outliers may provide some clues. Nursing, for instance, heavily involves people care and is dominated by women so fits that hypothesis. But the men seeking higher income predicts that men should choose nursing more often. Similarly, men seem to choose transit science, clearly a “things” field, more often than the income-motivation would predict. These outliers then suggest the “things vs people” is a stronger factor than income seeking, though both could be factors.

    This sets up a need to separate the “things vs people” factor from the “income vs balance” factor. Statistical analysis of available data may be sufficient (though not from what is provided here), or it may require experimentation. A first step is finding better data on classification of fields by “things vs people” categorization and just what that means.

    There are a bunch of references in Comment #10 here: http://www.macleans.ca/society/science/gender-inequality-in-the-sciences-its-still-very-present-in-canada/. That may be a starting point. (Perhaps I’ll copy and paste as a reply below.)

    An interesting second question would then arise, which is what forces would link higher paying jobs for “things” vs “people”. Such as study would also nicely removes any sort of inflammatory discussions of men vs women and all of the fighting that goes with it. The result might very well show that “thing” jobs earn more because they are dominated by men, confirming the male-bias hypothesis, or that it has nothing to do with gender, disproving that hypothesis, and that the gender difference is due instead to biological differences.

    A simple example of such a case would be if income relates to, say, area covered by a painter per hour. A taller person could reach higher and therefore cover more without going up or down a ladder and therefore taller people will tend to earn more, and biological evolution has made men and women sexually dimorphic in height, such that men would earn more statistically because of height, and if you grouped people by height and not gender you’d get a higher correlation with income.

  • Eleanor Siler

    Nice plots Randy! I don’t think you should discount the opposite causation — that quantitative fields earn more because they’re more associated with men, and society sees “men’s” work as more valuable. Maybe you could test this by using teacher gender ratios vs. relative teacher earnings in different countries? Or with any other job that would have cultural gender ratio variance.

    • Femghazi

      Well, that’s bull. Society doesn’t just see one work as more valuable than another unless it actually is. Engineers, Scientists, Doctors and other STEM subjects earn more because they provide more economic benefit to society.

      • Eleanor Siler

        That’s silly. For most jobs, especially well-paid ones, it’s impossible to decide exactly what the economic benefit is. For a doctor, what is the economic benefit of saving someone’s life? Does it depend on who’s life it is, or their age? For a scientist (like me), what’s the economic benefit in creating knowledge? You don’t know in advance how useful the knowledge will be. How valuable is it to educate children? How valuable is it to fix a bug in an operating system? The creater of flappy birds got lots of money- does that mean the flappy birds is valuble to society? Our values determine what’s valuable.

        Also, your premise is wrong. The money you earn depends on labor supply and labor demand, not on inherent economic benefit.

        • Femghazi

          What the hell?! Are you really saying that companies and businesses, hospitals and other institutions have absolutely no idea of what their employees bring to the table?

          The money you earn depends largely on your skill set, your willingness to work and availability of jobs. It’s a whole range of aspects including supply and demand.

          If feminists really want to sort out the issue of men getting paid more and they really think women do the same work at the same efficiency as men, they should start their own companies, businesses and only employ females because, quite frankly, they were work the same for less. A profit making enterprise if ever there was one. These female enterprises would be able to undercut any male enterprise due to the costs saved.

          Then again, you have no intention of actually starting and working that hard and never have, far easier to cry to the government.

  • idlewild

    Poop in my butt.

  • Aaron Clarey

    Pussyfooting around from stating the obvious is not going to close the wage gap any time soon, if ever. Harsh truths telling women they need to major in fields that require math skills and are in demand need to be stated, reiterated, and instilled in young women’s minds. I know being in academia you can’t say that (you already took enough of a risk doing this). But we in the real world can.

  • Femghazi

    Does all the data take into account exact hours worked, including overtime and such?

  • A factor not considered is that of the economy itself. “Free Trade” has lead to industrial, union-protect, high-paying, low-skilled, manufacturing jobs to go away. These jobs offered a “leg up” to their participants.

  • I really like your analyses – keep up the good work!

    I’d argue, however, that examining under- and unemployment rates to college majors isn’t the best analysis for evidence of wage suppression because those graduates are presumably working in unrelated fields (in fact, even the majority of STEM majors don’t work in STEM jobs: http://www.census.gov/dataviz/visualizations/stem/stem-html/). Analyses by the EPI strongly suggest that oversupply of labor (which you can assume with female-dominated majors) is associated with wage stagnation:
    http://www.epi.org/publication/shortage-skilled-workers/

    And you might be interested in the Federal Reserve Bank of New York’s analysis showing a female wage premium in 29 fields (for both male- and female-dominated majors) – but only when they’re recent graduates (age 22-27):
    http://libertystreeteconomics.newyorkfed.org/2015/08/when-women-out-earn-men.html#.VdeSlSxVhHx

  • Tegan Maharaj

    The problem is *Why* there are more men than women in quantitatively-oriented majors. As a woman doing an advanced degree in computer science, I strongly do not believe that this is because women are ‘not as good’ at quantitative stuff, or that they ‘prefer’ other fields. I think that this imbalance comes from a combination of cultural pressures and social norms that result in it being way more difficult for women to enter these fields.

    It’s like saying “Why aren’t there more poor people with high-paying jobs? It seems that poor people don’t go to university as often, and so they aren’t qualified for higher-paying jobs” … and then concluding from that that poor people just don’t want to work as hard to go to school, or just prefer lower-paying jobs. No, they don’t, they’re people just like you who may or may not want to go to university and may or may not be as intelligent or capable of high-paying jobs; it’s just way harder for them to afford the prerequisites.

    Historical precedent, workplace environment, cultural/social pressures, discrimination … getting through all of that is a prerequisite for a female person to work in a male-dominated field, and the emotional, social, and monetary cost of that is too high for many.

    • Henk Doorlag

      BS. You’ve got it the wrong way around.
      1) poor people in high income jobs is a contradiction. I expected a computer science major to have spotted this

      2) there are many places in the world where a top-notch education can be had for no more than a token cost – e.g. the Netherlands. And better yet, that token cost is both well within everyone’s means ( and else is further compensated by the gov’t) AND is not differentiated according to gender.
      In these places, low income still significantly correlates to less education. So it seems there is a link between those two. Maybe even causal?

      3) Historical precedent and cultural / social pressures cannot be regulated, that is social engineering and it has so far always misfired. Workplace environment and and discrimination are illegal and actionable. What would you suggest we actually do to improve these two?

      • Tegan Maharaj

        1) I’m sorry I didn’t clarify what I meant by the word ‘poor’. I meant ‘people from low-income backgrounds’, but ‘poor’ has significantly fewer letters. My bad.

        2) I am not a sociologist, but imagine you lived in a home where you often didn’t have enough food, your parents couldn’t afford enriching toys and books, or maybe had to work longer hours to support and didn’t have time to help you with your homework. I don’t find it surprising that kids growing up in this environment would have a harder time at school, would become discouraged, and be less likely to continue further. That low-cost higher education does not correlate with people of low-income background choosing higher education to me indicates that that solution does not solve the problem of equitable access to higher education, and support should probably start earlier in elementary school or something.

        3) ‘Historical precedent’ for the future is NOW. Social/cultural pressures can be regulated, by individuals changing their behaviour and by affirmative action programs. Justifying the continuation of non-egalitarian practices by saying ‘well that’s what we always did’ is BS.

        In terms of what I would suggest, I think since the 50s in North America we’ve created a self-reinforcing cycle of a lack of female role models in science for young girls, but happily I think this is changing – due to people believing that there is a problem and doing something about it, not by denying that the problem exists. On a personal level, you can realize that it may have been harder for your female colleagues to get where they are than it was for you. If you teach or are in a class with an under-represented minority (female or any other kind), try to notice what makes them uncomfortable (staring, jokes about getting back to the kitchen, attributing any differences or gaps in knowledge to their gender, questioning their work more than you would a male colleague’s, staring… ), and stop doing those things, and talk to/encourage others not to do those things. The overall solution is complex, but the steps to get there are so simple a lot of people don’t realize that they need to be taken.

        • Henk Doorlag

          1.1) social mobility is increased by lowering education cost (the american dream is alive and well in Denmark…)

          1.2) success in a meritocracy is partly determined by genetics. A good CEO maybe good partly because of genetics, a good bricklayer might be good because of genetics. To my knowledge, science has shied away from such eugenic research.

          2) true – but see my point 1.2

          and you misunderstand my point (after I misunderstood yours, i guess), lower cost education does certainly correlate to more low-income household offspring getting higher education. Lower cost higher education does not change the correlation between education level and income. High income individuals still tend to be highly educated.

          3) that is just a mess of non-sequitur.
          3.1) “can be regulated, by individuals changing their behaviour” – individuals changing their behavior is not regulation. Regulation is a group changing everyone’s behavior.
          3.2) “continuation of non-egalitarian practices”. are you talking about equality of opportunity, or equality of outcome. Because I think we achieved the opportunity one, and the outcome one is undesirable.
          what “non-egalitarian practices” are you talking about?

          3.3) We are all a minority of one. He has the wrong hair color, she has the wrong ethnicity, and the third one has a funny accent. Women seem to have broken through this… uncomfortable environment in many fields, as gingers, negro’s and hillbillies have. we are all patronized, marginalized and misunderstood at times – and no, not all of us equally yet (unfortunately). However
          The world doesn’t owe you comfort. The world is not responsible for your happiness; you are.

          Since you do support STEM support, but I didn’t hear you talk about similar regulations towards construction or transportation I’ll assume you’re just serving your own interest.
          if you really think men have it so much better, have a read: http://www.amazon.com/Self-Made-Man-Womans-Year-Disguised/dp/0143038702

          • Tegan Maharaj

            I never said ‘men have it so much better’ or that ‘women are oppressed’. Yes, things are much better than they have ever been, but that still doesn’t mean equal. My entire point was that, there historical and cultural artifacts that make it harder for women to get in to science, and that this is changing gradually by open-minded people recognizing that social change doesn’t happen overnight. You are defensive, but I am not attacking you. You didn’t read a single reference I sent you, and support your arguments based on how easy it is to keep doing what you’re doing.

            I hope that you are more supportive of your daughter than you seem to be of people on the internet who tell you they have had a tough time. I hope that you acknowledge her problems when she tells you she has them. I hope that by the time she is grown up she doesn’t have to worry about any of this.

  • Olsonic

    Randy,

    I highly highly suggest the Steven Pinker, Liz Spelke debate available on Harvard’s YouTube page. In it, I believe Pinker successfully argues about several innate gender differences (such as the ability to mentally rotate a 3D object) that affect the proportion of men and women in physical sciences. There is a lot more to offer, outside of these, but its probably the best talk available on gender in STEM

    https://www.youtube.com/watch?v=-Hb3oe7-PJ8

  • Mihai Cosma

    Isn’t an obvious application to calculate the gender pay gap due to choice of major? I got 13% using a simple weighted average

  • garg

    “Men are paid more because more quantitative because companies value that” is circular reasoning. How do we know that companies “value” quantitative jobs especially highly? Mainly because they pay more for those services. *Why* they pay more is still completely unanswered. Also, the excellent pay for nurses flies in the face of your hypothesis; nursing is hardly quantitative.

    I would suggest we fall back on simple supply and demand. Underemployment is higher among more female dominated majors; even if your simple linear correlation gives a low R^2, we can see from the graph that the points scatter much higher once we exceed 30% or so female. A bit more underemployment could be amplified by competition dynamics or waves of demand. This can explain the nursing outlier better – nurses are highly paid because of rapidly increasing demand driven by the aging population among other factors.

    • Emil Kirkegaard

      Actually, the correlation is .30 (which gives R2 of .09), so it is not weak. It is medium-sized. https://en.wikipedia.org/wiki/Effect_size#Pearson_r_or_correlation_coefficient

      It is larger than most results found in social science, which is somewhere around .20. Richard, F.D., Bond, C.F., Jr., & Stokes-Zoota, J.J. (2003). One hundred years of social psychology quantitatively described. Review of General Psychology, 7, 331-363.

  • Fred

    “It seems plausible that male-dominated majors are usually paid more because they are more quantitative in nature, which large companies tend to value highly.”

    I think you’re dancing around the real reason. It’s not that quantitative jobs are valued more. The higher earning jobs tend to *scale*, and are in *profit* industries.

    You have 1 elementary educator, and hire 9 more, you now have 10x more elementary education capacity. Neat.

    You have 1 software engineer, and hire 9 more, you now have more than 10x the capacity to do whatever it is your software does. Maybe more like 2^10. Maybe even more, because one of those might have been a really great software engineer, and a great engineer can put up huge productivity gains compared to the median. And unlike your elementary school, that translates directly into more profits.

    No matter how great your elementary teacher or psychologist or nurse is, they’ll never be able to double the productivity of your entire organization, and even if they could, that would never translate into double the profits for your bottom line.

    The top male-dominated jobs all have this feature. Mechanical engineering, chemical engineering, mineral engineering — all fields where one person can generate big profits, well beyond a linear scaling of their time. On the flip side, nursing, psychology, and education are all fields where it’s very difficult for increased personal productivity to translate into organizational profits, so it’s hardly surprising that they get paid less.

  • Fraga123

    “Why do American women continue to choose low-paying majors?”
    TIFIFY

  • Jay Phillips

    You waste your time arguing with progressive feminists. They actually believe that there are no significant physical differences between men and women, that women can perform the same jobs as men, and they believe that jobs that require logical thinking, like STEM, are sexists. They actually believe that a teacher with no math or science skills who basically does nothing but babysit other women’s children while they go off to their careers should be paid as much as someone who designs iPhones or programs computers or builds machines. Women for the most part choose jobs that are easier. Easier jobs pay less than hard jobs. Men prefer jobs that are challenging and more stressful. Those kinds of jobs pay more. Women who work in the same fields as men take more sick days, more PTO days, work fewer hours, work fewer weekends and holidays so they can devote more of their time to their families. Men work longer, get more training and experience, and thus are more likely to be promoted and earn more than women. Do we need to spend thousands of hours and millions of dollars on research to determine these obvious facts? I guess to people who don’t think logically and believe that science is sexist we do.

  • Juliann Shih

    Hey Randy,

    Here’s a dataset you could work with (takes some manual curation though so I’ve been too lazy to): http://www.bls.gov/cps/cpsaat39.htm (Median weekly earnings of full-time wage and salary workers by detailed occupation and sex).

    Given the additional level of selection when entering the workforce (e.g. work-life balance) beyond college, you’d probably expect the correlation between gender ratio and income to be even more pronounced. At the same time it’s difficult to deconvolute other reasons for the pay gap.

  • Maddog

    Hi Randy:

    I posted a reply to this interesting post on my blog maddogslair.com, I excerpt it here.

    I have long noticed that “male” jobs pay more than “female” jobs. Olson confirms this suspicion.

    “The trend that’s immediately apparent from this chart is that female-dominated majors make less on average than male-dominated majors. Some interesting exceptions to the trend are Nursing (90% women; $48k median earnings) and Transportation Science (12% women; $35k median earnings), where Nursing especially stands out as a relatively lucrative major despite being primarily women.”

    Olson then crunches numbers busting some myths until he finally crunches these numbers.

    “Perhaps this all makes sense now: It seems possible that male-dominated majors — such as Engineering, Physics, and Computer Science — earn more than female-dominated majors because male-dominated majors are often more quantitative in nature. These quantitative majors are often employed by large companies to design products, perform data analysis, manage the company, etc., and their salaries are higher to match the responsibilities of the job.

    It’s another question of whether businesses and governments should value the services provided by quantitative majors more than, say, Education majors, but I’ll leave that discussion for another day.”

    Provided the markets for employment are actually free, the difference in wage is due to the fact that employers must pay a higher wage for some of the jobs, but do not have to pay such a high wage for other jobs.

    Olson does a good job here but I think there are a few more things happening.

    Men and women want different things out of work. The same can be said for the college majors they choose. Women generally want more security, men generally want more opportunity, men are willing to take far more physical risks, and work more physically demanding jobs, as well as work longer hours, take fewer and shorter vacations.

    http://www.aei.org/publication/blog/carpe-diem/

    http://www.aei.org/publication/new-bls-report-womens-earnings-17-9-gender-pay-gap-2013-explained-age-marriage-hours-worked/

    The wage gaps, and the analysis discussed above are not exactly on point with the question asked by Olson but they offer some insight into what seems to be happening. Men and women seem to view work differently. Men choose majors which will lead to long term careers, with more opportunities, and which will pay well, they seem to view their purpose as to make more money. Women simply choose majors differently, likely because they view their purpose differently than men. Women also want a career, but they appear to view their purpose as raising a family, and secondarily as a career. The choices each makes are understandable.

    So, men are more quantitative than women, willing to work longer hours with less vacation time, at more dangerous, difficult and more physical jobs, and place themselves at the command of the employer more readily, and they are unlikely to leave the workforce to rear children. Woman on the other hand are less quantitive, want more flexibility in the job, and more security, with fewer physical discomforts, and a greater ability to leave to rear children and later return.

    What is going on is a self selection process where each sex is generally choosing jobs which maximize their goals, the more “male” jobs pay more because it is much more difficult to find high quality workers for these jobs, the people who take the jobs work longer, under more difficult conditions, and with fewer breaks in work status, while the “female” jobs pay less because they are more flexible, require less work time, offer more work status breaks, etc. We also have to remember that this process is happening early in lives of the individual, and as such they are likely to want to maximize their opportunities for both work, career, marriage, family, etc. This will make men even more likely to shift towards earning careers, and women to shift towards secure, flexible careers.

    Why we are surprised when people maximize their economic potentials, and outcomes is unclear. We find it difficult to understand why, as a group, men or women would choose a particular career path, yet we do not find it difficult to understand why a couple with a large family would choose a Chevy Suburban as opposed to a Chevy Corvette, while a couple with no children are might choose the Corvette over the Suburban. This commonly goes so far that childless couples who intend to have children will purchase a large vehicle even if they intend to wait on family for a few years.

    Thanks,

    Mark Sherman

About this blog

This blog is my labor of love, and I've spent hundreds of hours working on the projects that you'll read about here. Generally, I write about data visualization and machine learning, and sometimes explore out-of-the-box projects at the intersection of the two. I hope you enjoy my projects as much as I have.

If you would like to use one of my graphs on your website or in a publication, please feel free to do so with appropriate attribution, but I would appreciate it if you email me first to let me know.

Archives

Subscribe

Enter your email address to subscribe to this blog and receive notifications of new posts by email.