Open threads at the Open Thread tab every Sunday and Wednesday

2D:4D Ratio And Psychological Traits: Results From The LW/SSC Survey Sample


2D:4D ratio is the length of someone’s index finger divided by the length of their ring finger on the same hand. It seems to correlate with some kind of prenatal hormone exposure, which makes it a unique way to explore the effects of purely-biological factors on various traits without having to do hormone assays or go through an ethics board.

Past research suggests that men generally have lower 2D:4D ratios than women, and that within each sex people with various stereotypically-male characteristics have lower 2D:4D ratios than people with various stereotypically female characteristics. For example, aggressiveness, penis length in men, and lesbianism in women have all been correlated with more masculine 2D:4D ratio. But there have also been other studies that challenge some of these results, and the whole field is maddeningly inconsistent.

Maybe the weirdest paper along these lines is Madison et al’s 2014 paper showing that feminist activist women have more masculinized 2D:4D ratios than other women, which suggests something about how “neurological gender” influences the way people respond to social gender roles. Or something.

I wanted to look at this more, so I included questions about digit ratio in the 2014 Less Wrong Survey and in the 2014 Slate Star Codex Survey. Both surveys also included demographics questions, questions about sex and gender identification, and questions about political beliefs.

Methods and Results

407 people on the Less Wrong Survey and 122 people on the Slate Star Codex Survey gave plausible digit ratio measurements (I rejected measurements outside the range of 0.8 – 1.3 as implausible), for a total of 529 people. Of these, 454 were biologically male and 75 biologically female. Although participants gave digit ratios from both their right and left hands, the differences were not significant and I averaged them together to create a less noisy measure.

The average male digit ratio was 0.972; the average female digit ratio was 0.975. The difference was not significant, didn’t trend toward significance and actually was the opposite direction on right vs. left hands.

Average Bem Femininity was 42.2 for men and 45.8 for women. Average Bem Masculinity was 42.3 for men and 40.5 for women. The sex difference in femininity was significant (p = 0.01) but masculinity wasn’t (p = 0.19) though it trended in the expected direction.

On a scale from one (strongly disagree) to five (strongly agree), the average male opinion of feminism was 3.4; the average female opinion was 4.1; the difference was significant at a p < 0.001 level. Other political opinions also had gender differences, always with women more liberal, but the gender difference in feminism was the strongest.

In a subsample limited to men, digit ratio did not correlate with Bem sex roles at all, but did correlate with positive opinion of feminism (p = 0.001). Bem sex role did corelate with feminism; more masculine men had lower opinions of feminism (r = -0.277, p < 0.001) and more feminine men had higher opinions of feminism (r = +0.277, p < 0.001).

In the opposite subsample limited to women, digit ratio correlated strongly with Bem masculinity (r = -0.381, p = 0.005) but not with Bem femininity (r = +0.123, p = 0.38). Opinion of feminism did not correlate with either sex role or digit ratio and there wasn’t even a real trend.

I looked at eight other political issues not clearly related to gender. None of these had a clear connection with digit ratio the same way feminism did. One issue, immigration, had a single correlation significant at the p = 0.05 level, but was likely a false positive.

There was a limited sample of transgender people in the study; because most participants were biologically male I compared cismen to transwomen. Difference in digit ratio was not significant and showed no clear trend. Bem sex role inventory found transwomen to be much less masculine and much more feminine than cismen (and in fact less masculine and more feminine than ciswomen), but because of the low sample size the trend didn’t quite reach significance. Transwomen were more positive towards feminism than cismen (p = .047) but not quite as positive as ciswomen. I tried something similar with straight men versus non-straight men and the trends were arguably in the expected direction but didn’t come close to significance. There weren’t enough biologically female people to be worth running the analyses with that sample.

A binary autism variable (does vs. does not identify as autistic) had little correlation with sex roles or opinion of feminism, but approached significance for digit ratio in men (p = 0.051) and achieved significance once men and women were combined (p = 0.047). Autistic people had slightly lower (more masculine) digit ratios. I am wary of this result since it was so weak and not replicated on the left hand.


This was weird.

The whole point of digit ratio work is that men are supposed to have lower digit ratio than women, and then you use that to determine whether other characteristics are associated with more male-typical or female-typical hormone balances. But we couldn’t even get the basic finding of men being more male-typical than women.

In fact, our entire sample was heavily feminized compared to those of every other study; our men had more feminine digit ratios than other studies’ women. This is really weird because it’s a sample selected for very high mathematical ability – average IQ of 139, many math and physics PhDs, computer-related jobs as the most common occupations – but mathematical ability is usually linked to more masculinized digit ratio. Likewise, presumably this sample’s much more autistic than normal, but that’s also supposed to be more masculinized. I think maybe the measurement technique on the survey predisposed towards overreporting the ratio. Or maybe I screwed up somewhere and divided when I was supposed to multiply. Really that’s all I can think of.

Despite our failure to pick up what should have been the most drop-dead obvious finding, we still got strong signals on psychological traits that should have been much more subtle. The two clear results were a correlation between digit ratio and opinion-of-feminism in men, and between digit ratio and masculinity in women. P-values for both were very low and unlikely to be coincidences despite the multiple tests performed. But the pattern is hard to explain.

I don’t understand the biology of digit ratio very well, but it’s certainly only one part of sexual differentiation; after all, nearly all studies show wide overlaps between men and women in digit ratio, but except for a few intersex people the sexes still clearly differentiate anyway. So maybe women’s level of masculinity is determined mostly by the thing behind digit ratio, and men’s level of masculinity is determined by that plus various other male-specific masculinizing processes?

More annoying is the apparent partial-disconnect between digit ratios, sex roles, and feminism. In men, digit ratio affected feminism but not sex roles. In women, digit ratio affected sex roles but not feminism. The results weren’t ambiguous either. But why would digit ratio make men less feminist if not by making them less gender-conforming themselves? And is it really plausible that digit ratio makes women much more masculine without shifting their opinions toward feminism at all? And remember, this whole area was started by someone who found a connection between digit ratio and feminist activism in women. While it’s possible that the original study was really finding that masculine-digit-ratio women were more activist rather than more feminist per se, that seems like kind of a stretch. I’m going to blame this one on low sample size and high measurement error until somebody forces me otherwise.

In conclusion, this study was a mess, but somehow managed to find clear signals in weird places anyway. I don’t know.

If you want to see for yourself, you can find the public Less Wrong survey data here and the public SSC survey data here. You could also get my haphazard combination of the two here but I would warn against that if you’re really trying to double-check the results as the combination process was one of the most potentially error-prone steps and I’d rather see it independently replicated.

This entry was posted in Uncategorized and tagged , , . Bookmark the permalink.

230 Responses to 2D:4D Ratio And Psychological Traits: Results From The LW/SSC Survey Sample

  1. Alsadius says:

    Proposal: Any study of laypeople asking about digit ratio must, by Internet Law, include a Youtube video showing how to properly measure it. I answered the LW survey, but I wasn’t sure what I was doing, and just guessed as to what proper technique was. I mean, I *think* it was a reasonable guess, but I have very little confidence in my results given that I had a ratio below female-typical, and I’m a pretty masculine dude in most respects.

    • Evan Þ says:

      +1. I answered the SSC survey and just eyeballed it looking at (IIRC) the back of my hand. If I’m anywhere near typical, I’m astonished you got any signal at all.

      • Rachael says:

        Not much point analysing the data to 3 significant figures if people are just eyeballing it.

        I didn’t answer the question because it looked a lot of effort and I didn’t have access to a photocopier.

        I wonder if there’s a correlation between gender and ignoring the instructions? :p

        • DensityDuck says:

          Also it’s not correct to measure something to two decimal places (hundredths of a centimeter) and then report results to three decimal places.

          • pngwn says:

            Just measure it in millimeters and convert to kilometers.

            Problem solved.

          • Peter says:

            I’m not sure actually. If you’re taking an average, then rounding error is going to introduce some noise but averaging is going to take some of that noise out again.

            I mean, suppose I have a source of random numbers, with a mean of 1.014 and a uniform distribution from 1.009 to 1.019. Suppose I sample lots of those numbers, round each to two digits, and and average. I should get more results saying 1.01 than 1.02, so the average should be below 1.015. In fact, if I did another run with a batch of numbers, mean 1.013, range 1.008 to 1.018, if the sample sizes were large enough, I could show with a t-test a statistically significant difference between the two distributions.

            I think it would work with gaussians and other distributions as well as uniform distributions, but I’m less sure about that. I mean, I think I can prove the initial bit with uniform distributions (I don’t think I can prove the bit about the t-tests), but the maths is harder for gaussians.

          • DensityDuck says:

            An investigator could certainly deal with variation by taking a sufficiently-large sample and averaging, but the instructions don’t say “measure every possible definition of ‘length of your finger’ and use the average”, they say “measure the length of your finger”.

          • Bassicallyboss says:

            In physics, at least, we just include the uncertainty, or report to the last certain digit. For an average, then the uncertainty (standard error) is just the standard deviation of the mean*, so I’d give the error as [average] +/- [SD of mean], both rounded to the first digit of the uncertainty. Or if there’s some reason not to use the average, just report to the least significant certain digit.

            In terms of significant figures, if we had an average of .948753 and an uncertainty of .00027548, then you’d report it as
            .9488 +/- .0003 or just .949. If the first digit is 1 or 2, it’s acceptable to put 2 digits of uncertainty–just make sure that the value and uncertainty are reported to the same decimal place.

            I don’t know what the conventions in social or biological sciences are, though.

            *This is only true with two assumptions:
            1) Every measurement is equally uncertain. A weighted average is a better value if we trust some measurements more than others, but since we don’t have reason to here, we can grant this assumption.
            2) We have good reason to expect that we’re sampling from a Gaussian distribution. I’d say that’s true for the male and female subgroups, but not true in general–we expect something bimodal-ish here. I don’t know how to do it for a non-Gaussian though. Since there’s so much overlap, I’d probably just do it this way anyway, since the overall average isn’t terribly important.

          • Peter says:

            OK, a few things.

            1) Decimal places are a red herring – as pngwn says, decimal places depend on your choice of unit. You don’t even need to convert – for digit ratios you could measure your fingers in Planck lengths or light years and the ratios would still come out the same. But obviously significant figures are the real issue here.
            2) Technically a few of the numbers spill out into four significant figures, but whatever.
            3) The numbers that Scott is quoting are averages. The instructions may indeed just say, to each individual, “measure your fingers”, but in the end lots of fingers ended up getting measured. Nowhere did I imply that one finger was being measured multiple times. All that is needed is for there to be a distribution of digit ratios, and for that distribution to be sampled multiple times – i.e. by measuring multiple fingers from multiple people.
            4) Based on Basicallyboss: Presumably by “standard deviation of the mean” you mean “standard error of the mean” – i.e. if you calculated the mean repeatedly by resampling, the standard deviation of that. This is equal to the sample standard deviation divided by the square root of the sample size. I had a Excel tab open with just results from the LW survey. I got right-hand values, finding a sample standard deviation of 0.046, 399 data points, so a standard error of 0.002. The sample mean was 0.971, so it seems that, for that sample, quoting to 3 significant figures appears to be correct. For particular subsamples it may be that only two or fewer significant figures are appropriate.

            Sampling and averaging can be great for dealing with noise. No matter how much noise there is, you “can” cancel arbitrary amounts of it by using a large enough sample (getting a large enough sample, OTOH, may be a problem, especially if “large enough” is bigger than your population). On the other hand, it won’t help with systematic errors. One of the worries about slapdash digit ratio measurement is that it has systematic error (or worse – systematic errors that correlate with variables of interest), and a big sample won’t help with that.

          • Bassicallyboss says:

            Yes, that’s what I meant. I didn’t realize “Standard Error” was the more common term.

            And agreed absolutely about systematic error. I would normally think that since the measurement technique was so inconsistent here, the random error would dominate. But I know that before I learned the proper technique was to photocopy and measure from the bottom crease. Had I not known, I probably would have measured both fingers about a half-centimeter higher, and decreased the ratio correspondingly.

            Based on quick estimates, that’s an error of ~.5cm of a finger length of ~10cm. If the fingers were actually length 9.71cm and 10cm, adding .5cm to each would add about .0014 to the ratio. Propagating that through the average, we’d expect a systematic error of ~.0014/399 = .0000035. This is 2 orders of magnitude below the random error and therefore negligible, but it’s not likely that everybody erred in the same direction as me. But how large a systematic error would it take before it surpassed the random error? By the same calculation, it looks like it would take an error about the length of a whole finger for the systematic error to overtake the random error. I think we can safely ignore any systematic error that affects both fingers equally.

            Index and ring fingers look different at the base, though, and perhaps only one is mis-measured, or they both are but in opposite directions. I don’t want to figure that out right now, but it would probably be a much larger effect than the one I described above.

    • Viliam says:

      So the confounder of this study result could be something like “feminism correlates with watching youtube videos about gender-related topics… which increases the chance of measuring the digit ratio correctly… which is somehow also connected to gender because… uhm… maybe because some of the male readers already participated in a study about penis sizes… and therefore… [insert your favorite conclusion]”.

    • I followed the directions, but it was not convenient to go to a xerox machine.

      I suggest adding “how carefully did you follow the digit-measuring directions?” to the upcoming survey.

      • HeelBearCub says:

        There were directions when you did it?

        • Deiseach says:

          You people are like my mother – she never bothered reading the instructions either 🙂

          • I totally read the instructions!

            … they were complicated and a significant amount of trouble since I didn’t have a photocopier handy. Also, it was about the second or third post I’d read, and I wasn’t very invested yet. So I left that bit out. >.> In hindsight kinda regret that.

  2. Outis says:

    Did the survey describe a precise method for measuring the length of the fingers? The picture on the Wikipedia page has the lines not following the longitudinal axes of the fingers, which seems weird to me. And what is the base reference? The skin crease? The space between the fingers?

    On my hand, different methods result in a difference *in ratios* of more than 0.05, which is more than double the difference between the male and female means given on Wikipedia!

    • Scott Alexander says:

      The survey quoted the University of Cambridge’s instructions:

      1. Place your right hand firmly on the plate of a photocopier or scanner with fingers straight. Close cover of place a sheet of paper over your hand to prevent glare from overhead lights. Ensure that the bottom crease and finger tip can be clearly seen in the photocopy.

      2. Use a ruler or calipers to measure the distance from the middle of the bottom crease to the tip of the finger to the nearest hundredth of a centimeter.

      3. Once you have the measures for both your ring and index finger, then divide the length of your index finger by the length of your ring finger. The result is 2D:4D (2nd digit divided by 4th digit).

      If possible, please give three digits – for example, 0.915. Some people may have digit ratios slightly greater than 1, which is okay.

      I’m skeptical of too much survey error, because if the results were completely meaningless we wouldn’t have gotten the couple of really nice correlations we did.

      • William O. B'Livion says:

        > I’m skeptical of too much survey error, because if the results were completely
        > meaningless we wouldn’t have gotten the couple of really nice correlations
        > we did.

        Yeah, because that couldn’t have been Lt. Murphy kicking sand in Claude Shannon’s face, right?

      • keranih says:

        I’m skeptical of too much survey error, because if the results were completely meaningless we wouldn’t have gotten the couple of really nice correlations we did.

        Now see, I really don’t see how that follows.

        • Izaak Weiss says:

          Sure, but that’s why we have p values. p=.001 means that we should expect 1 spurious correlation for each thousand examinations. I count four correlations with p <= .001. That means that we should expect Scott to have done 4000 analyses, if these are all spurious. I doubt Scott did 4000, and if he tells me how many he did do, I'll calculate the probability that all four are spurious.

          • et.cetera says:

            “p=.001 means that we should expect 1 spurious correlation for each thousand examinations”

            Only if that is what p values actually mean… But it ain’t.

      • Deiseach says:

        Since I have nothing better to do at work and a photocopier is conveniently to hand 🙂

        Right hand:

        Index finger – 6.6 cm
        Ring finger – 7.0 cm

        Ratio: 7.0/6.6 = 0.943

        Left hand:

        Index finger – 6.7 cm
        Ring finger – 6.8 cm

        Ratio: 6.7/6.8 = 0.985

        Measurement error probably included, because I don’t have a calipers and you try measuring to the finest degree with a crappy plastic ruler on a black-background photocopy.

      • DensityDuck says:

        I can get a +/-.5 variation in the digit ratio just by changing which side of the digits I measure, which is miles beyond the supposed sex-linked variation.

        So, yes, measurement error is a huge factor in this one.

        • NN says:

          If you think you can choose which side to measure, you are not using the procedure described in the comment you’re replying to.

          • Anthony says:

            No, but how do you determine the middle of the crease? On my index fingers, it’s pretty obvious where the crease is, and the crease is pretty straight (though it’s diagonal to the axis of the finger), but on both my ring fingers the crease is curved, and there’s a secondary crease about 5mm closer to the tip than the main crease. And all the creases are close to a millimeter wide – how the heck are you supposed to get 0.1mm precision?

      • Jeremy says:

        A couple issues that I considered when I filled out the survey:

        1. What does “straight” mean? Does it mean all fingers are parallel? Does it mean all fingers are touching together as closely as possible? Does it mean all fingers are not bent from their relaxed position? Does it mean all fingers point in the same directions as the tendons attached to them?

        2. What does the “middle of the bottom crease” mean? My bottom crease is not a single line, it is composed of many smaller creases, which span a distance of about a quarter inch. If I choose the center of that region vs the very bottom of that region, that will be a difference of an eighth inch, or about 5% change in the measurement. Even the individual creases are wide enough to affect the measurement by another 1%.

        Considering I ended up with an answer in the 99th percentile, I probably did it wrong, as I suspect many other people did.

        • Tibor says:

          I did not participate in that survey, but I’d have quite a bit of trouble with that. Both of my index fingers are slightly bent towards the other fingers of their respective hand so while all other fingers line up perfectly, leaving zero visible space between themselves if I straighten them and put them together, there are small “holes” between my index fingers and my middle fingers. So, given that my index fingers are slightly bent, I don’t know in what way I would calculate their length (they are not curved that much, but since the difference between the male and female digit ratios is so small, it could introduce too much noise).

  3. suntzuanime says:

    Is it possible that masculine digit ratio is correlated with deciding “to hell with all that effort” when instructed on measuring digit ratios? Are there any interesting correlations about who answered the question vs. who didn’t?

    • brad says:

      I came to that question in the recent gender attitudes survey, never having heard of finger ratios before. My first thought was “are they going to ask about the bumps on my skull too” and then I saw the elaborate directions I was supposed to follow to get the ratios and went nope, not going to happen.

    • Saro says:

      This! Based in my real life interactions (all highly masculine including the women), no-one bothered with the ridiculously complicated instructions. It must take a certain kind of person to actually do it…

      Someone please do this analysis!

      • Tibor says:

        This, at least superficially, goes along with the “common knowledge” that women are much more likely to go to regular doctor check-ups whereas men usually don’t bother with those.

  4. Sonya Mann says:

    I’m kind of stunned by how many more men than women responded (which likely indicates at least something about the readership). I would have guessed maybe a sixty-forty split, not this drastic. It’s a little disturbing.

  5. Pku says:

    Are there any studies on cases where different hands have different digit ratios? (Mine are visibly pretty different). Going by the hormone-exposure theory it seems like that shouldn’t matter (the hormones would probably affect both hands equally, then random noise would take care of the difference) – But it does seem to imply that measuring the difference between the hands might give you a good measure of the noise level in the experiment.

    • Deiseach says:

      This digit ratio thing is chiromancy (not so) pure and simple. While we’re measuring our length of ring versus index fingers, we could also indulge in a little chirognomy, too?

      Difference in digit ratios between hands could be difference between dominant hand and other hand, and chiromancers always read the dominant hand.

      So gentlemen (and with 454 male to 75 female, I feel confident in addressing you by that term), what is your hand type?

      Conical, square, spatulate, pointed, knotted, or mixed?

      Now I really want to see those penis length answers, I could use a good laugh 🙂

      • Anthony says:

        Now I really want to see those penis length answers, I could use a good laugh

        If people didn’t follow the directions for finger length, what makes you think they’ll follow the directions for penis length?

  6. Alraune says:

    I realize this isn’t a real study, but the “approached significance”s still make me wince.

    Anyway, I’d assume we just all screwed up the measurements. If we didn’t screw up the measurements? You found another trait your readership are severe outliers on, congrats.

    maybe women’s level of masculinity is determined entirely by the thing behind digit ratio, and men’s level of masculinity is determined by that plus various other male-specific masculinizing processes?

    That’s not utterly implausible, but before that I’d look into “N=10 White Male College Students” effects in the original study: that connection between digit ratio and feminist activism may have been valid, but only held for a particular population, a particular time, and a particular movement in feminism.

    • Peter says:

      Throw in a Bonferroni correction or one of those related controls for False Discovery Rate or Familywise Error rate or whatever and it doesn’t approach significance. Otherwise you’re straying dangerously close to the land of p-hacking.

  7. Speaking from personal experience, highly intelligent, mathematical men are effeminate. Highly intelligent, mathematical women are masculine.
    Feminism is more popular with women, period, and your survey did not include a sample of less-masculine women who could be compared to masculine women, just because those kinds of women don’t read your blog/LW.
    Among men, masculine men react more antagonistically to feminism’s anti-male messages and value traditional masculinity. Effeminate men try to placate feminists.

    • …and if effeminate men have life experience that causes them to perceive feminists as implacable – which I suspect may correlate with the typical life experience of highly intelligent individuals – what will their reaction to feminism be? That’s what I think is going on in these results.

    • Deiseach says:

      Okay, let’s stir the pot here.

      What exactly is meant when we’re talking about effeminate men and masculine women?

      Delicate boys who look a bit girly and limp-wristed versus gruff-voiced women with hairy legs?

      Or do we mean ‘not exhibiting stereotypical traits of gender roles, e.g. men are all reason, emotional control, logic while women run on instinct and burst into tears at the drop of a hanky’?

      Because I don’t think intelligence, courage and honour are purely masculine values and compassion, tolerance and sensitivity are purely feminine virtues and neither sex shows or shares the traits associated with the other.

      • True – the best approximation is that people have two independent sliders. Push both the F and M slider high and you get Jacko or a hot MMA fighter chick, have them both low and probably you get something really similar to depression.

        I mean, the female body does make a lot of testosterone too, just aromatizes most of it to estrogen.

        So low T alone just makes a man who is less of a man, but there is still nothing womanly in him. Effeminacy is not an accurate term, but it is just part of popular culture to use simple contrasts, if you aren’t manly then you are like all the other not manly people, who are women. But it is not accurate.

        Not low T alone, but low T and high E(strogen) makes a woman.

        This really explains hot yet brave women – high T, high E.

        Sensitive men are sort of more difficult to explain. Note how estrogen has just about nothing to do with compassion or sensitivity and especially not with tolerance. It’s primary psychological effect is as far as I know is these typical pregnant mood swings and food cravings. Pregnancy is NOT a period characterized by high tolerance of putting up with the husbands shit 🙂 And those mood swings are something you will not find in a low T man.

        It is more likely that tolerance or compassion are traits that are simply suppressed by T, and not caused by its absence.

        Of course, it is really easy to be a little bit afraid and rationalize it as being nice (compassionate, tolerant). And afraid could be low T.

        Also, I think I am engaging in primitive and vulgar speculation now, teenage magazin psychology. I mean, you were just about to say that anyhow I guess, so I figured I’ll say that I know 🙂

        • Deiseach says:

          Well, I think if we’re going to be throwing terms like “effeminate” and “masculinised” around, we really need to get some definitions going, otherwise we’re reduced to “glasses wearing nerd who gets punched in the face by the real manly guys because he’s too wussy and nine stone weakling to defend himself”.

          Definitions of “what makes a real man” and “how does a real man behave?” have changed over the years; if a ‘real man’ is an aggressive loudmouth who gets in fights, then all the ‘real men’ are in prison serving sentences for common assault.

          On the other hand, to quote “The Three Musketeers”:

          To be obliging and polite does not necessarily make a man a coward. Look at Aramis, now; Aramis is mildness and grace personified. Well, did anybody ever dream of calling Aramis a coward? No, certainly not, and from this moment I will endeavor to model myself after him.

          Or Sir Gawaine, who in “The Green Knight” faces the test of courtesy to his host’s wife, which is just as much part of proving his knightliness as being willing to get his head chopped off – and a lot trickier to navigate, between the risks of showing himself a churlish boor by insulting the lady in the game of courtly love, and dishonouring his host by taking the stylised flirting too far and being in danger of adultery.

          • Look if we want to avoid all cultural stuff, then it is really just hormones. So we should not see it as “I think politeness is masculine, now what causes it” but more like “masculinity is defined as something created by androgens, now politeness could be one of those things?” And I think yes, if armed societies are polite soceties, more martial minds are more polite for the same reason.

            We understand how testosterone is closely associated with status-seeking that they even uses it to test how stereotype threat affects girls.

            We also know there are two kinds of status, dominance and prestige and we know on one hand T makes one pursue them more and on the other hand winning points in either increases T probably generating a loop.

            I would say everything closely related to exclusve masculinity is probably dominance. It is the guys who punch the nerd, it is the inability or unwillingness to defend against them, it is the muscles, it is the guns, it is the harley bikes, it is the pit bulls, it is every 2pac type gangsta hip hop song ever, it is every Turisas type viking metal song ever, it is every man ever who asked another one “what are you looking at?!” or “are you dissing me?” and so on. Do I have to go on? It is basically everything one associates with the lower class men. It is obvious. It is non-obvious only if you stick to upper class people only and perhaps only in the US because e.g. in Britain there is a certain lovely tradition of making upper class kids play really brutal sports in order to avoid that smart wuss effect, rugby, boxing etc. And in your case, it is probably four out of five male social housing clients so you probably know this very, very well. And this tends to be incredibly exclusive to men.

            The other hand, prestige status, is a tad bit more difficult. The best way to approximate it is art, as most art, especially most of this modern contemporary abstract fsckery called art these days is just crystallized, distilled prestige status. It is simply cool and it has almost nothing but coolness in it.

            And while most artists through history were men, clearly they often tended to towards the more effeminate or Leonardo gay, women always seemed to really love art, probably the original function was men buying fine artworks to impress their wives mostly, as women to like everything nice and fine, and probably they reason it used to be a male domain because it was used to impress women, think troubadours, but once women were allowed to participate they really quickly got into it and now it looks like there is some adult art class in the college it is 90% women. Art seems to be very well fitted for women and thus it seems like seeking status through prestige is fitted well for women. Amazons or shieldmaidens weren’t real but Sappho’s poetry was real, that sort of thing I have in mind. But it still requires testosterone to really be hardly driven for prestige status. Still, so much about femininity is really prestige status. Of course there is a lot of male prestige seeking too, but it is an upper class thing. In the upper classes they say saying X makes you a bigot, in the lower classes they say saying X makes you a wuss. This is prestige vs. dominance.

            So as a summary, the masculinity is more clustered around dominance status seeking, and it is something that happens far less in the upper classes, but prestige status seeking is also there, mostly in the upper classes. While feminity is both about reduced status seeking and about basically prestige only, not dominance. This means feminity is something that in the upper classes just looks like being normal. There is hardly any difference between Jack and Jill if they are both in the art class, the children of lawyers, and Jack is probably gay or at least the kind of super nice guy that would be called gay in the lower class pub anyway.

            Will that do for a beginning?

          • TD says:

            The glasses wearing nerd weakling stereotype should probably be considered unmasculine or tending towards neuter. This is reflected in how stereotypical nerds are often seen as asexual or have low sex drives. The nerd’s color is gray. His movements are stiff and awkward. Doesn’t talk much. His voice is weedy, nasal and monotone.

            Pardon the offense, but “effeminate male” brings to mind more the flamboyant homosexual stereotype, who is definitely not asexual and passionless. That character has womanish traits that stand out rather than neutralized ones. The stereotype gay color is pink or a rainbow. His movements are floppy and comically theatrical. Gossips constantly. His voice is weedy, lispy and with a wide pitch variance.

            So, there are two different pathways to failing masculinity stereotypically speaking. (In reality, you could be a gay nerd or a macho gay jock or fit into neither nerd or jock easily, or be anything under the rainbow, but this should go without saying, so I can’t believe I’m typing this, but I am anyway.)

          • Cord Shirt says:


            Up until 1990-1995 or so, the US popular mind conflated those two forms of male “effeminacy.” Revenge of the Nerds, 1984 In 1984, most Americans would’ve thought the character of “Lamar Latrell” was just another nerd.

          • Deiseach says:

            it is the muscles, it is the guns, it is the harley bikes, it is the pit bulls, it is every 2pac type gangsta hip hop song ever, it is every Turisas type viking metal song ever, it is every man ever who asked another one “what are you looking at?!” or “are you dissing me?” and so on.

            TheDividualist. and that is so boring. You are correct, these are a lot of the guys we see in social housing, and they’re not your guidos, they’re not smart guys with money who like the party party party lifestyle, they’re losers.

            I’m not impressed by fellas that can punch and posture, and the girls that are generally end up regretting it, one way or another.

            And there are also women who punch and fight and stab, and they’re doing it for dominance as well as prestige (the type who is the terror of the estate, who gets away with murder because the neighbours are scared she’ll break their windows or come after them in the street).

            And they’re losers, too.

          • @Deiseach

            > it is boring

            It is – and it can be very destructive. However, it at least cannot do one thing: masquarade as something totally ethical and holy. The problem is, prestige competition can, and it can outcompete actual ethics:

            Programmers have a saying – it is a good idea that potentially buggy code should also look aesthetically bad, so that it is easier to spot. Dominance can look bad enough to spot the bugs. Prestige cannot, this is why it is so destructive. That prestige signalling can entirely eat ethics while still looking cool.

          • Tibor says:


            One thing I’ve always had trouble understanding is why one has quite a moderate number of female singers in non-classical music but very few instrument players. All that while anyone is melting over a girl who can play the drums/guitar/bass even moderately well (they would not be so enthusiastic about a guy at the same skill level).

            I’m not sure what is so much different about using your voice as an instrument and using a piece of wood and metal as a instrument that one attracts women disproportionately less than the other. Also, both have to be practiced if you want to be any good and do not if you just want to play in your local punk band.

        • science says:

          Also, I think I am engaging in primitive and vulgar speculation now, teenage magazin psychology.

          Admitting you have a problem is the first step. Congrats!

          • 1. We admitted we were powerless over vulgar speculation—that our ~~lives~~ Internet debates had become unmanageable.

            2. Came to believe that a Power greater than ourselves could restore us to sanity.

            3. Made a decision to turn our will and our lives over to the care of God as we understood Him.

            4. Made a searching and fearless moral inventory of ourselves.

            5. Admitted to God, to ourselves, and to another human being the exact nature of our wrongs.

            6. Were entirely ready to have God remove all these defects of character.

            7. Humbly asked Him to remove our shortcomings.

            8.Made a list of all persons we had harmed by vulgar speculation, and became willing to make amends to them all.

            9. Made direct amends to such people wherever possible, except when to do so would injure them or others.

            10.Continued to take personal inventory, and when we were wrong, promptly admitted it.

            11.Sought through prayer and meditation to improve our conscious contact with God as we understood Him, praying only for knowledge of His will for us and the power to carry that out.

            12.Having had a spiritual awakening as the result of these steps, we tried to carry this message to other vulgar speculators, and to practice these principles in all our affairs.

        • Sarah says:

          If you are going to talk about hormones, talk about actual hormones, not metonyms for the gender archetypes in your head. For instance, progesterone and estrogen both rise during pregnancy. Does estrogen cause irritability? or is it progesterone? or is it, perhaps, having your back hurt and constantly needing to pee? Perhaps someone would have to do *experiments* to find out!

          It would be interesting to get a summary written up of the cognitive/behavioral effects of the sex hormones, but you cannot safely assume they do exactly what your stereotypes suggest.

          • Good point but if we study estrogen simply at puberty, we get another confounding effect of the subjects being teenagers/children. Perhaps study trans people. And you know what? Transititioning seems to show gender roles are pretty solid and hormonal. Get a T blocker, take estrogen, these are the most important aspects of transititioning M2F.

            Besides. Who made gender roles? Patriarchy. What is patriarchy? Male dominance. What makes people want to dominate others? Testosterone. What happens when people successfuly dominate others? They get more testosterone. The later two statements are objective facts, the first two are obviously up to a lot of social “science” and philosophical interpretation. But basically, if you claim patriarchy creates gender roles you are just saying testosterone creates gender roles in a roundabout social way.

      • Sorry for the late response. I’m going to have to be lame and wave my hands and say “It’s a vague thing.” I don’t mean “feminine” or “masculine” as an insult–mathy people are honestly the only people in the world I enjoy being around, not to mention the only ones I find attractive. So their relative personalities are things I value about them, not dislike. I really just don’t like traditionally masculine and feminine people.

        But if pressed: masculine men are large, loud, and overbearing. Their bodies are muscular squares. They like sports, cars, and punching things.

        Intelligent, mathy men, by contrast, are small, not loud, and shy/polite. Their bodies are lithe rectangles. They are more interested in math and science than in sports and cars, and they tend not to punch things.

        Feminine women are demure around men and vivacious among women, seek agreement rather than conflict, and pay a lot of attention to their hair, makeup, and clothes. They like fiction with lots of emotions and dislike numbers.

        Mathy women tend to be shy around normal people but happy among other mathy people (who tend to be male.) They care about the ideas in their heads, not the clothes adorning their fleshy parts, and will defend their opinions even in the face of discord.

        Obviously I am making really big, sweeping generalizations. Hopefully the general idea is clear, even if the words are imperfect.

    • Tracy W says:

      Speaking from personal experience as someone who went from a highly academic single sex girls school to electrical engineering (90% male), your claims are not borne out by my observation.

      • Alsadius says:

        My fiancee has precisely the same life path, and effectively the same experience.

      • unsafeideas says:

        You two might just have two different definitions of femininity and masculinity. Also, your definition of how highly intelligent person behaves may be different from his one, so you might assign intelligent label to different kind of person.

        My other observation would be that if your single sex girls school was high school (and not another major on college), its population is self selected out of mainstream bunch. The decision to go to single sex school already suggest opinions, attitudes and decision making much different from what overwhelming majority of population does.

        • Tracy W says:

          I’m a New Zealander. Single-sex schools are not that uncommon here.
          And my observation is that within the population of mathematical intelligent people of one gender there is massive variation in personality, so I don’t think this is a matter of different definitions of masculine and feminine.

          • unsafeideas says:

            I will remember the New Zealand and single sex school thing. I genuinely through such a thing practically do not exist anymore.

            I would agree with massive variations in personality. However, there were differences in behavior compared to art students. It is hard to describe what the difference is (not effeminate vs masculine nor necessary extrovert vs introvert nor anything easy like that), but the communication patterns were definitely different.

          • There is this stereotype about NZ that basically you just go there as a good looking man and then it is zero effort, women propose to you openly. True? Related?

          • Tracy W says:

            TheDividualist: you were misinformed. You don’t need to be a good looking man, you’ll get that effect purely because you don’t smell of lanolin.

            NZ: where the men are men and the sheep are scared.

          • Tibor says:

            When I finished my high school and went on to study maths, I was surprised by a couple of things:

            1) About 30% of the students were female. This remained more or less constant throughout both the bachelor and the master…with one twist – women overwhelmingly ended up choosing more “practical” subject, such as finance maths in their Ms. and very “theoretic” fields such as algebra were almost exclusively male (and finance math ended up with some 50/50 ratio) and fields such as numerics or statistics somewhere in the middle (although analysis somehow ended up attracting more girls than probability theory…but it attracted more people in general, so that might have been it). Now, finance maths is also easier (I mean the study programme we had there) than algebra. But there were female students who would not have a problem studying it (and low confidence was not a problem either, one of them had a stipend for excellent grades for example, hard to imagine she would feel like she was not good enough for algebra) but still chose the more applied stuff. On the other hand, really applied fields, such as electroengineering, are heavily male. Maybe boys like machines…And girls like money 😀

            2) There was only quite a low number of guys who would fit into the “nerd” stereotype (not many Sheldon Coopers). I have not met any aggressive guys there or anything of this kind of “masculinity” (but as was noted above, this probably has to do more with stupidity than masculinity). This is also why I have a quite an ambivalent relationship towards the “Big Bang Theory” series. It is quite funny (although much less so in the seasons 3+) but it also tells you “this is what maths and physics types are like” and that is just not true. At best, there is a higher incidence of these types in those fields than in (the students of) law or economics, but nowhere near even a sizable minority.

            Intelligent men may seem less masculine than stupid men because they are in fact more rational and able to control basic instincts. Likewise a stereotype of a stupid woman is an extremely feminine bimbo and an intelligent woman, while being feminine does not display these “overblown feminity” as often, because she knows better and is able to control herself more.

      • Personal experience is personal, of course.
        Though with all due respect, part of the difference in experience may be that I’ve attended schools that drew from a much larger pool of students. The entire country of NZ is about the size of a large US city; my highschool drew from a larger pool than this. (Then again, perhaps NZ schools are real magnets for international talent. I am not very well informed on the subject.)

    • unsafeideas says:

      It is possible that high intelligence is simply correlated with lower willingness to observe traditional gender roles/posturing among your group of friends.

      Stereotypical traditional femininity involves quite a lot, but also pretending to be dumber/weaker then you are because that is supposed to be cute. Highly intelligent women either wont do it, or will succeed and you wont consider her highly intelligent.

      Stereotypical traditional masculinity involves being dumb in other ways – being overly aggressive/tough when you gain nothing out of it and tons of pointless posturing. Again, highly intelligent guy will be less likely to do it simply because he has brains or will succeed and you wont consider him highly intelligent as a result.

      In both cases, there is also opportunity cost. Maintaining proper masculine/feminine image takes time and effort. It don’t just happen randomly for all that many people – no cultivated image does. It is quite possible that highly intelligent people simply choose to spend their time and effort differently.

      • “Streetwise, shrewd, cunning, foxy”. Terms like this sound like being highly intelligent in a way not approved by the higher prestige upper classes.

        So, you are highly intelligent, figure out trad roles bring you hotter sex partners at the price of getting frowned upon by the upper classes. As the upper classes define language – that is the whole point of being upper class – you are now seen as dumb, stupid. OF COURSE. THEY get to define who is stupid. That is the whole point of being upper class intellectual elites. But OTOH…. you are now sometimes seen as surprisingly “streetwise shrewd cunning” ? And yes, when you think about it, this actually seems like being a thing.

        This is part of why I use the term “guido” in a positive way. Besides that it is amusingly uncomfortable to US intellectuals, you could say many of the men are on (of course, you see far more guidos in Europe but the term is not used, it is just being “party guy”) are “shrewd cunning streetwise” – they are intelligent, they figured out partying and picking up hot girls is fun, they ignore the opinion of the elite and not care that the elites consider them stupid: they are having a good time, getting what they want, so why care. Apparently they can afford beachside houses, which suggests their finances are okay as those places must be naturally scarce and desirable, hence expensive. Sounds about “shrewd cunning” ?

        • unsafeideas says:

          How does all that relates to what we have been commenting about? This thread was not about sexually frustrated people looking for hot partners. Neither it was about class.

        • Tibor says:

          Looking like a dumb macho might get you a certain kind of women, but if you in fact are smart and look for more than just good looks (I am not saying you don’t want that, but you want other things too), this kind of behaviour is going to hurt you rather than help you.

          But like I pointed out above – extreme masculinity and extreme feminity is probably mostly just stupidity. But you can appear both very masculine and sophisticated – think Sean Connery. I doubt anyone would say that Connery is a brute and at the same time, nobody would suggests that he is not masculine. Masculine can mean a brute, masculine can also mean a “stand-up guy”. You don’t find many brutes among smart men, so you find fewer among smart men who fit into this particular stereotype of masculinity…so you (not you you, general you) conclude that they are less masculine, while their different behaviour may simply stem from their higher intelligence itself.

    • Cord Shirt says:

      Studies of high-IQ children going back to the 1920s have usually found them to be more androgynous than unselected children. Often these studies either don’t follow them into adulthood or don’t check the adults for androgyny/femininity/masculinity…so dunno about adults.

      Effeminate men used to be one of feminism’s constituencies. Guess that’s changed with feminism’s popularization/oversimplification. :sigh:

  8. Daniel Speyer says:

    I note that in the LW data, there is no relationship between masculinity and femininity as measured by BEM. I would have expected a negative correlation.

    • Scott Alexander says:

      The real studies find a correlation, but it’s surprisingly small. I don’t know if that reflects “reality” or Bem’s test being bad.

      • Deiseach says:

        Having had a quick gander at an online Bem test, I’d say it badly needs updating. Originating from the early 70s, the way it classifies is as much based on stereotypes as anything else – anyone else willing to bet they assign “assertive” as masculine, when we’ve had about three decades of assertiveness training since then and it’s a trait encouraged in women (assertiveness is not bossiness or nagging, gentlemen) especially for success in work?

        Also, what does it mean to score yourself as “masculine” or “feminine”? If Jane plays a lot of sports, is she being more “masculine” than the average woman?

        Anyway, I got 35/100 Masculine points, 36.842/100 Feminine points, and 56.667/100 Neutral/Androgynous points.

        Oh, dear: now I am having a gender identity crisis! (not) 😉

      • TMK says:

        Why would it that reflect on either? Why should the scales of Bem test correlate with each other?

  9. Possibly some added fun: The Armored Rose (a book about women and SCA fighting) has it that there are skeletal differences between men and women, though with a 10% overlap. The only one I remember is that women’s line of knuckles typically slants down towards the outside of their wrists, while men’s run parallel to their wrists. This is easier to see if you make a fist. (It’s important for SCA fighters because it affects sword strokes.) I have no idea whether knuckle angle correlates with digit ratio.

    Another gender difference mentioned in the book is that women’s adrenaline revs up more slowly than men’s, so training methods which are dependent on raising adrenaline fast don’t work well for women.

    • Deiseach says:

      *makes a fist*

      *line of knuckles is at a slant*

      Hey, that’s true!

    • Ydirbut says:

      *Makes a fist*

      *Line of knuckles depends on how far I pull in my pinkie*

      Hey, that’s meaningless!

    • The slant actually does shorten the ring finger, 4D or am I misunderstanding something? For what it worths, I am a pretty slanty man, apparently, and is a good thing because in boxing I am supposed to hit with the first two knuckles, not doing so results in the infamous boxers fracture, and the slant is really handy at keeping the last two knuckles out of the way.

      In fact, if it is true that the man jaw evolved as a way of protection from fist punches, if prehistoric men really “boxed” a lot, the knuckle slant looks like something mighty useful for brawly men to avoid that boxers fracture. You really really don’t want your last two knuckles meet someone’s jaw or skull.

      But prehistoric men probably didn’t box actually, there are whole cultures, like Middle East, where fighting with slaps is far more common than with wrists. I am not sure how is that even supposed to work but the French based a whole sport on slapfighting (and kicking each other with shoes on, ouch) called savate so there is perhaps something to that. BTW its history is that the government tried to curb brawls by making a law that hitting someone with a fist counts as an armed attack. Naturally the Marseille sailors were having none of it and proceed promptly to slap and kick each other. I find it highly amusing.

      • Eric Raymond says that untrained men use overhand blows at each other’s shoulders, which makes sense in Konrad Lorenz sort of way– dominance fights are supposed to *not* cause injury.

        Now that I think about it, untrained isn’t the same thing as innate and not influenced by culture.

  10. Martin Lindfors says:

    You can’t pick and choose and look for correlations within the data. This is one of the most important root causes for the problems with reproducibility within social sciences. You should either

    1. Formulate the hypothesis
    2. Collect data
    3. Test the data for whether or not it supports the hypothesis,


    1. Collect some data
    2. Search for interesting hypotheses
    3. Formulate a clear hypothesis
    4. Collect new data independently of the old
    5. Test the *new* data for whether or not it supports the hypothesis

    • nico says:

      Yes, it seems like Scott’s on step b2.

    • Scott Alexander says:

      My main prediction going into this was that we would replicate the digit-ratio/feminism correlation found previously. Technically I did replicate it in the whole sample, but not in the female subsample.

      • Deiseach says:

        That’s based on 70s test results, though, yes? We’ve gone through several iterations of feminism since then. I think those original results are most useful as historical data but I wouldn’t use them for modern-day predictions.

        • Tibor says:

          Is it? I am no expert on feminism, but it seems to me that 70s feminism was more or less women saying “it is ok for women to have a career too” and modern feminists today saying, if not directly, “it is wrong for women to stick to the traditional female role”. Then not differentiating between those two is like equating goes-to-church-on-christmas “Christians” with people who give money to televangelists.

    • Jaskologist says:

      Doesn’t that run against a lot of what Google and modern machine learning do? They like to take a whole lot of data, and then find every correlation they can. It can be surprisingly effective.

      • suntzuanime says:

        Yeah the real answer is not “you can’t do this” but “you need a more rigorous standard than p<0.05 if you're going to do this". the good old interocular trauma test is pretty robust.

  11. Neanderthal From Mordor says:

    For correlation purposes you should have asked about penis length in men, and lesbianism in women.

    • William O. B'Livion says:

      How do you reliably get a penis length measurement?

      Slapping it on a photocopier is going to cause some issues.

      • Siah Sargus says:

        You measure on the top of the shaft, from the pubic bone to the tip, usually with a tape measure. The precision of this method is about the same as other body measurements, like waist circumference for instance. Of course, we really should just use super-detailed six-step process, as outlined by a prestigious dick-measuring study, for optimal results.

      • Urstoff says:

        The photocopier method hasn’t worked for me yet, but I’ll keep trying.

    • Scott Alexander says:

      I *have* data on lesbianism in women, but there weren’t enough lesbians to make a decent sample size.

      • Deiseach says:

        That might make another good survey question, for the person in the open comment thread asking for suggestions.

        “Could you – purely as a thought experiment or fantasy, without this drawing any conclusions or suggesting anything about your sexuality, or suggesting you want to or ever would do so – imagine yourself having sex with someone of your own gender?”

        I have half a notion women would be more likely to say “yes” or at least not “of course not, the very notion turns my stomach and makes me want to scrub my brain with bleach” than men, and I have half a notion this is tied up with very complicated ideas of sex and gender roles and who does what to whom (i.e. the Classical Roman division between the penetrator and the penetrated, not by gender of same) and that it needn’t have anything to do with homophobia (that is, I’m quite sure that men who don’t get triggered by a disgust reaction, don’t think gay sex is filthy and are all for equal rights for LGBT persons still couldn’t even imagine themselves being sexual with another man).

        • TrivialGravitas says:

          I don’t see how you can get that from the Roman idea. Roman influenced disgust would only provoke disgust at catching, not pitching for male/male relationships.

          Testable though! Also ask men how they feel about getting pegged by a woman. I would predict a substantially larger number of men are disgusted at being with a man regardless of the act than getting pegged.

          • Deiseach says:

            Roman influenced disgust would only provoke disgust at catching, not pitching for male/male relationships.

            Well, that’s part of it: if the immediate reaction is “Yeah, but who’s on top? I’m okay with that but I’m not bottoming for anyone”, then that’s part of the complications I mean.

            I don’t necessarily think women would go immediately to “Am I going to be on top or on bottom?” as the determinant there, while I think a lot of men would, because of what TheDividualist says – the notion of dominance and “strongest dog fucks”.

        • Note: the Classical Roman division is extremely widespread in lower class cultures, even leading to things like Russian prison logic: “punish the filthy gay by raping him”, at some level I find CR division “natural”, at least, “natural” in the sense that if you care about dominance status it really follows from it. And it actually IS the most widespread reason for homophobia, at least sure it is the reason of mine and most of my friends. Religion is just an excuse to save face amongst the upper classes, for those who even care.

          And it is not complicated, it is the most basic simple uncomplicated dominance based “the stronger dog fscks” thing, if you find it complicated then at some level you really don’t understand how dominance status works… You understand this concept “culture of honor? It is the same thing…

          Culture of honor = dominance = CR separation = masculine gender role in its most simple basic uncomplicated form.

          I think maybe it can look complicated because it looks extremely unethical and you are kind of desperately trying to see it from a more ethical angle and that is complicated. But historically speaking being good at dominating others is a huge part of ethics, consider how “virtue” comes from “virtus” and that from “vir” which means man but that comes from “vis” meaning strength, see “viribus unitis”. So in the classical sense, “morally virtuous” really does mean, at least partially, at least some undertone level “stronger dog”.

          Not trying to depress you – I really like you – and I am afraid this is another round when I inadvertedly do that. Sorry. I am just geographing the City of Man…

      • 27chaos says:

        Why are you acting so frequentist? 🙁

  12. Douglas Knight says:

    This isn’t the first time the Bem sex role inventory has been mentioned on this blog, but you really should say that Sandra and Daryl are different people, indeed, married.

  13. For reasons which open a can of worms, but should be obvious to many of your readers, I think SSC readers include a disproportionate number of men who react badly to feminism but do not meet other masculinity criteria – and in particular, react badly to feminism BECAUSE they do not meet other masculinity criteria and see feminism as having made and broken promises to men like themselves. It’s unsurprising to me that the study results would reflect such a population.

    • Scott Alexander says:

      The effect persisted on the LW subsample taken alone.

      • I don’t know enough about LW’s readership to judge how similar its population may be to SSC’s, but even “persons who read Web logs” is already a population quite narrowly selected from the world in general.

        • Viliam says:

          I could imagine a relationship between “is a man who reads internet a lot” and “is a man disappointed by feminism”.

          At least in my experience, in real life you usually meet people saying “feminism is about equality”, while on internet you often meet people saying “this guy used a wrong pronoun once, hey everyone, let’s get him fired!”

          Also, outside of internet, it is difficult to find any mention about men’s rights. So the men who are pro-equality but anti-#killallmen don’t even realize there is an alternative.

    • Why don’t I see any gamergate activism here or on LW? As that is a 100% perfect sample of that…

      • My guess for LW is there’s a consensus that gender divisiveness simultaneously grabs attention while being too much work for too little gain. It’s like having a basilisk without the coolness factor.

  14. chaosmage says:

    Hadn’t you explicitly told us about your expectation that digit ratio might predict attitudes towards feminism? Maybe a few of us were influenced by that and, already having measured our digit ratio, rated our support for feminism in the desired direction. That’d explain why this particular correlation could turn up despite the other data pointing at digit ratio not actually being very correlated with anything.

  15. Harald K says:

    I’m guessing that the rise of men’s rights ideas may be an explanation for what surprises you about gender roles conformance vs. attitude to feminism. Men’s rights activists tend to be both strongly against traditional gender roles, and yet negative to feminism.

    • Orphan Wilde says:

      My experience with MRAs is that most of them tried feminism first, and found it unwelcoming at best, and most often outright hostile.

  16. James says:

    I didn’t measure my digit ratio at the time of the survey, but did a couple of weeks ago, and was surprised by how femme it was—somewhere over 0.98, putting me above the female mean. (I’m a man.) I was surprised for the same reason Scott is surprised at his data, because I have a highly ‘masculine-style’ intelligence (mathematical, work with computers, etc). Basically, I went through a miniature N=1 version of the perplexity of this post.

  17. I have low-T all through my childhood, I have every sign of low prenatal T, and from my teenagerhood I made a conscious effort to man up with workout and other means. This apparently influenced my view about politics and suchlike as well far more than the initial low T. I thought this must be a rare, weird and bad combination. In fact, there were cases when I felt a bit fake, because you should somehow accept all the deficiencies you were born with and fixing them is somehow wrong and similar irrational feelings. Then I learned about Ted Roosevelt. Ooops. This is apparently a thing, and a succesful thing, at least in some cases. Then I researched a bit more and now it is almost funny. Apparently interviews with just about every body building champion begin with “well I was a weak boy and…”

  18. Ilya Shpitser says:

    Perhaps nerdy men tend to be low T?

    • Zykrom says:

      Theoretically nerds are supposed to have low T but high natal T, which is what determines the digit ratio.

      • Mathy nerds might need high prenatal T but anime fans, D&D players etc. not. If someone is a nerd who is into D&D but never liked math and studied history instead, I would bet low T on both factors.

    • Perhaps the sky is blue? Intellectual interests typically mean a withdrawal from the physical-competitive activities that increase T, and the other way around, having a lower T and disinterest in those activities is perhaps required in order to be able to invest time into learning.

      I mean, “normal” boys hate sitting in a class studying, they want to go out and climb trees.

      E.g. one guy told me “I used to be a huge neckbeard and when I joined the wrestling team in college that slowly fixed it”.

      Also, the nerd who isn’t smart is simply invisible – yes, there are such people as lower-IQ D&D fans, anime fans or World of Warcraft addicts and they are simply considered losers.

      The ambiguous meaning of “nerd” or “geek” largely comes from a surprise factor: “hey, I didn’t think this social loser here is so smart!”

      I was never able to fully understand myself but I think it was about some kind of neurological near-disability made me have really low dexterity, I still cannot juggle oranges, I just don’t understand how is that even possible. So I stayed out of the physical games and read books instead which caused low T. And I had a classmate who was exactly the same except he wasn’t even smart and didn’t read and he was not seen as a geek but as a non-person basically…

      • science says:

        Back to step 0 I see.

        • Sorry doc, will try to improve!

          Joke aside, there is a certain problem with survey methods and the LW love of statistics. It is gathering data from a lot of people, but shallow data, like “how much you like cake, 1 to 5” Sometimes you need to do the opposite, dig deep in a few selected spots like hours long deep interviews with selected subjects asking “but really, really, just what do you find in cake that you like? describe the liking feeling in detail. when did it start?” This is something easier done on yourself and your friends.

  19. Muga Sofer says:

    >digit ratio did not correlate with Bem sex roles at all, but did correlate with positive opinion of feminism (p = 0.001). Bem sex role did corelate with feminism;

    First, misspelled the second “correlate”.

    Second… how is this possible? Are the correlations involved just that weak?

  20. Michael Watts says:

    we couldn’t even get the basic finding of men being more male-typical than women

    There are hints of this in some other comments, but I’ll say it plainly: the “basic finding” you mention should be hard to see, because you’ve preimposed a heavy masculinity filter on the entire surveyed population. See: 529 people responding, of which 75 female. Include some feminine women, and the “basic finding” should show much more clearly.

    This comment does not express an opinion on whether men actually are more male-typical, as defined in the post, than women; it is only meant to point out that under the assumption that this is the case, a LW survey can’t be expected to show it very well.

    • Zykrom says:

      How does this explain “In fact, our entire sample was heavily feminized compared to those of every other study; our men had more feminine digit ratios than other studies’ women?”

      • Michael Watts says:

        To reconcile those, I believe you’d need a masculinity-femininity space of more than one dimension.

        Consider something Steve Sailer wrote, which I will here paraphrase from memory: “being a nerd is one way of expressing masculinity”. If nerds are all[1] men, but the prototypical manly man is not a nerd, then you can’t fit everyone onto a line going from masculine to feminine no matter how you try.

        Consider also the idea of a “bad girl”. This bad girl’s behavior is unladylike — she sits with her feet up on the desk and is physically familiar with several different boys. The terminology here, “bad girl” and “unladylike”, suggests that there are problems with her femininity. However, I claim that while sitting in unorthodox positions like feet-on-desk can be reasonably thought of as a masculine trait, slutty or suggestive-of-sluttiness behavior is “bad” in a particularly female way, and doesn’t mean that the girl’s feminine nature is inhibited. Rather, it is an expression of femininity along a different dimension than the “correct”, decorous one.

        [1] almost all

    • Scott Alexander says:

      Yes, but I wasn’t picky – if it had at least trended toward significance, I would have been happy like this. But 75 women and no trend.

      • Peter says:

        What’s the confidence interval? You should be able to get values of Cohen’s d from other surveys, and see if your confidence interval on Cohen’s d includes that value.

        • Izaak Weiss says:

          This seems wrong? The p = .05 boundary is customary, sure, and I don’t mind adopting it, but If we’re analyzing data, trying to find trends that we would like to try and replicate with larger sample sizes, more rigorous measurement, or whatnot, talking about things that aren’t significant in our sample, but that might be false “negatives”, is worthwhile.

        • 27chaos says:

          In this context, Scott’s claim makes sense. The commenter claimed that Scott is making a false negative claim, Scott’s saying he doubts it because the way he assessed p values erred more towards accepting false positives.

  21. stargirl says:

    Can you please not use the term “biologically male.” It is definitely going to upset a decent number of readers. “Male assigned at birth” is a much better way of getting across what you want to.

    • Anonymous says:

      Why exactly would they* object to that? Which part of that isn’t true?

      * For that matter, who are the “decent number of readers”? You imply that you yourself are not offended, but appear to also claim to be some sort of spokesperson for this group. With all due respect, you seem to be a concern troll.

      • stargirl says:

        I found the term pretty off-putting personally. In addition a reasonable number of this blog’s readers are trans. A given trans person is very likely to find “biologically X” offensive.

        I think you are pattern matching me to someone I am definitely not. Anyone on this blog who remembers my comments can confirm I am not a SJ type. In fact I strongly disagree with alot of SJ arguments. And have been doxxed by SJ people for disagreeing with them. There is however no rule of the universe that says that the out-group is wrong about everything. I happen to think the SJW are right about the use of terms like “Biologically X” and people should not use such terms.

        Your comment points out why I do not like the term “concern troll.” As far as I can tell the term mostly functions to make criticism difficult without getting painted as a hated outgroup member. There is no reasonable definition of “Concern troll” under which I fit. I have posting comments on this site for well over a year. The vast majority of my comments are not criticizing Scott for being insufficiently SJ-approved.

        In addition Scott has repeatedly posted worries that the blog is moving too far right. Or at least the readership is. Given Scott’s concerns it is a very bad idea to use “biologically X.” Using that term tends to signal views on gender that many on the left (Correctly imo) see as harmful to trans people. It is not a large cost for Scott to change his usage here. And I think changing his usage would better reflect Scott’s actual views.

        I will note you did not actually disagree with me that the term will offend a decent number of readers.

        • Berna says:

          Aha, I get it. Until now, I’d always thought that ‘[sex] assigned at birth’ meant the baby was intersex, and that the doctor had to make an explicit decision about whether to call it a boy or a girl.

          • Douglas Knight says:

            That is exactly what it means. The whole point of using it generally is to assert that everyone is intersex.

          • Peter says:

            @Douglas Knight

            Nope. Your comment is the first place I’ve heard that suggestion, and I’ve heard quite a few things in my time from all sorts of different people. Intersex conditions are a grab-bag of semi-related things, but most people don’t have them.

            What it means, is that with some intersex conditions it’s obvious that an assignment is being made, with others[1], and with non-intersex people, it’s less obvious; open to debate, even.

            [1] Quite a few conditions aren’t obvious at birth, but come to people’s attention later on.

          • Anonymous says:

            Ah, so it is about the assignation at birth socially, and not the karyotype decision at the moment of conception. But surely then “biologically male” and “male assigned at birth” are *not* synonymous, so can’t be substituted in good faith (whence the hostility toward stargirl), as it is possible for instance via complete androgen insensitivity, or some kind of pseudohermaphroditism (ie not genetic mosaicism), to say be biologically male and yet be assumed female at birth. Or am I in fact wrong in assuming “biologically male” really does refer to 46,XY humans, as distinct from 46,XX humans?

        • Anonymous says:

          I will note you did not actually disagree with me that the term will offend a decent number of readers.

          Nolo contendre. It might offend them, just like it offended you, but it’s irrelevant – taking offense where no offense was clearly intended is a problem with the parties being offended, not the supposed offender.

          • wysinwyg says:

            That’s a value judgment, not a fact about the state of the universe.

            As Stargirl explains at length, using different verbiage could very well help Scott obtain results that he desires. That seems relevant to me.

          • Peter says:

            I don’t think this is the whole of it. One can take offence at disrespect, and one way to be disrespectful is to fail to pay proper attention to something you should be paying attention to. In a sense, no offence was intended; the very lack of intentions of the disrespecter towards the disrespected is precisely what the problem is.

            I mean, it’s a well established principle in law that mens rea can cover negligence (or in plain English, carelessness) and recklessness as well as things done knowingly or deliberately. In terms of everyday politeness, there seems to be a think where ordinary slip-ups – at the level where only extraordinary carefulness could have prevented the slip-up (Adam Smith has an interesting paragraph on this, if you can get past the example being about accidentally killing someone’s slave) – prompt a small apology (maybe just an “Oops!” or “Sorry!”). There’s also the related “well, you weren’t to know, but now you do know, don’t do it again” effect.

            A key point where I differ from the SJ movement is that I think that in the fog of culture war, it can be hard to tell reasonable from unreasonable requests – both for the requester and the requestee. So I’m prepared to cut people slack, both because it may take them time to sort out what’s what, and also to allow for the possibility that I’m the one in the wrong; I’m not trying to claim I’m in a position to “educate” him or anything like that. Being aggressive about this would also be a massive breach of the Golden Rule, and, y’know, a large part of the reason for coming and staying here is because I have problems with the SJ movement and it’s remarkably therapeutic to hear from someone who feels the way I do about various issues. SSC is a hell of a lot safer for me than some nominally “safe” spaces, and I like things to stay that way. Of course, this doesn’t mean I can’t try to influence people.

        • I remember your comments being generally very reasonable and have a lot of sympathy for your being targeted in such a way. I hope you don’t let them stop you doing what you do!!

          I would say though I think if something is factual and doesn’t significantly malign people it should probably off-limits for being supressed for potential offensiveness. Do you feel “biological male” does not meet this criteria? I don’t mean to tread on anyone’s toes, but I’d like to be able to discuss XX and XY and if there is or isn’t relations to specific stereotypical gender traits, for example. This seems to prevent that discussion, Or perhaps I am confused about your meaning?

          • Peter says:

            If actually you want to talk about someone being XY, talk about them being XY. If you want to use some proxy to guess whether someone is XY, talk about that proxy, and say what you think it’s a proxy for, and say how the proxy could fail. In particular, your proxy should be something that is actually observed. Be prepared to discuss relevant edge cases; for example, if you think XY specifically is important, be prepared to argue why XY is important even in individuals with no function SRY gene on their Y chromosome, or XY with no functioning androgen receptors, or whatever.

            If you just mean “the biological correlates of male assignment at birth, whatever they may be” then say that. But be prepared for various correlates not to correlate perfectly with each other, and for the actual story of the correlates to be more complicated, messy and harder to know than you had imagined. Note that there are various studies of those biological correlates that find that some of them are better correlated with gender identity than birth assignment; to pick a relevant example, there was a study that found that the 2D:4D ratio in trans women was similarly distributed to that of cis women, and not cis men. The information comes in in dribs and drabs, and if you like you can try disputing some of the statistics, but you can’t just brush it away and assume the opposite of what it says.

            One point – there’s an interesting nature/nurture debate to be had to do with stereotypical traits in the cis population. It’s tempting to use the trans population to settle this, by assuming that they have the same “nature” as cis people of the same birth assignment. Don’t do this – you can’t just assume this works, there are studies that contradict the assumption, and it pisses people off.

          • Anonymous says:

            Shorter: Gender and sex are murky enough, particularly for a survey aimed at its borders and edge cases, that the term “biological male” may be question-begging, and should be avoided on that basis.

          • Tracy W says:

            Peter: you are asking for Scott, and anyone else writing about sex and gender, to add an awful lot of words, none of which are fundamentally going to change his points. That’s something that seems minor but in fact is an awfully a big ask. Good writing always assumes a large amount of background knowledge, and focuses on the new information the speaker wants to convey.

          • Peter says:

            I think an issue here is what you’re trying to do.

            For correlational studies on the general population, that aren’t concerned with cis/trans status – correlations with gender identity are likely to be very similar to correlations with birth assignment, current possession of sperm-producing testes, XY karyotype, possession of a working SRY gene, etc. Basically, whatever you’re doing, you’re going to be using some proxy for the thing that’s the actually interesting thing. So, for correlational work, why not go off gender identity? Or split it into a four (or more)-way comparison? I mean, for 2D:4D ratio, assuming that the relevant biological factors correlate better with those that cause a penis to be there at birth than with gender identity is a) unwarranted and b) contradicted by recent studies, so probably wrong.

            For general discussion of nature/nurture debate on sex and gender – if you want to go beyond correlation into causation – you can talk about how biological factors may or may not influence various traits without having to commit yourself to any particular people being male or female. Are men more interested in sex than women, and if so, is it biological? My answer is, “quite possibly, it seems to have something to do with testosterone”. Asserting that trans men were biological women here would not only be unnecessary, but counterproductive; lots of trans guys have male-typical testosterone levels.

            I make the correlation-causation distinction – being assigned male at birth doesn’t cause hairs to grow on your chin, but it’s correlated. And mentioning something that so obviously doesn’t cause the things of interest is a good way to remind people of how correlation works. If they’re shocked or startled by it, then evidently they need to improve their ideas of correlation to understand what’s being discussed properly.

          • houseboatonstyx says:

            @ Tracy W
            “Peter: you are asking for Scott, and anyone else writing about sex and gender, to add an awful lot of words, none of which are fundamentally going to change his points.”

            Adding words for a sort of PC disclaimer padding, is bad, I agree. But what about using shorter, or at least more exact, wording in the first place? Such as “XY-SRY-2d:4d-C3PIO” or whatever.

            “Good writing always assumes a large amount of background knowledge, and focuses on the new information the speaker wants to convey.”

            Good writing for what audience? An audience with the background knowledge to know that ‘biologically male’ means to convey [technical term] will understand a more technical term. Where there is a more general audience, a term like ‘biologically male’ confuses some readers, who may then derail the conversation.

          • Tracy W says:

            Firstly, knowing your audience is a part of the art of good writing, and a good piece of writing for one audience may be totally inappropriate for another audience, because of differences in background knowledge. If writing covers only things that the audience already knows, then the audience will get bored (unless the author writes more poetically and gives the reader the mental exercise of working out the allusions), if writing contains too little stuff that the audience already knows the audience will get lost. Parents are often bored to tears by books their pre-school children demand avidly (some writers have the skill of pulling off books/movies/etc that appeal across the ages by writing on multiple levels, but that’s genuis-level work). A popular (as measured by sales) science book aimed at educated adult readers will almost always be written assuming far more background knowledge than a popular science book aimed at 5-7 year olds.

            To support this claim, if you look at great pieces of writing aimed at adults but discussing every day topics, they are generally not straightforward. Eg: Jane Austen wrote “It is a truth universally acknowledged that a single man in possession of a good fortune must be in want of a wife”, rather than “Women often want rich husbands”, “Three score and seven years ago our fathers …” (Abraham Lincoln) rather than “In 1776…” . “For sale: baby shoes, never used.” (Ernest Hemingway), merely implies the tragedy, and draws on background knowledge that “never used” is wording for second-hand sales, not department-store ads. (Writing aimed at introducing new information for adults is much more straightforward, even when it starts with a poetic idea like “We are made of star-stuff”, explanation quickly follows.)

            Secondly, readers have limited working memory – the common rule is that people can keep in mind 7 +/- 2, based on the average number of digits that people can keep in mind during psychological testing. (And digits are relatively simple, and highly practised given their relationship to things like money, for more complex ideas this is going to be smaller). A good writer aims to use that working memory space most efficiently to present whatever idea they want to convey. So, yes, Scott could introduce a more technical term to his audience and they probably could follow that in abstract, but until that technical term is practiced, in a variety of different circumstances, it’s going to be taking up working memory space and thus reducing what other information Scott can present.

            Yes, a term like “biologically male” will confuse some readers. The question is how many readers will it confuse compared to any alternative form? And what is the trade-off to Scott’s writing goals? If Scott wrote exclusively about, say, the details of sex and gender and posession of a working SRY gene, and etc, he might avoid confusing any readers on that point, but also might wind up driving far more away via boredom: if I wanted to read technical discussions of working SRY genes I’d be reading science journals in biology. You are asking Scott to change his writing focus drastically and thus probably his audience. As I said, this is a big ask.

            But what about using shorter, or at least more exact, wording in the first place? Such as “XY-SRY-2d:4d-CSPIO”

            In this case there is a trade-off between shorter and more exact, and accessible. If Scott had said “men” and “women” it would have been shorter, but less exact. If he had used a longer, more exact term than “biologically male”, it would be adding length and thus detracting from the points Scott wants to make. If you make up a word like “frext”, and define it to refer to the same thing Scott means by the term “biologically male”, then that’s short, but it doesn’t tap into any readers’ existing knowledge of related terms and thus is less accessible, so it would need to be redefined when re-used several times.

            If Scott had used “XY-SRY-2d:4d-CSPIO”, well, there’s a reason that English outside of technical journals doesn’t use words like that: there’s an optimal level of redundancy in language. If I make a small typo or misspelling in English generally (not always!) tyhere’s enuf redundancy you can work out my meaning. If I use a word you’re not familar with, a fluent English speaker can often pick up the general gist of the meaning from surrounding words. A term like “XY-SRY-2d:4d-CSPIO” doesn’t offer that redundancy.

          • @Peter & others

            Your discussion about what objects are relevant and how they correlate is fine, and I may or may not agree. But what was raised was that a term was *offensive*. That’s quite different, and I think as rationalists we should be very limited in how they label things in that way.

          • Peter says:

            @Tracy W

            The issue isn’t just background knowledge: it’s background ignorance (and meta-ignorance), background assumptions, background misconceptions.

            Citizensearth said they’d like to be able to discuss XX and XY; a lot of my initial reply to them was taking that suggestion literally, that they actually wanted to talk about XX and XY. As far as I can tell, other people have taken “XX” and “XY” to mean “some well-defined biological thing synonymous with ‘biologically male'”. I’ve seen the made-up term “XY-SRY-2d:4d-C3PIO” come up. I do not know exactly what that term is meant to mean, but I suspect it may be an attempt at “a well-defined biological thing synonymous with ‘biologically male”. I gave some examples of why “XX” and “XY” might not be the thing they were looking for – those examples were in no way meant to be comprehensive. Given our current state of knowledge, I don’t think it is possible to be comprehensive.

            I offered a few things which seemed to me like neat dodges. One was for correlational work – to point towards a correlate of all that stuff which we’re on pretty firm ground with – birth assignment. But oh no, we can’t use that because we have to pander to the background ignorance of some of our readers. Another, not for correlational work, was to talk about “biological stuff that causes maleness” which people don’t seem to have picked up on – given that this post is about correlational work, I can see why, but given that when you talk about correlational work people say they want to go beyond that into causation, I thought I should provide a way to do that too.

          • Tracy W says:

            Peter: All of what you say may be true. But, if Scott was to write about all that, he would be detracting from the main message of his writing, the topic he actually wants to write about – the outcome of his digit testing. The price of including longer language about biological sex means having to simplify other bits of the writing, or having readers not following the main message because they’ve been lost in a bunch of unfamilar words about XY karyotypes, etc.

            Your framing of “pandering to the background ignorance of some of our readers” is an odd way of wording “know your audience”. I don’t know about you, but I actually prefer to read stuff that (a) presents me with things I didn’t know before and (b) does so in a way I can follow, which means tying back to my existing knowledge base. If that’s pandering to my background ignorance, then please, Scott, more pandering. If I only wanted to read things that are comfortably within my existing knowledge base, my local library has a very large kids’ section.

        • Whatever Happened To Anonymous says:

          I, too, (for whatever it’s worth) will vouch for stargirl not being a SJ type person, or at the very least being an uncharacteristically good sport about it.

          On the other hand, I don’t think it should matter.

        • Such a long post and you don’t even explain why it is offensive. Biological is supposed to mean “the genitals”. Of course the brain is biological as well but ideas, culture, behavior and identity as are customarily put in a different category, software not hardware. So biological male just means male hardware, penis, while the brain can be female, even biologically, which is weird. But instead we interpret it as a person on a software level choosing a female gendered identity and behavior. Perhaps the issue is that it suggests it is software and not brain hardware?

          • Peter says:

            One of the issues is that the software/hardware distinction is better for computers than the brain. Yes, every difference in software implies a difference in hardware. Neuroanatomy, though, differences in size big enough to see with MRI, or in slices from a dissected brain, that awkwardly straddles the software/hardware divide.

            There are some intriguing results showing anatomy congruent with gender identity and discongruent with birth genital configuration. There are also suggestions that differences in prenatal or early-postnatal hormone exposure may be related to the condition. As I’ve said elsewhere, wikipedia cites a 2D:4D ratio study that found finger length consistent with gender identity rather than birth genital configuration, so it looks like there’s some interesting biology going on.

            To be perfectly honest, some of the “why do you find it so offensive” answers may well induce eye-rolling, and as mentioned elsewhere there are a few things I’ve rolled my eyes at in my time. It’s surprising how wound up I’ve let myself become about something that seems like such a silly little thing, and to a certain extent I’m defending other’s right to be offended when perhaps it would be best to let it go. Anyway:

            Part of it is the “not being taken seriously” thing – ‘biologically’ having overtones of ‘really’.
            Part of it is getting annoyed at an odd fixation with genitals, which shows up elsewhere too. I mean, if you’re shagging someone then genitals may be a big deal, but generally hormones and social transition are a bigger deal – and generally you can’t get genital surgery until you can show you’re taking well to hormones and social transition. Some people find questions about what’s down your pants, or what used to be down your pants, a bit personal. Cis people can have gonads or genitalia removed for various reasons, for example (e.g. accidents, cancer surgery, etc.), and no-one asks about that.
            Part of it is trans women not wanting to be lumped in with cis men, and trans men not wanting to be lumped in with cis women.
            Part of it is what I like to call the “psychological irreducibility” of the thing. Lots of people rack their brains for “why” and fail to come up with anything – it’s a common theme that comes up. It sort-of feels like if it was a “software” thing then it would be something you can be talked out of. I think quite a few people spent quite a while being talked out of something, and feeling miserable because of it, and when they allowed themselves to stop being talked out of it then they got happier (but often more frustrated with the world around them).
            There might be something to do with the whole vexed “what you were pre-transition” thing (which is a whole story in itself and more complicated than some loud activists like to make out), but the links seem more tenuous than the stuff I’ve mentioned above.

            I may have missed something, and other people would likely be more forthright (to say the least) than me.

            To me, none of this feels strong enough to demand anything, although other people’s mileages may vary; it feels like enough to request things – or to request giving the matter reflection and consideration. Of course, some people will call me all sorts of nasty names for saying this but there you go.

    • thirqual says:

      Assumption: Scott meant “bone development under a androgen-dominated system”. That makes MAAB not quite right either (intersex people, arguably depending on how they are counted, are as numerous as trans people in the general population).

      … and then a quick search showed me how “biologically X” is used typically and I get your point.

    • Diadem says:

      I have to admit I never understood this objection, or the new term.

      To me “Male assigned at birth” sounds like the doctor flipped a coin to decide whether to call the newborn a boy or a girl. That’s not what happens. The doctor uses a criterion or set of criteria to determine whether to call the baby a boy or a girl. Presumably they’ll generally just look at what’s between the baby’s legs. The very existence of trans people proves that this is not 100% reliable. But on the other hand it’s clearly much better than chance.

      Doctors don’t randomly assign gender. They assign it based on some biological criteria. What’s wrong with call those criteria ‘biological sex’? Seems like a logical term.

      • Deiseach says:

        What’s wrong with call those criteria ‘biological sex’?


        You have no idea what you’re asking. No idea.

        I am not going to venture into this morass because I’m already fighting enough on Tumblr, I don’t need to start a fresh set of rows on here 🙂

      • Peter says:

        The phrase doesn’t say anything about randomness or capriciousness or whatever; it says it was assigned, you say it was assigned, what’s the problem?

        The thing about checking for the presence of a penis; in many ways it’s an odd thing to pick out as the determinant of biological sex. There are lots of other things that are clearly biological which correlate strongly but not absolutely with the presence of a penis, some of those may or may not be more important. A common theory, with some empirical support, is that some of these things are neurological, and push people in the direction of male or female or other identities. After all, in the rest of biology, there are lots of species that never have penises at all, yet we’re happy to assign “male” or “female” to them, or parts of their anatomy, or stages in lifecycles, or whatever. The thing that zoologists and botanists like to do is go by gamete size; big gametes female, small gametes male. Except for some reason we don’t usually try to harvest gamete samples from newborns and examine them under a microscope.

        The reason I’d like to specifically recommend the phrase to Scott is that I think it might be the rationalist way to go. “Rationalist taboo” is a useful tool for clearing up all sorts of conceptual confusions, and what I’m proposing is a variant on this; tabooing the use but not mention of “male” and “female”. The map is not the territory, but various maps may be parts of the territory, “male” and “female” (note the quote marks) are properties of maps. See the old LW blegg/rube post.

        One thing I will say is that mileages seriously vary. One surreal experience I’ve had the privilege of having is of hearing people on the internet complain bitterly about some term to do with trans* people, and then going along to my trans* support/social group and hearing people, cis and trans, use unironically the very same terms I was hearing being decried on the internet like there was no problem at all. Very odd.

        Disclaimer: I’m biased. What I’m calling my variant of “rationalist taboo” is basically how I make sense of myself; “by convention male, by convention female, in reality atoms and void”. Part of self-discovery was that finding out how I wanted to live my life, how I wished to present myself and be addressed etc. was relatively[2] easy compared to the nightmarishly hard question of what I ‘am’, for me it’s easy to doubt that this is even meaningful. But I worry that I’m treading on other people’s toes if I try to universalise this.

        [2] As in, “actually quite hard and still a work in progress, but nevertheless something I could get somewhere on, unlike other problems”.

        • What if the doctor thinks it is the balls that matter, not the penis, because they produce all that T that is going to determine the neurology and behavior and all that of the boy? Sounds correct to me?

          • Peter says:

            There’s an intersex condition called “micropenis”. Decades ago, doctors tended to think that the penis was the thing, that a micropenis wasn’t big enough to do the job, and reassigned those people. These days it seems that they don’t think that way. That said it’s one of those structures where the line between a micropenis and a macroclitoris is hard to draw and people tend to go by other anatomical structures in the region.

            To a certain extent, “what the doctor thinks” isn’t the thing, so much as “what the procedure is”.

            There are a few other conditions where there are testes, but they’re hidden. In 5-alpha reductase deficiency you have undescended testes and various other signs that cause people to be assigned as female, but when puberty hits the balls descend and all of the other signs of male puberty come with it. In CAIS, there are testes and indeed plenty of T, but it doesn’t do anything as the androgen receptors aren’t working.

      • Anon says:

        There is nothing inherently wrong with the phrase “biological sex”, however in the blogosphere there’s a pretty strong correlation between ‘uses the phrase “biological sex”‘ and ‘believes trans women are Really Men, and vice versa’, while “assigned male/female at birth” is used mostly by people who believe trans women are women, and vice versa.

        Like how “n*gger” seems like a perfectly fine derivative of the Latin word for black, yet people will assume certain things about your worldview if you go around using that word.

        • Tracy W says:

          Obvious solution: all the people who think trans woman are woman should start using “biological sex” at every opportunity, reclaim it and thus remove the stigma. It is evident that the last thing the English language needs is *more* ways to insult marginalized groups.

          (N*gger is a lost cause but just because the bigots won one battle is no reason for us to roll over on our backs and show our throats whenever they appear.)

        • I believe in a stronger version of it, namely that people are not free to choose what they are, they are whatever others see them, because it is a selfish thing to determine what you are for yourself, the properly social thing it define yourself as how you interact with others and be useful for them, so they get to define you.

    • “Male assigned at birth” strikes me as a less clear label. Suppose someone makes a mistake in the hospital and writes a boy baby down as a girl or vice versa. That isn’t what “biologically male” is intended to mean.

      In Rabbinic law there are two categories of ambiguous gender, hermaphrodite and tumtum. We aren’t certain what a tumtum was, but it was pretty clearly some sort of oddity that made it initially impossible to tell which gender the baby was, something concealing the genitals. A tumtum who got classified as male isn’t at all the same thing as a boy baby who might later decide to become trans.

      “Biologically male” isn’t perfectly unambiguous, since there are people whose genetics are neither XX nor XY and others who are genetically one gender but physically the other. But it doesn’t depend on how someone is assigned at birth.

    • Scott Alexander says:

      I think in this case I mean what I say. The point at issue is whether they have a body such that they would be exposed to male-typical androgen levels vs. female-typical androgen levels. I don’t care what gender their doctor assigns them socially, and anything “assigned at birth” doesn’t matter in a study where the relevant dimension is their sex in utero!

      • Peter says:

        I think there are a number of assumptions there that aren’t quite right.

        The whole reason we’re talking about 2D:4D is that variation in prenatal androgen (and estrogen) levels is more complicated than “male-typical levels” and “female-typical levels” – except that it’s probably more complicated than whatever it is that 2D:4D tracks too. I’ve seen prenatal hormone status implicated in all sorts of conditions, to do with gender identity, sexuality, autism spectrum status etc. and you get all sorts of weird and wonderful combinations of them.

        I doubt that any of us have hard data on our prenatal androgen levels. The one thing we do know is what the doctor’s decision was based on what they saw. A good thing to do is to report the observed relationship between observed things first, and then theorize about it later.

      • Chrysophylax says:

        Why not say “biologically male (by which I mean ‘exposed to male-typical levels of androgen in utero’)”? That reduces offensiveness and is more precise.

        The general objection to “biologically X” is that X is a cluster, not a dummy variable. Sex and gender have multiple continuous components, just like sexuality.

        It’s similar to defining species (see Wikipedia on the species problem), but even more difficult. With species, we look at things like DNA, physical appearance, frequency of mating and fertility of offspring. With sex, we need to look at DNA (which is a *lot* messier than XX/XY), levels of multiple hormones over time, genital structure, fertility/sterility, body shape and age as a minimum. (For particular applications, we can look at subsets – for example, there are cases where the relevant classes are “can ejaculate fertile sperm” and “all others”, and others where the relevant categories are “can succeed as a women’s underwear model” and “all others”.)

        • Tracy W says:

          The general objection to “biologically X” is that X is a cluster, not a dummy variable.

          If you look at English, this is incredibly common. “Blue” is a cluster, not a dummy variable. “Food” is a cluster, not a dummy variable (eg eggs can be nutritious to some and deadly poison to others.) “Water” is a cluster and not a dummy variable – the water out of my tap has some impurities in it such as flouride. And so forth.

          With sex, we need to look at {snip long list}

          Only if you’re a strict Aristotlian. The rest of humanity can live with a higher error tolerance for the efficiency gains.

    • While I agree that some people are offended by it, I think the effect of their taking offense is very bad. Much as if someone were to say, “Could you please not use the term ‘true’? I am offended by the fact that it implies that some of my beliefs might turn out to be false.”

    • keranih says:

      It is definitely going to upset a decent number of readers.

      While I would quibble with the assessment of “a decent number of readers”, I don’t think it’s generally useful to use terms that upset people.

      However, I share the concern of other commentators who find “male assigned at birth” to be very far from a “much better way of getting across what you want to.”

      In particular, revising definitions on the grounds of inaccurately describing a portion of the population which is of such a small size that it doesn’t even register on the plots does not strike me as particularly useful, and is a distraction from the already sizeable issues with this study.

    • Leonard says:

      Using “male assigned at birth” is also going to upset a lot of readers. “Biologically male”, or just “male”, is a much better way of getting across what Scott wants. Indeed I would bet this is so of the vast majority of readers.

      (Do you have any real trouble interpreting unmodified “male” as meaning “biologically male” or your “male assigned at birth”?)

    • Winter Shaker says:

      While I can happily get behind calling people by their preferred terms to avoid causing unnecessary offense, I’d still think that it’s useful to have a simple, relatively short term that groups [cis men + trans women], and a corresponding term for [cis women + trans men] for any occasion where people are discussing physiological, as opposed to psychological, characteristics that tend to vary with assigned sex at birth, as we are here. Is there any serious attempt to do this? I mean, something Latin derived, like maybe something like ‘virisign’ and ‘mulierisign’, would be reasonably user-friendly, yet just awkward enough to avoid jumping on something like a euphemism treadmill and becoming used to berate trans people?

      Or is there no way to avoid such a treadmill?

      [Edit – or ‘femisign’ – that would save another syllable, though at the risk of veering too close to ‘female’ and thus hitting that treadmill]

      • houseboatonstyx says:

        Somewhat out of context, I applaud the term ‘just awkward enough’. For a general audience, to convey a technical meaning it is advisable to use an obviously technical-sounding term. When ordinary people can reasonably conclude that you mean some ordinary thing, they will jump to different conclusions about which thing it is, and unhelpful disagreement will arise. If it’s a technical-sounding term, we can file it as ‘technical-unknown’ and refrain from jumping.

    • Izaak Weiss says:

      Interesting. I’m Male assigned at birth and not male, but this didn’t really bother me at all.

      What did bother me was the use of “transwoman” as a noun instead of an adjective noun combination, ie “trans woman”. I don’t know why – it’s obviously pretty illogical, but visceral reactions are visceral.

      • Peter says:

        Back when I was first getting to grips with the issue, there was a lot of spilled ink, or, err, spilled pixels, over the issue of “the space”. Prior to that discussion, no space seems to have been the norm, but people who preferred the space seem to have won out.

        There were various cited reasons which I didn’t agree with, and I had a friend who was both trans and a linguist and they agreed with me, but shrug. Connotations are a weird thing and in cases where something isn’t important it’s sometimes better to let people have what they want even if their stated reasons don’t always make sense.

    • Lignisse says:

      It didn’t offend me (I’m a trans woman) and I wasn’t going to mention it, but it did seem factually misleading, yes.

      There are a lot of biological sex differences, and with regard to many, but not all of them, I’m “biologically female” (mostly the differences that are mediated by hormones). For example, I have breasts, I have soft skin, I don’t develop visible muscles, I have fat deposits on my hips, I need about 2000 kilocalories a day, I get cold easily, I don’t sweat very much except with heavy exercise, and when I do it smells like woman sweat (yes, there’s a difference). (Also, there’s a difference in how I think – hard to describe, harder to prove, if you don’t believe it, you needn’t).

      With regard to other differences (the ones mediated by historical hormone levels during puberty), I’m “biologically male”. For example, I have a penis, I have broad shoulders and a broad ribcage, I have a visible lump on my trachea. (Odds are 90-98% that I have an XY chromosomal type, but never had it tested)

      But honestly? The differences in the former paragraph are way more salient in my life than in the latter. So yes, “biologically female” is closer to true than “biologically male”, though I wouldn’t choose to use either term myself – just “female” will do. “AMAB” (assigned male at birth) seems like the most useful term for the distinction Scott was trying to make.

  22. Deiseach says:

    I’m laughing, Scott, but not at you.

    All this biological determinism stuff strikes me as the modern version of phrenology: measure your finger lengths and we can predict your opinions on various topics is like feeling for he bump of philoprogenitiveness to see how many kids you are likely to have.

    • suntzuanime says:

      “Your claim that your lighter contains some sort of magical ‘fluid’ that causes fire sounds an awful lot like the discredited phlogiston theory.”

      • wysinwyg says:

        Not a great analogy. Any time I want I can pull out the lighter and start a flame quite consistently, but it doesn’t seem as though anyone can consistently replicate these digit ratio results.

    • Peter says:

      A: “I’m going to pick up C from the airport”
      B: “Why? Can’t he get here himself?”
      A: “Well, maybe, but public transport is less than ideal, it’s too far to walk or cycle, and he doesn’t have a car there.”
      B: “Well, how did he get there then?”
      A: “He flew.”
      B: “Oh, don’t give me that silly ‘human flight’ nonsense, you’ve seen cargo cults just as well as I have. Don’t say you believe that people actually fly by plane.”

    • Urstoff says:

      I doubt Scott is claiming that individual traits can be predicted with strong confidence based off of digit ratio. Statistical variation is a fact of the world, and these numbers are just statistical aggregates. There are lots of masculine men with large digit ratios and lots of feminine women with small digit ratios.

      • Did people who believed in phrenology claim that traits could be predicted with strong confidence or only statistically?

        • Peter says:

          The thing about phrenology – it became thought of as unscientific in about the 1840s (or so says wikipedia), which looks like it was before a lot of modern ideas about statistics came into play.

          Reading through, it looks like there’s a real mix of stuff. Some of it seems to be based on the mere idea that there are faculties that occupy areas of the brain, without needing to say what those areas or faculties there.

          I found “Thousands of people consulted phrenologists for advice in various matters, such as hiring personnel or finding suitable marriage partners”, which sounds like the sort of thing Deiseach was lampooning. I’ve not heard anything about this going on with digit ratios – although it wouldn’t surprise me if various people not connected with the research scene had got hold of it and were peddling advice loosely based on it.

  23. giant nanosanta says:

    In fact, our entire sample was heavily feminized compared to those of every other study; our men had more feminine digit ratios than other studies’ women. This is really weird because it’s a sample selected for very high mathematical ability – average IQ of 139, many math and physics PhDs, computer-related jobs as the most common occupations – but mathematical ability is usually linked to more masculinized digit ratio. Likewise, presumably this sample’s much more autistic than normal, but that’s also supposed to be more masculinized. I think maybe the measurement technique on the survey predisposed towards overreporting the ratio. Or maybe I screwed up somewhere and divided when I was supposed to multiply. Really that’s all I can think of.

    I have an alternative hypothesis: The LW/SSC sample is heavily selected for people who have spent their entire adult lives touch typing on cramped laptop keyboards, and their finger lengths have adapted to the ergonomic conditions in a Lamarckian fashion.

  24. Jaskologist says:

    If general surveys are any indication, I think you should really be dividing the women into “married” and “single” to find substantial gaps on feminism and politics. But I suspect that in this case, the sample sizes would end up far too small.

  25. Scott,

    You would find better signals from current, not prenatal T. Some people do get blood work done or saliva tests and have an accurate number, and for everybody else maybe you can use upper body strength. Guesstimate your max bench press, perhaps. (For the gym-averse: a push-up, according to my bathroom scale, is a 70% body weight bench press. Try it with legs on the sofa or a chair to have the same angle as a bench press. From the floor it is far too easy and not readily comparable.)

  26. God Damn John Jay says:

    “I looked at eight other political issues not clearly related to gender. None of these had a clear connection with digit ratio the same way feminism did. One issue, immigration, had a single correlation significant at the p = 0.05 level, but was likely a false positive.”

    ^ The post that launched a thousand blogposts containing a thousand variations upon “cuck” ^

  27. keranih says:

    How many “implausible” measurements were discarded, and what were the differences between the discarded observations and the missing observations and the observations you ended up using?

    • Scott Alexander says:

      I got rid of digit ratios outside the interval (0.8, 1.2), because most of them were people who didn’t understand the question and were saying things like “95”, which I assume was 0.95 but didn’t want to mess with, and (0.8, 1.2) was my eyeballed guess at what kind of digit ratios are possible in normal people.

      • keranih says:

        I was not clear, sorry.

        Of all the people who gave a digit ratio, how many did you discard?

        How did the characteristics of the people whose responses were discarded differ from the characteristics of those whose responses were kept, and how did they differ from the people who gave no response?

  28. Attaining Personhood 158 says:

    Aren’t the “typical” digit ratios wildly different for people of different ethnicities? This might be part of the explanation why even the average male respondent had a more “female” digit ratio than the average female respondent of the studies you read.

    • Scott Alexander says:

      The sample was mostly white, and most other people’s samples are mostly white.

      • Attaining Personhood 158 says:

        Wikipedia quotes a researcher saying “There’s more difference between a Pole and a Finn, than a man and a woman.” ( Both Poles and Finns would count as white.

        However, I’m quite unfamiliar with the literature and don’t know if this quote exaggerated – I’m posting because I’m hoping someone with more knowledge will chip in 🙂

        • TrivialGravitas says:

          Wikipedia’s source for that statement provides no result son a ctrl+f of ‘pole’ ‘finn’ ‘ethnic’ or ‘race’. I did find Mannings paper on Jamacian kids but it’s about the change in digit ratio with age, not racial differences.

          TL;DR Wikipedia is being Wikipedia.

  29. Peter says:

    Analysis question – how about confidence intervals on some of those correlation coefficients?

    The issue here is that if you say “A correlated with B” and “A’ did not correlate with B'”, but your confidence interval for the correlation between A’ and B’ includes the strength of the correlation between A and B, then you can’t really say that there’s a difference; there might be a difference (in fact there almost certainly is, if Meehl is to be believed) but your sample size is too small to detect it.

    (In particular, note that your female sample is much smaller than your male sample, so correlation sizes which are statistically significant for your male sample may not be statistically significant for your female sample.)

    What you need is something like a t-test but for correlations. Apparently there’s something called a Fisher r-to-z transform which may be what you want, but don’t trust me. I’m not a statistician, I’m just a data scientist.

  30. Urstoff says:

    What’s the correlation between digit ratio in men and the ability to grow a sweet beard?

  31. Apologies if I’m missing something basic here, and probably this is unneccesary to hypothesise on the limited data set, but could a plausible explanation be that masculine digit ratios are associated with factors that *force* a group of traits, whereas feminine digit ratios means that it leaves it up to environmental factors to decide? Then making a selection specifically from the feminine end of the spectrum would be selecting for a range in which the data was all over the place, with disproving or being in conflict with the overall correlations in the full spectrum? Or in other words, nerds do and think all sorts of crazy stuff that don’t fit the overall pattern?

  32. Orphan Wilde says:

    I think “Feminism” fails to capture “Feminism”. I think you really need to break that down into four elements; support for Sexual/Gender Equality, Sexual/Gender Self-Determination, Women’s Reproductive Rights, and Men’s Reproductive Rights.

    I think Feminism generally does a decent job on equality and women’s reproductive rights, but find it to be generally opposed to self-determination (not just limited to trans people, but with regard to anybody who likes a particular pre-existing gender role) and men’s reproductive rights.

  33. Deiseach says:

    Honestly, after re-measuring my digit ratios and reading that Wikipedia article, no socio-biologist ever gets to laugh about horoscopes ever again.

    “Your low digit ratio makes you good at maths and music, likely to be aggressive, and puts you at risk of ADHD” might as well be “As an Aries, the influence of Mars in your sign makes you a natural leader but prone to leap before you look”.

    If I’m doing it right, I have a masculine right hand (smaller index than ring finger) and a feminine left hand (equal more or less). Which do I take as my “real” ratio and what traits does this augur for me?

    The better explanation for my mixed hand ratios is that as Gemini, I fall under the influence of Mercury, which is an androgynous sign 🙂

    I genuinely think this digit ratio thing is best regarded as “Huh. So that’s a thing” and nothing else, particularly no conclusions drawn from it re: masculinity and femininity. Leave it on the level of reading your horoscope: a bit of silly fun (“oh, according to this, I should be good at music!”).

    • Peter says:

      I searched through the wikipedia article and failed to find the word “Your”. Could we have an actual quote please; it’s remarkably easy to point out something that sounds like astrology if you made it up to sound like astrology.

      (Yes, there are problems with the article, which a more pedantic style could have avoided. The first problem I spotted is “The 2D:4D digit ratio is sexually dimorphic: while the second digit is typically shorter in both females and males, the difference between the lengths of the two digits is greater in males than in females” where the second bit is missing a “typically” – a bit of googling suggests Cohen’s d gives values of 0.5 or so, which isn’t huge, much smaller than for adult height, and we all know short men and tall women. But I’d like to see people complaining about the actual problems rather than made-up ones.)

      • Deiseach says:

        People reading the Wikipedia article on digit length are going to be applying it to themselves: “So what does my high/low ratio mean? Oh, according to this – such-and-such indicates this and that”.

        If they write the article that sounds like a grab-bag of astrology traits, complain to them. Or the studies that throw the kitchen sink in – ADHD, music and aggression!

    • Urstoff says:

      Again, don’t confuse statistical aggregates as predictions of individual traits.

    • Scott Alexander says:

      I was careful not to say that digit ratio was linked to testosterone exposure in the post above.

      Also, what is up with his section numbering system?

  34. Troy says:

    Any correlations with charitable giving?

  35. Kyrus says:

    I don’t think that people were very careful about their measurements. What we can conclude from this “study” is that you can’t make studies on the internet that ask people anything more complicated than clicking boxes. I would advise against putting much weight on these results.

  36. Quite Likely says:

    How about this explanation: people who read this rationalist blogs and respond to these sorts of surveys are nerds. Nerd men tend to be less stereotypically masculine than non-nerd men. Nerd women tend to be less stereotypically feminine than non-nerd women. Maybe we’re just a bunch of self-selected womanly men and manly women and thus not a representative sample for how this sort of thing works in the general population.

  37. TomA says:


    Are you presuming that the samples collected within LW and SSC cohorts are representative of the US population as a whole? Your data suggest otherwise. Once you dip into small self-selected cohorts, statistical science becomes much less reliable as a source of inference.

  38. Goatstein says:

    Hey I rebutted your “Manufacturing Consent” review

  39. Sarah says:

    I think the obvious conclusion to draw from this an update in the direction of “digit ratio isn’t a gendered thing”, an update in the direction of “we did our study wrong”, a small update in the direction of “LW/SSC is atypical”, and no updates at all about feminism.

  40. Decius says:

    The conclusion I drew was that there was something different about your sample than about the samples used in the studies that you would have cited in a more formal context.

    I think that suggests a whole slew of questions.

  41. Besserwisser says:

    Feminine men being more likely to support feminism wouldn’t suprise me since feminists despise masculinity or, at the very least, gender-conforming. Women presumably don’t have to worry about being feminine enough and I actually have some examples of feminists despising femininity in women almost as much as masculinity in men. But then, men with high digit-ratios weren’t actually shown to be more feminine, so I’m at a loss.

    Transwomen being more critical of feminism than ciswomen is hardly surprising, given such wonderful works such as “The Transsexual Empire”. Some feminists accepting transsexuals is a relatively recent development. Then again, so is the case with most of society but I can’t fault them for being careful with whom they sleep. Also, transwomen being more feminine than ciswomen is indicative of both feminine traits being highly associated with dysphoria and some people who would be happier being trans unwilling to admit it because of societal pressures.

  42. Ruben says:

    In short, it’s likely that your study and all other studies are marred by (unconscious) biases in measuring, the garden of forking paths and publication (and more so than in other realms). These biases are in my view sufficient to explain your results. I’m a researcher who could have worked on this and I decided not to after finding out how horrible the research literature is. It’s building on sand IMHO.

    ## Details:
    1. I’ve measured hundreds of finger lengths using scanned hands for psych research. It is ambiguous, even if you follow a research manual to the t. People’s fingers curve differently, some have less pronounced/no visible creases at the palm. I was unaware of subject identity but not sex (it was in the file identifier, but you can also kind of recognise some male hands) while coding. It would have been easy for me to unconsciously bias the results slightly towards a desired state. You/LW have posted on digit ratio, so some people are somewhat aware of the research. This could bias correlations in expected directions and could make them significant. Recall the Bargh replication, where the experimenters could replicate the “old people priming makes you slow” when using stopwatches, but not when using objective light barrier timing. In general and counter-intuitively, there may be a lot more ambiguity about how to measure, aggregate and analyse digit ratio than neuroticism. Basically people can pick either hand or both and get it published and that’s not the only researcher degree of freedom. You can always go back to the hands and remeasure if you’ve got scans. I actually measured hands in this context, because results were not significant and the PI thought the previous coder might have been sloppy. I’m sure he would have double-checked in the reverse case too, ha ha ha.

    2. Digit ratio research is bad at its core. Martin Voracek and his collaborators have a research program that just shows all the ways it’s bad. The evidence showing it’s an index of (prenatal) testosterone isn’t even good. A lot of research comes from one group surrounding Manning. Most other groups can’t replicate their findings. When they meta-analytically examine the data, they find evidence of publication bias and true zero effects. They also find that people keep citing the flawed significant studies and ignoring the published evidence to the contrary. Worse, they’ve presented some research at conferences finding evidence of data-fakery for one group using Benford’s law (they didn’t name it, but I believe everyone present knew who it is). It’s also odd in that this sort of research keeps being done by non-anthropologists (ie. the people who don’t have calipers/know what they’re doing). Probably because it’s “so easy”. But even with callipers, there’s soft tissue to be pushed. At least for one study with calipers from the Manning affiliates, I know via an RA that one researcher flipped the caliper so he could see the readout while deciding how to measure. You shouldn’t do that and protocol said not to. That way it’s again easy to have bias sneak in.
    Voracek has also been given trouble trying to publish his p-curves etc., so there’s a mafia. If you look at his publication history he’s gone from believer to skeptic, which is a good sign I guess.

    3. As others have pointed out, your language regarding p values is misleading in places (they don’t approach/trend). It’s hard to talk properly about p values without sounding like a prick, so it’s probably best to avoid them in most fuzzy research like anth, psych, soc but I can’t say the Bayesian alternatives are as easy to use yet.

    • TrivialGravitas says:

      It sounds like what we need is software to do the job, and reduce the whole thing to ‘step one put your hand on the scanner’.

      • brad says:

        It sounds like what you really need is x-rays to measure bone lengths not optical scans measuring fuzzy skin features. But you aren’t going to get that one by an IRB.

    • Thank you for the level of detail.

      I was wondering whether it was worth taking the trouble to measure my ratio, but I’ve made my contribution to finding out that science is much harder than it looks.

  43. Really forget anything else but measuring hormonal levels or their older, like, prenatal effects. Everything else doesn’t have a rat farts chance in a hurricane of being objective. Even if we were trying 100% honestly we would be influenced by old vs. changing culture, further confounded by immigration or simply globalization with cultures that are LESS feminist than the 1950’s West, but of course a lot of people are not even going to try it honestly, just inject ideology and counter-ideology into it so basically it is the worst clusterfsck evar and just forget things like these tests.

    While if I just make a saliva or blood hormone test, at least I get an objective number. is apparently shipping this Xmas! I was never so excited about a Silicon Valley fad product for a long time. I wonder if Scott or the Mealsquares guys are paying attention to Really they should, IMHO. I would be pleasantly surprised if one of the inventors would now show up…

  44. JayMan says:

    Did you know that there’s probably nothing to digit ratio? It’s probably just hot air.

    For one thing, it certainly doesn’t measure in-utero androgen exposure, as fraternal twin girls with brothers don’t have more “masculine” digit ratios than those with sisters.

    Digit ratio is in fact heritable (as shown by twin studies), as are most things supposedly associated with prenatal hormone exposure.

    As well, I haven’t seen much by way of large meta-analyses of studies of the behavioral correlates of digit ratio to check to see if the hits aren’t all the results of publication bias.

  45. Q says:

    Sorry for not reading other comments, maybe I repeat something. But I think it is correct finding, that higly intelligent, math oriented men are less masculinised. The scientist Daniela Ostatnikova observed the same with gifted boys. She tested those pictures of mental rotation if the cube. Having less sensitive testosterone receptor was an advantage for the task. It was published in Neurosciencia or similar journal name. The research team was from Slovakia.

  46. squf says:

    I realize I’m much lower IQ than average around here, but I assume I’m also more masculine (wanna fight about it? I kid). For me its a simple nominal issue, why should I support feminism? I’m not feminine. We can move from object level to meta and dig deeper, but the preceding statement honestly suffices. I also find it practically stupid to correlate a high mathematical ability with masculinity, was Sulla known for his love of geometry? “Mad Dog” Mattis is manly. Terence Tao is not manly. Further, Heinlein isn’t manly either, and he was fervently contemptuous toward those who cannot “cope with mathematics”. Is math masculine? Signs point to no.

    It seems more likely to me that maths ability correlates with higher logical ability, logic is typically considered a masculine trait, and it should be easily agreeable that Mattis and Sulla are logical thinkers at least. Maths seems to be much more abstracted though, perhaps “purely logical”, its not like I have any clue what I’m talking about here — I can’t even pass remedial algebra. What I do know is man however; my paternal grandfather is a murderer, my father is a Marine, my same-father brother somehow wound up in a black street gang, and with several violent felonies under his belt. My Y-chromosome appears strong.

    Masculinity does not exist solely in the noumenal realm; – hence mathematical ability has no correlation to masculinity (math being solely noumenal). The entire point of what is manly is that it rests on phenomenal destruction, destruction which can be perceived with the senses. The point of it is to fight, to kill, and to die — preferably in defense of that which you love. Make an equation out of that if you must.

    I’ll also pitch in some further noise, might as well at this point, and agree with Alsadius. I can eyeball that my ring fingers are noticeably longer on each of my hands, but this is uselessly non-rigorous.