Here’s the talk presented Friday night by HJ Hornbeck. Below the fold is the full transcript including all notes and links, which he’s helpfully provided.
I believe if you visit the Youtube page you will see that HJ has adequately performed his intended task — dredging for MRA trolls and making them take anti-scientific positions. Because that’s what it takes to deny that feminism is right about its foundational claims: you’d have to be anti-science.
[This is a direct copy of my shooting script, including all my cues. It’s a condensed version of a much longer source document, which also contains all the bonus content, full citations, illustrations, and links to PDFs. To get your paws on that, click here.]
Hi, I’m HJ Hornbeck, a co-host of the Legion of Reason podcast and a former president of the UofC Freethinkers. Before I get talking about the evidence for feminism, I must let you know that I’ve made a rough transcript of what I’m about to say here [tablet]. This should a big help to the deaf, or the vast majority who think I’m talking too fast. I’ll also toss in a few footnotes, illustrations that you can actually see, and a little bonus content.
On with the show. Of all the published works I’ve looked at, Helen Woolley’s has been the most fascinating. The introduction to her second meta-analysis complains of a flood of new papers, mostly due to the feminist movement.
Summing up the wealth of data, she declares [tablet] “the truest thing to be said at present is that scientific evidence plays very little part in producing convictions.” She spends a fair bit of time discussing the contradictory theories, results, and interpretations that plague the current research.
One reason why I dig this review is that it was published in [tablet] 1914. Yes, for over 100 years science has been studying culture and gender. For twice as long, feminism has also been making claims about culture and gender. If feminism’s claims are out of step with the scientific evidence, then it’s at best a cargo-cult, taking the aspects of reality it likes and constructing wonderful fantasies around them.
So let’s cut through the fantasy and get to the fact. What does science have to say about the claims of feminism?
We’ll start big, with the feminist view of gender. The most definitive statement I know of comes from Simone De Beauvoir’s 1949 work, “The Second Sex.” I quote: [tablet]
“The biological and social sciences no longer admit the existence of unchangeably fixed entities that determine given characteristics, such as those ascribed to woman, the Jew, or the Negro. [tablet: 1949!] … But does the word woman, then, have no specific content? This is stoutly affirmed by those who hold to the philosophy of the enlightenment, of rationalism, of nominalism; women, to them, are merely the human beings arbitrarily designated by the word woman. Many American women particularly are prepared to think that there is no longer any place for woman as such; if a backward individual still takes herself for a woman, her friends advise her to be psychoanalysed and thus get rid of this obsession.”
A quick aside: most of what I’ll reference for the time being was written before the term “gender” was used to describe humanity. When you hear “sex,” think gender instead.
Anyway, the vast majority of the feminist movement still agrees with De Beauvoir. However, Woolley’s complaints about the pace of studies are even more true today, so I’m forced to rely heavily on meta-analyses. These take a big-picture approach, gathering as much high-quality data as they can and rigorously summarizing it. While the meta-analysis approach has issues, it remains the best method we have for digging up truth.
Helen Woolley’s 1914 paper was the first good meta-analysis of gender differences. She mentions two competing theories of difference: gender differences were either inherent to our biology, or they were socially constructed. [tablet]
“All of this group of men, in spite of their wide differences of opinion as to the nature of the psychological characteristics of sex, are convinced that they are inherent and are not to be explained by environmental influences during the life of the individual. … Differences which remain constant at different ages and in different countries must, they think, be inherent in sex itself. They do not seem to have considered whether or not there are factors in the social environment of sex which remain constant in all modern civilized countries.”
“When one turns to the books written more largely from the historical and sociological point of view, the trend of opinion is that mental differences of sex are of social origin. … They all lay stress on the view that social conditions account for most of the traits ordinarily considered feminine, and particularly for the limited accomplishment of women in art and science.”
Woolley gives no general summary, but she usually concludes there is no significant difference, or not enough data to conclude anything. Woolley does find some minor differences though: girls grew up faster than boys, but caught more diseases; girls were better at complex motor tasks, memory tasks; they occupied the higher parts of the IQ curve, and were absent from the lower parts; and finally, girls were better at school. =05:00= Boys had better reaction time, and excelled at simple motor tasks; better at reasoning, math and logic; better at exams; and better at free association, which might explain the gender gap in rap.
The next major review after Woolley’s came in 1927, from Florence Goodenough. Her overall conclusion: [tablet]
“The most outstanding impression which one gains from a comparison of these studies is the inconsistency of the various findings. Sex differences are shown, it is true, and while they are in general small, there are still a number of instances in which a sufficient number of cases have been included to give a reasonably high degree of reliability to the differences shown. The direction of these differences varies, however, according to the type of test used, and with the age and school status of the subjects.”
“As several writers have been at pains to point out, the practical import of sex differences in mental traits is negligible, since the amount of overlapping is so great that the small differences between the sexes are completely overshadowed by the great variations found to exist between members of the same sex.”
What about the non-significant differences? Again, we find girls excel in verbal skills, and boys exceed in math, but only in the later grades. Goodenough concurs with Woolley about memory, but is more cautious about motor skill, stating [tablet] “there is, unfortunately, but little in the data at hand which serves to indicate whether the popular opinion as to masculine superiority in tasks of this sort would be justified if the material used were equally familiar to both sexes.”
The next major review was from Harvey Lehman and Paul Witty in 1930, and they blast much of the literature for perpetuating falsehoods. They find seven key flaws, including: biased samples, small sample sizes, poor analysis, overgeneralizations galore, and an automatic assumption of differences being biologically innate, when social factors have not been ruled out. Overall, they agree with Goodenough’s review.
The next major one came in 1946. [book] The “Manual of Child Psychology” contained a single chapter devoted to sex differences. Titled simply “Psychological sex differences,” it was written by Lewis M. Terman, among others, and unlike the previous authors I’ve mentioned they favor biological explanations over social ones.
You sense this was a minority position, though; their leading theory was that hormone levels explained difference. While animal models provided good evidence for this, the technology didn’t allow them to accurately measure endocrine levels in human beings, depriving them of a smoking gun; in contrast, the social construct side had mountains of evidence in their favour, ranging from cross-cultural studies to the difference between stereotypes and actual performance. The best the innatists could do was declare the situation to be complicated and messy, with no single explanation for difference being sufficient.
Terman and his co-authors found significant sex differences, but for various reasons, they have been led [tablet] “to emphasize sex differences in central tendencies rather than the extent and frequency of sex overlap.” Besides, “the results are sufficiently consistent to leave little doubt as to what are the essential facts.”
When it comes to verbal ability, for instance, there is no difference. While small-scale studies sometimes contradict that, they usually have deep flaws that makes them unreliable.
As for memory, there is far more evidence for similarity than difference, with the studies that fail to find a difference outnumbering those that do. On general intelligence, the authors state that [tablet] “although the groups tested are often large[,] they can never be viewed as strictly random samples of the general population, since a number of factors affecting school enrollments[sic] are known to operate differentially for the sexes. In general, however, the sex differences found were so small and so inconsistent in direction that no positive claim could be made for the superiority of either sex at any age.”
They also have a brief note on variation. The notion that men vary more than women began fifty years earlier, to explain why the historical record contained nothing but high-achieving men. They just occupied the extremes, the rationale went, while women were less variable. As early as 1914, however, there was considerable evidence that showed little if any difference in variation existed. Every meta-analysis I know of concludes there’s no significant variation.
1974 brought us the most cited review of gender differences, creatively titled “The psychology of sex differences.” I feel bad about just quoting from the summary chapter, but I’ve gotta watch the clock. The following beliefs about gender were found to be false [tablet]
Girls are NOT more social than boys.
Girls do NOT have lower self-esteem.
Girls are NOT better at rote learning, and boys are NOT better at higher-level cognitive tasks.
Boys are NOT more analytic.
Girls are NOT less motivated to achieve than boys.
Girls are NOT biased to auditory sensation, and boys are NOT visual.
=10:00= The following are actual gender differences that Maccoby and Jacklin found good evidence for [tablet]:
Girls have greater verbal ability than boys. The differences don’t show until 11-ish, however.
Boys exceed in visual-spatial ability. Again, the differences don’t conclusively show until adolescence.
Boys excel in mathematical ability. There’s a lot of caveats, though; Maccoby/Jacklin admit the difference is small and could easily be explained by social factors.
Boys are more aggressive. The pair don’t give a magnitude here, but make it clear the difference is big, starts early, and gradually declines over time.
A number of claims didn’t have enough data to say either way, including [tablet]:
Tactile sensitivity. Apparently that was a thing?
Competitiveness. Culture, here, seems to be more important than biology.
Dominance. If the genders are segregated, boys tend to be more dominant. When they’re mixed, the evidence is ambiguous, and may point to any short term difference levelling out into long-term equality.
Maternal behavior. Cultural studies suggest girls are more nurturing, while observational studies show no difference whatsoever. If we use altruism as a proxy for nurturing, then there probably isn’t a difference.
Now, this meta-analysis marks the end of an era. Until until here, every review has been summarized by eyeball and intuition, permitting bias to creep in. The increasing spread of computers in the 1980′s, however, combined with easily searchable databases, gave us an effective tool for cutting through bias, and dramatically improved the quality of analysis.
The first to champion this new approach was Janet Shibley Hyde. In a 1981 paper, she reexamined the Maccoby/Jacklin review and found:
The main conclusion that can be reached from this analysis is that the gender differences in verbal ability, quantitative ability, visual-spatial ability, and field articulation reported by Maccoby and Jacklin are all small. Gender differences appear to account for no more than 1%-5% of the population variance.
Hyde next focused on the strongest claim of those two, that men are significantly more aggressive than women. More surprises appeared: adults displayed no difference in aggression, while pre-pubescent children displayed the most.
In conclusion, it appears that gender differences in aggression, like cognitive gender differences, are not so large as one might assume from the conclusion that they are “well-established” (Maccoby & Jacklin, 1974). Within-gender variation is far larger than between-gender variation.
But what really caught everyone’s attention was Hyde’s finding that the gender gap was narrowing with time. The floodgates opened, and everyone started pumping out their own large-scale meta-analyses.
In 1988, Hyde examined the verbal gender gap with improved methods and more data, and declared:
We are prepared to assert that there are no gender differences in verbal ability, … in American culture, in the standard ways that verbal ability has been measured.
This study also found no clear increase in verbal skill around age 11, contrary to Maccoby/Jacklin; that female researchers were more likely to find a gender difference; and yet again, the gender gap was decreasing over time.
In 1990, Hyde’s latest data told her the gender math gap was small, but when she only included studies that reflected the general population, it completely disappeared! Hyde did note a wider gap in some areas, but could explain those exceptions with social factors. And yet again the gender gap was decreasing with time.
Other authors have also studied the same topics, and more, and backed up Hyde’s findings of small-to-no gender differences.
Hyde, however, has been a favorite target of critics. Her infamous paper “The Gender Similarities Hypothesis” triggered four rebuttals in the very same journal it was published, of which I’ll mention two.
Richard A. Lippa criticized her for starting from the wrong hypothesis. He himself ascribes to what he terms the “gender reality hypothesis;” he assumes that there will be some gender differences, then once he finds a difference he tries to look for a reason it exists. Hyde’s assumption of the null hypothesis was too simplistic, for him; she should move past this heated and polarized debate over the existence of difference, and look instead for greater meaning.
G. E. Zuriff criticized Hyde for being too scientific. After all, everyone sees huge differences between the genders, including many keen observers of humanity. They don’t come to that conclusion by playing with numbers, running experiments, or weighing the evidence, they use psychology, not science. Hyde’s methodology is doomed to fail, and she should consider other ways of knowing. =15:00=
Whatever your stance, we find ourselves in a very weird place. It’s been over one hundred years since Woolley’s crack at the subject, yet we still don’t have a consensus on gender differences. Even the meta-analyses I’ve covered don’t always agree. Woolley found girls were better at language, Terman found no difference, Maccoby found they were better, Hyde and others found they weren’t. Woolley found almost difference in math ability, Terman and Maccoby found a difference in higher abilities, Hyde would flip between a difference due to higher education, no difference at all, then a difference due to high school. Outside of meta-analyses, we still find stark differences in individual studies that point every which way. How can we conclude anything from this mess?
Pretty easily, actually, with the help of firearms.
Machine guns are not precise weapons. Pull that trigger and bullets go flying everywhere. A rifle, in contrast, is precise; pull that trigger twice, and both bullets will land in roughly the same spot. That spot, however, may be a few feet to the right of where you were aiming. If the scope is misaligned or the barrel warped, a rifle becomes a biased and inaccurate weapon, even if it still precisely lands a bullet in the same spot.
OK, now let’s say I snag a paper bullseye from a shooting contest. [tablet] Bullets are scattered all over the place, but there are few clusters. What can you tell me?
Well, everyone could have used imprecise weapons. That’s not likely, and we’d expect an imprecise weapon to throw bullets around more randomly.
Or maybe they used very precise weapons, but on a very small target. A weapon’s accuracy would then dominate, and any bias would have a similar effect to imprecision, though we’d generally expect less clustering. Generally.
Or maybe the weapons were precise but inaccurate. The target was large, but bullets went almost every which way because everyone’s gun was biased. Almost, though, because bias itself can be biased.
I think you can see where I’m going here. If we were unbiased and accurate, we’d have long settled the sex difference question. If we were biased, or the differences were small, or our tools were inaccurate, then we’d have results all over the map. But the vast literature and use of meta-analyses rule out inaccurate tools, and the wild clusters of different and shifting conclusions point to significant bias.
So De Beauvoir called it; either the differences are so small they can be treated as non-existent, or we’re biased and in need of a shrink.
It’s important to note that “bias” isn’t synonymous with “personal prejudice.” [tablet] Here’s the standard deviation of variously sized random samples taken from a larger population. Note that it underestimates the population’s standard deviation for small values, making the distributions more concentrated and exaggerating difference.
Of greater concern is this horn surrounding the deviation. A random sample is a random sample, after all, and sometimes you pick values that have more or less variation than the overall pool. The inner horn is your 50/50 line; there’s a 50% chance you’ll pick a value from within there. The outer horn contains 90% of the values. Even with a sample size of 100, there’s a 50/50 chance your standard deviation will be off by 5% or more.
That’s a bit of a problem, thanks to the “file drawer effect.” Scientists are human, after all; we want to have an impact on our chosen field, push the boundaries and come to solid conclusions. Publishing a study that says “we couldn’t find anything” doesn’t seem to do that, so instead the results are stuffed into a file drawer.
[tablet] Problem: if the effect is relatively small, then tossing those null results distorts the scale of what we’re studying, letting the extremes dominate. My simulations show that published studies will tend to inflate the mean.
Similar effects apply to the standard deviation. [tablet] Publishing only the results that are statistically significant tosses out trials with high deviation, which underestimates the population deviation and exaggerates difference. This effect gets worse with smaller sample sizes, or effects. And random samples random – horns – et cetera.
That’s bad enough, but there’s another effect in play. Richard Feynman’s famous “Cargo Cult” lecture points out that scientists are more critical of their results if they’re unexpected or too different from the norm, causing them to look for ways to bring the data back “into line.” Flawed initial results do get corrected, but only slowly and gradually. =20:00= Feynman claims that nowadays “we don’t have that kind of a disease,” but Maccoby and Jacklin demonstrate a counterexample.
You might think the cost of a study would discourage tossing out results, but those two also point out that most studies aren’t targeting gender differences; their primary goal was to look at another question, the analysis around gender just happened to be on the path to that goal. They lose nothing by filing the gender analysis bit away.
[tablet-REDUX] We can now build up a picture of this system: a small study finds a big difference and publishes it. Other researchers jump on the bandwagon; those that get a weak effect will dustbin or “retool” their work, while those that agree, or disagree in a spectacular fashion, will get published. Later researchers increase the sample size or do a meta-analysis to cut through the confusion. Greater statistical power allows weaker effects to get published, and the observed difference shrinks. Disappointed, future researchers either look elsewhere, or split their meta-analyses into sub-populations in search of something they may have missed. Sample sizes drop, and suddenly differences re-emerge. The cycle thus repeats.
But just because our biases have a greater effect than gender differences, doesn’t mean gender differences are small. We need a way to judge the degree of difference. Let’s take a step back.
When we say “men are taller than women,” what do we really mean? I figure it declares “pick a random man and woman, and more often than not the man will be taller.” It says there’s this property that we call “height” that tends to be associated with men more often than women.
Note this goes both ways; if I was to hand you someone’s height, you should be able to predict their sex. Demographic data confirms this. For each possible height, we calculate the number of men and women who have that height. We chose our guess based on the biggest share, accept our losses, and then update the integral. In graphical form, we divide the biggest share by the total of all shares. [tablet]
For Australia in 1995, we find we can correctly guess if a person is male or female about 83% of the time. [tablet] Sri Lanka as of 2005 gives the same results. That means our failure rate is 17%, though, which seems pretty high next to the standard “p value” of 5%. Then again, height isn’t the sole way we determine gender. We’re fine with the existence of tall women and short men, because we have other ways of determining gender.
With our tool now in hand, we can measure gender differences! We’ll start with a null result, the difference in mathematical skill that Hyde found in 1990 within the general population. [tablet]
As you’d figure it’s 51/49, barely different from a coin toss. Hyde breaks down the studies into several categories, and not surprisingly found bigger differences in the smaller samples. Here’s the largest she found, which is for graduate students and certain colleges and has a solid social explanation: [tablet]
That’s still pretty small, we’d be wrong 40% of the time if we tried to predict gender that way. Still, it establishes the range of differences that could be explained or generated by social factors.
Let’s move on to spatial ability, a more fertile ground for gender differences. Here’s the difference in spatial visualization, the smallest according to the Voyer’s 1995 meta-analysis. [tablet]
That’s only 54% predictive. The largest difference they found was in “mental rotation ability:” [tablet]
Wait, our failure rate is 39%? How can they state “men are significantly better at mental rotation,” then?
Because the word “significant” has multiple meanings. A statistically significant result is one that we wouldn’t expect by chance. To properly test whether or not a coin comes up heads 11 times out of 20, I need to flip it over a thousand times. That result is highly significant, in a statistical sense, but is it personally significant? If I gave you a coin with that much bias and told you it was fair, would you ever notice I’d lied? I doubt it. A 60/40 split might be noticeable, in certain contexts, but if it only shows up with large sample sizes in laboratory conditions, we probably couldn’t detect it in our messy everyday lives.
But to some extent I’ve been talking from Mars while sociology has been focusing on Venus. Contemporary research has abandoned gender differences in favor of sex differences.
What’s the feminist view on those? Here’s what the Transfeminist Manifesto says: =25:00=
Though the second wave of feminism popularized the idea that a person’s gender is distinct from her or his physiological sex and is socially and culturally constructed, it largely left unquestioned the belief that there was such a thing as true physical (biological) sex. The separation of gender from sex was a powerful rhetorical move used to break down compulsory gender roles, but it allowed feminists to question only half of the problem, avoiding the question of the naturalness of essential female and male sexes. Transfeminism holds that sex and gender are both socially constructed; furthermore, the distinction between sex and gender is artificially drawn as a matter of convenience.
Sounds like there’s no consensus, with some branches of feminism content with sex differences. I don’t know which view is the most common feminist one, but within my peers and readings the Transfeminist view seems to rule the roost, so I’ll defend that.
How does sex differ from gender? Up until 1955, it didn’t; the meaning of “gender” relating to humans wasn’t even invented until then. Even afterwards, there was no consensus definition; depending on what you read, “sex” was anything that remained constant across cultures, or anything tied to biology, where “biology” meant anything from chromosomes to hormone levels. “Sex” was basically a synonym for ignorance. Time and time again, our knowledge would advance, an unexplained difference would get a cultural explanation, and it would be moved to the “gender” pile.
But the problem goes far deeper than that. Here’s Woolley again [tablet]
The ancient idea that the female is essentially an undeveloped male seems to be finally disproved by the fact that it requires more determiners—usually one more chromosome, or a larger sex chromosome—to produce a female than a male. When the additional sex chromosome was first discovered the assumption was that it determined maleness, doubtless because of the idea that the male was a more highly developed type.
The assertion that there are exactly two sexes is pretty new. The ancients observed that while humans generally fell into two categories, there was variation and exception. Plato thought there were three sexes: male, female, and “hermaphrodite,” or people who combined male and female characteristics. He thought category three was extinct, however other ancient historians note them as fact.
The most common solution to this problem was to declare there was just one sex, which developed in different ways. In the ancient Middle East, for instance, they thought everyone stored sperm in their brain. Coming from strong sperm meant you had hairy weights that drew your sperm down your spine to be ejected from a handy tube. If you came from weak sperm, your tube was inside out, no weights dropped, and you instead developed long hair on your head. This acted as a wick, drawing any sperm deposits into your body where they would compete with your own sperm.
Incidentally, that explains why Middle Eastern religions ask women to cover their heads; no-one likes it when you wave around your genitalia.
All variation was neatly explained: there was an ideal form, which we’ll call “male,” and all those “women” and “hermaphrodites” were just defective men!
That didn’t sit well in some quarters. By the 1800′s, scientists were trying to separate science from religion, and some proposed a newer model of sex that didn’t claim women were sin-filled defects: they were instead fundamentally different, a full-on second sex. Wanting to have and eat their cake, however, they argued the differences didn’t matter because the two sexes were equal where it counted most: our minds.
As a side effect, intersex people were tossed under the bus. The bigotry of the one-sex model came from asserting there was an ideal human nature; remove that, and the hate goes bye-bye without killing the model. In contrast, the two sex model demands a clean biological division, and those that don’t follow it must be defects otherwise they’d threaten the model. Opponents of the two-sex model used this to their advantage, and while the binarists looked for biological difference, they looked for biological similarity. For nearly a century, the one-sex hold-outs were victorious in biology, even as they lost ground in psychology.
That changed when sex chromosomes were discovered just before 1910. Many scientists including Woolley were quick to claim the X chromosome as the female counterpart to the Y. This belief pops up even today, and it’s absolutely wrong.
As far as we know, every single human being has an X chromosome. =30:00= It contains genes critical for the development of our skin, eyes, nose, intestines, muscles, and other necessary anatomy. The X even contains more genes related to testicular development than the Y chromosome. You cannot make a human being of any sex without an X chromosome.
On the flip side, you don’t need a Y chromosome to make a man. XX males are rare, constituting maybe 1 in 20,000 births, but they’ve been well documented by science. Nor, for that matter, does the presence of a Y chromosome guarantee maleness, as XY females also exist.
It’s tempting to ignore these exceptions. But how did we determine their sex? If chromosomes were the one true division, we’d have an error rate of exactly zero; instead, we’re overriding chromosomes and flipping the sex. This minority cannot be dismissed, if we want to discover a division of sex.
The next big hope was hormones. While there was some early work done in the 1850′s, the study of sex-related hormones didn’t take off until the 1930′s. Testosterone and estrogen compounds were found to have far-ranging effects on the human body, and in animal models were also found to alter behaviour. Scientists eagerly started digging here in search of difference.
This search has also carried into modern times. [tablet] Lauren Rosenberg and Sohee Park tested two groups of women, one set on the pill, one set not, and found a significant difference between them; one group was better at verbal memory tasks.
Wait a minute. I covered verbal ability earlier on, and back then we found no gender differences. But if there’s no difference between men and women, who have different estrogen levels, why would we find a difference within women? What’s the sample size here – oh, eight?! Wait, and they tried three different types of contrast analysis to extract a correlation? And- OOOOO, look what’s buried in the back: “we did not perform hormone assays to verify the actual hormone levels”
The field is littered with examples just like this, small sample studies which find sharp gender differences that larger studies have shown to be minimal or non-existent. On top of that, meta-analyses are few and far between here.
Take Premenstrual Syndrome. The science on mood and menstruation cycles is surprisingly lousy. Sarah Romans and her co-authors couldn’t even perform a meta-analysis on the evidence; out of 646 articles on the subject, only 41 reached their minimum requirements. The resulting old-school qualitative summary found only six studies of those 41 which demonstrated PMS as we think of it. Eighteen found a complicated relation between mood and menstruation, while 16 found no relation. PMS probably doesn’t exist.
Testosterone, though, provides a different story! There’s plenty of data for a meta-analysis, and so far four have been done. The latest is a rehash of an earlier analysis, done in 2001, which found a predictive rate of… [tablet] 54%.
I never did provide much detail on XY females. Some of them have “streak gonads,” or lumps of inert tissue. Others, however, have fully-functional gonads within their chest cavities, which pump out the same level of testosterone compounds you’d find in a man… and yet, no-one would dare call these people men.
But don’t worry, another potential division has come to the rescue: brain structure! See, the past few decades have given us some incredible tools for examining the brain, and a few neuroscientists have used them to dig for differences. As one study put it:
“Sex differences in brain anatomy may explain some documented differences in behavior. Women perform better than men on verbal and memory tasks, whereas men excel in spatial tasks”
Hmm, deja-vu. We’re presupposing a difference, and then searching for that difference. This is a big problem with neuro-imaging; these tools are fire-hoses of data, making them ripe for data-mining and false conclusions, such as discovering brain activity in a dead fish. And because of time and expense, these studies tend to have small sample sizes.
One major exception broke in December 2013. I’m sure you’ve seen the headlines: [tablet] “Male and female brains wired differently, scans reveal: Maps of neural circuitry show women’s brains are suited to social skills and memory, men’s perception and co-ordination.” The study in question is certainly the most definitive study of neurological difference to date, and it found that if you used brain wiring to predict the sex of a person, you’d get it wrong… [tablet] 40% of the time.
And that was the best-case scenario. There’s a lot more I and others have said about this study, but I’ll only mention one other objection. =35:00= Did these researchers find brain structure steering our social roles, or were social roles causing our brains to rewire and adapt? If you want these small differences to be intrinsic, you must assume the brain’s structure is static and eternal. Even a neuroscientist should know that’s poppycock.
But don’t worry, we’ve got ANOTHER candidate waiting in the wings: genes!
Decoding the structure of DNA started a minor revolution in biology. Thinking from a gene’s eye view explained many previously puzzling problems, like altruism or colonial insects.
Robert Trivers was at the forefront, and in the 1970′s he published an influential series of articles, one of which covered parental investment and sexual selection. A simple set of assumptions could be used to explain a wide variety of behaviour, which sparked a huge interest and created a new discipline: evolutionary psychology.
Trivers made three critical assumptions: there are only two sexes, fundamentally different from one another, and that a primary driver of that difference was sexual selection. Those assumptions weren’t necessary, but they made it easier to do the math and find examples.
As one, David Buss argues that men are bigger than women because parental investment demanded that men had the hunting duties, while women had no choice but to gather, which means that we should expect to find major biological differences in… spatial ability. Huh. He concludes a difference exists based on five studies, completely ignoring the Voyer’s meta-analysis of 286 studies, which as I showed earlier found a difference small enough to be explained by social factors.
Buss asserts that mental rotation is the best measure of hunting skill, arguing it was necessary for chucking spears and route finding. But mental rotation is just the ability to rotate things in your mind; that has nothing to do with manipulating objects while dealing with distractions, which is called spatial reasoning in the literature, or memorizing landmarks, which is a spatial memory task. In a remarkable co-incidence, mental rotation shows the largest gender gap of all types of spatial ability, while spatial reasoning shows less and Buss himself argues that women have better spatial memory than men.
Buss is cherry-picking the data to fit his theory. He must presuppose there are sex differences, though, as he’s using the same assumptions as Trivers. The theory predicts them, he believes they exist, so that settles it! If he doesn’t find a difference, he just keeps looking and waiting until one arrives.
Consider aggression. Scientists used to find big differences there, which they explained through evolutionary theory as the result of men competing for mates. Women weren’t thought to be aggressive, with David Buss himself declaring female aggression wasn’t worth studying. When they were studied, however, the difference in aggression disappeared. Not a problem! Human beings are social creatures who spend a lot of time in groups, hence getting along with one another would enhance our survival, and based on that evolutionary theory predicts there should be no difference in aggression. And hey, we don’t find a difference! Our theory is still correct.
If there was any difference in aggression, it had to be by type. Men preferred physical aggression, while women preferred indirect aggression through flexing their social superiority. But didn’t Maccoby and Jacklin fail to find superior social skills? Hmm.
No-one’s bothered to study female physical aggression exclusively, but they have studied aggression between the sexes in heterosexual relationships, and found… [tablet] women were just as prone to committing physical violence as men.
On the flip side, women have more parental investment than men, so we’d expect them think in terms of care-giving. Carol Gilligan proposed exactly that in her book “In a Different Voice,” which triggered a decade of research. Hyde neatly summed it all up in 2000. Based on 160 studies, she concluded that the sex difference in a moral orientation directed towards care was… [tablet] 56%.
Ok, what about risk-taking? Women invest quite a bit more in their offspring than men, so you’d expect them to be more cautious. There was a meta-analysis on this back in 1999, and the difference found in gambling risk was [tablet] 54% predictive.
The observed difference in risky physical activities, like climbing steep hills or playing in the street, was [tablet] 53% predictive:
The difference in risky sexual activities was [tablet] 51%:
But the biggest difference was in observed risks taken based on physical skill, like deciding which peg to toss a ring on, which was [tablet] 58% predictive:
=40:00= Ok, rewind a bit: there wasn’t much difference in sexual activities? Admittedly, that statistic is a bit misleading, because it groups all age categories together. If you divide the data down into smaller chunks, you can get a sex difference that’s [tablet] 62% predictive:
However, that’s based on a single study of 183 individuals, aged 10-13. While it’s tempting to ignore the small sample size, since that age range is roughly when menstruation begins, if you have a good diet, you then have to explain why a much larger sample pool of slightly older children showed a difference in sexual risk-taking that was only [tablet] 54% predictive:
The sex differences in… well, sex were at least interesting. Maybe here we’ll find the differences predicted by EvoPsych? Hyde has crunched the numbers, and the sex difference in the number of sexual partners was… [tablet] 55% predictive.
The sex difference in attitudes to extramarital affairs was… [tablet] 56% predictive.
This is freaky. We’re talking about sexuality, an area where EvoPsych claims to reign supreme over all other theories, and yet we’re not finding the sex differences it says should be there. Maybe we need to focus on specific claims.
Consider casual sex. In 1989, Russell Clark and Elaine Hatfield unleashed a few “confederates” onto a college campus, who would size up random strangers in an area, then walk up and proposition them. The results were striking: three-quarters of men agreed to an out-of-the-blue offer of sex, while not a single woman did the same. Conclusion: men are far easier to get into bed.
Problem: Women have been told, repeatedly, that men could sexually assault them and thus can’t be trusted. Maybe this social difference is at fault? Terri D. Conley decided to find out; by substituting written tests for the live experiment of Clark and Hatfield, she was better placed to figure out why women were rejecting these men. Sure enough, men were consistently perceived to be far more dangerous by women, then women were by men. Conley then spent quite a bit of time playing around with the basic scenario. When the person proposing sex was someone the person knew, and therefore trustworthy, the sex difference disappeared entirely. She also found that women preferred attractive men to wealthy men, by quite a margin, and that men ignored whether or not a woman was fertile when deciding to share the sack, both of which contradict the predictions of EvoPsych.
Women are also claimed to be more discriminating about potential mates, because of their greater cost in child rearing. Several large studies, most notably one by Robert Kurzban and Jason Weeden in 2005, have consistently found that in speed dating, men are less choosy than women. Their study in particular had a huge number of participants, in contrast to Clark and Hatfield’s piddling sample size. It looks like a solid result…
… but, Eli Finkel and Paul Eastwick noticed that every speed dating session had women stationary, while the men moved around. Their study flipped that around, and observed what happened when men sat around instead. The men not only became the choosy sex, they became just as choosy as women. This difference wasn’t due to sex at all, but a seemingly trivial social custom.
Now, you do not create consensus through a single study. Those two studies need multiple follow-ups before we can treat them as anything beyond interesting. The only way to achieve more in a single step is to kick out a core premise.
[book] In the intro to one of his textbooks on EvoPsych, David Buss writes:
It is part of the male lion’s nature to walk on four legs, grow a large furry mane, and hunt other animals for food. … It is part of the porcupine’s nature to defend itself with quills, the skunk’s to defend itself with a spray, the stag’s to defend itself with antlers, and the turtle’s to defend itself with a shell. All species have a nature; that nature is different for each species.
Buss is arguing some things related to sex are universal for all human beings. We’ll call these objective sexual truths.
But how would you establish their existence? Suppose I assert that all women possess vaginas, and find that out of 100 women, 100 possess vaginas. Have I demonstrated that the 101st woman will possess a vagina? Nope. How about a million? A billion? For what value of X do I demonstrate that example X+1 possesses a vagina, and why do I fail at example X-1? =45:00=
Philosophy majors know this as the problem of induction, and demonstrating objective sexual truths exist would solve it. What are the odds of Buss succeeding where two thousand years worth of philosophers have failed?
But there’s a bigger problem here. No other animal has the genetic code I do, and given how many possible genomes there are, none probably ever will. The same is true for you, your sisters and brothers, your pets, the fly zipping around your room, and so on. How then can you and I share an essential nature, but not our pets or that dang fly? I suppose you could argue parts of our genetic code are unique to us, but which parts? And if a future human being is born without a part, do they still share our nature? Ten thousand years ago we would have declared that drinking milk was contrary to our essential nature, and yet today quite a few of us enjoy a cold glass, because we…. “evolved.”
For the mind encased in Platonic blinkers, a rabbit is a rabbit is a rabbit. To suggest that rabbitkind constitutes a kind of shifting cloud of statistical averages, or that today’s typical rabbit might be different from the typical rabbit of a million years ago or the typical rabbit of a million years hence, seems to violate an internal taboo. …
The word “essentialism” itself wasn’t invented till 1945 and so was not available to Darwin. But he was all too familiar with the biological version of it in the form of the immutability of species, and much of his effort was directed towards combatting it under that name. Indeed, in several of Darwin’s books – more so in others than On the Origin of the Species itself – you’ll understand fully what he’s on about only if you shed modern presuppositions about evolution, and remember that a large part of his audience would have been essentialists who never doubted the immutability of species.
David Buss has been pushing evolutionary psychology for thirty years. What does it say about EvoPsych if their most vocal advocate doesn’t understand evolution? I think it says this is a cargo-cult science, which goes through the motions but only when they reinforce an existing bias.
Now, where does that leave feminism? Let’s return to XX males. From a gene’s eye view, the most plausible explanation is that some of the sequences on the Y chromosome jumped ship for the X. This means we could find a division between the sexes by analyzing the genes of an XX male.
Science has done that, and found the minimum number of genes you need is one. Sex-determining Region Y, or “SRY” as I prefer to call it, kicks off a long, complicated trail of dominos which alter fetal development. Cool! Except, there are about 20,686 other genes in the human genome, the vast majority of which are not contained on the Y. My body has the complete instructions for building a uterus, plus every gene that controls an evolved behaviour in women. What’s stopping it from activating the behaviour genes, by accident or design? Having a single switch sex determiner is actually an argument for similarity, not difference.
So it’s too bad we don’t have that single switch. While about four fifths of XX males have genes known to be unique to the Y chromosome, the remainder do not. We have no idea what’s up with that, but we do know there’s at least six genes involved in sex determination, bare minimum, and only two are on the Y. Human development is just a very messy, complicated affair, and most of us are ignorant of its diversity. Have you heard of micropenises? Aphalia? Ovo-testes? Mosaicism? And I’ve deliberately steered clear of sexuality and transgender people, as their inclusion adds entirely new dimensions of complexity.
At the same time, there are patterns in how we develop, some of which are more likely to happen than others. To represent these patterns, we have jointly developed some terms to tame the complexity, terms like “male” and “female.”
Sex really is a social construct, then, exactly as the Transfeminist Manifesto claims.
There are a lot more claims out there I could cover, like rape culture and objectification, but I see I’ve run out of time. Another time, perhaps?
In the meantime, any questions?