Article 1: Science Totally Debunks That Shocking Manifesto That Got a Google Employee Fired
Current research generally does not find evidence that variations in preferences, psychology, or personality stem from genetic or biological factors. Rather, they’re primarily attributed to culture and socialisation.
In his manifesto, however, Damore suggested the gender differences he lists do have biological components. One justification he gives for this belief is that the differences he mentions are “what we would predict from an evolutionary psychology perspective” and are “universal across human cultures.”
Damore didn’t cite any sources to back up his reasoning. However, a 2001 analysis of responses to a prominent personality inventory test found that “contrary to predictions from evolutionary theory, the magnitude of gender differences varied across cultures” – a direct contradiction to his argument.
Note that the paper is cited not to support that differences are small, but to say that there is no evidence that variation in preferences, psychology, or personality is genetic or biological. BUT NOWHERE IN THAT PAPER IT DISCUSSES THE ORIGIN OF PERSONALITY.
That paper featured in my first article about the Google memo. Read it. I extract a relevant bit:
- Differences in preferences (N=500,000) are found to be gigantic (Su et al. 2009). Women prefer working with people, men prefer working with things. This effects is also present in monkeys.
- There are 70% more dual talented (in maths and verbal skills) girls than boys, boys are twice as likely as girls to have strong math skills, but not strong verbal skills. People who only had only strong math skills as students were more likely to be working in STEM. People with both strong math and verbal skills as students were less likely to be working in STEM. Thus
“Results revealed that mathematically capable individuals who also had high verbal skills were less likely to pursue STEM careers than were individuals who had high math skills but moderate verbal skills. One notable finding was that the group with high math and high verbal ability included more females than males…
Our study provides evidence that it is not lack of ability that causes females to pursue non-STEM careers, but rather the greater likelihood that females with high math ability also have high verbal ability and thus can consider a wider range of occupations than their male peers with high math ability, who are more likely to have moderate verbal ability.” (Wang, Eccles, Kenny, 2013)
- Boys and girls are equally good at math on average. But the variance of the distribution of male skills is higher. (Lindberg et. al 2010)
- At the highest levels of mental ability, there tends to be two men per each woman (Johnson et al. 2008, Deary 2007)
- Male-female differences in math skills are not due to stereotypes (Flore & Wicherts, 2015)
- Women want to work less and have more flexibility (Lubinski et al. 2014), and this explains why there is no ex-ante wage gap in industries with constant return to hours, and large ones in industries where working more increases productivity more than proportionately (Goldin, 2014)
- Finally, differences in the mean can become magnified when one gets to the tails of the distributions.
About the second paper, Varinsky forgets to mention another key bit: That personality differences are greater in developed countries
Contrary to predictions from the social role model, gender differences were most pronounced in European and American cultures in which traditional sex roles are minimized.
And contrary to what the article says, the article does not deny (nor confirms, fair)a genetic origin for the personality differences that the paper finds. But what is more damning, the paper does say that
the general pattern of gender differences is similar across cultures, there is also variation across cultures, especially in the magnitude of gender differences.
Also, if you cite Hyde (2004), don’t forget Hyde (2014)
Hyde (2014), for example, reviewed several psychological sex differences and concluded there are relatively moderate to large sex differences in spatial rotation abilities, agreeableness, sensation seeking, interests in things versus people, physical aggression, certain sexual behaviors (e.g., masturbation and pornography use), and attitudes about casual sex. Smaller sex differences exist in measures of gregariousness, reward sensitivity, conscientiousness, negative affectivity, relational aggression, and self-esteem. Some of these sex differences persisted in size across cultures and time periods, others did not (see also, Lippa, 2009; Schmitt, 2014).
Rather than taking the average sex difference across each psychological dimension on its own, Del Giudice et al.’s multivariate method is to examine all psychological dimensions under consideration simultaneously (controlling for collinear overlap among dimensions). From a multivariate perspective, lots of small ds may be “additive” and create “planetary-size” sex differences when examined together (e.g., Del Giudice et al., 2012, found less than 10% overlap in men’s and women’s personality traits when looking across 16 dimensions simultaneously). By thinking about sex differences in terms of multidimensional space, this approach is probably a fairer evaluation of whether men and women differ, overall, within a particular multidimensional domain (such as “personality” or “cognition”).
Most importantly, the variance ratio, a measure of how “fat” the tails of a distribution are not one: men show more variation in mathematical, verbal, and spatial performance. This matters because even if the means were equal, higher variance alone could lead to overrepresentation of men.
(ht Garett Jones)
If that is not enough, a more recent cross cultural study(De Bolle et. al 2015)
Sex differences in college-age and adult personality have been systematically examined across a wide range of cultures (Costa, Terracciano, & McCrae 2001; McCrae & Terracciano, 2005b; Schmitt, Realo, Voracek, & Allik, 2008). Results suggest that the overall pattern of sex differences is similar in most cultures if one considers only the direction of effects. For example, women are rated as higher than men in Neuroticism in 49 of 50 cultures (McCrae & Terracciano, 2005a).
The biological changes during adolescence, such as the hormone-driven development of secondary sex characteristics, typically begin around age 11 for girls and age 13 for boys (Marshall & Tanner, 1986). These biological changes have been found to contribute to changes in how individuals behave and interact with the social environment. More specifically, whereas girls tend to display a stronger affiliative style than boys even before adolescence (characterized by a preference for close emotional communication, intimacy, and responsiveness within interpersonal relationships), this affiliative orientation intensifies during adolescence (Larson & Richards, 1989). Research in nonhuman mammals has suggested the existence of biological processes (i.e., changes in circulating gonadal hormones) underlying this intensification of affiliative behavior (Cyranowski, Frank, Young, & Shear, 2000). From a FFM perspective, affiliative behavior is closely linked to A. Therefore, it is expected that the sex differences for A (with females typically scoring higher than males) should already be observable at age 12 and become larger thereafter
Clearly, one cannot examine hypotheses about cultural variations in sex differences in personality traits or their developmental course unless one samples a range of cultures. At the same time, it is important to note that cross-cultural studies are valuable even if no cultural variation is found. If studies are confined to a single culture, there is no way to determine whether sex differences are attributable to cultural norms, recent historical events, or human evolution (Buss, 1995, 1997). To the extent that similar patterns are seen despite differences in culture, biological and evolutionary theories of sex differences become more plausible.
In conclusion, the present study indicates that culture is probably not a substantial source of the inconsistencies that characterize the literature on sex differences in adolescents’ personality. Age, by contrast, was found to significantly affect sex differences for several personality traits
She then says
To back up the “people over things” hypothesis, Damore cited a study published in the journal Social and Personality Psychology Compass in 2010; however, that work never suggests that the gender differences it lists have a proven biological basis.
In fact, the study says the opposite: “Although most biologic scientists accept that sexual selection has led to sex differences in physical traits such as height, musculature, and fat distributions, many social scientists are sceptical about the role of sexual selection in generating psychological gender differences.”
The second paragraph is facepalm-inducing. The paragraph she quotes is part of the introduction, where theories are presented. It is not a conclusion. In the exact same section it also presents the other view:
Evolutionary theories of gender difference propose that because women and men have somewhat different reproductive natures (e.g., women invest more in offspring than men do, both physiologically and behaviorally), the two sexes evolved to have somewhat different traits, particularly in domains related to reproduction (Buss, 2008; Geary, 2009). In the realm of personality, higher male levels of aggressiveness, risk-taking, and status-seek ing presumably evolved as sexually selected traits that fostered male dominance and helped ancestral men attract mates.
Damore did well in citing the paper, because it does show that the differences are universal, and big
For the people–things dimension of interests, the results in Table 1 are clear, strong, and unambiguous. Men tend to be much more thing-oriented and much less people-oriented than women (mean d = 1.18, a ‘very large’ difference, according to Hyde (2005) verbal designations). The Su et al. (2009) meta-analysis generated the smallest effect size (d = 0.86). However, as Su et al. note in their paper, a number of the interest inventories that fed into their meta-analysis used item selection strategies intentionally designed to reduce gender differences. Thus, the Su et al. estimate for the overall gender difference in people-versus-thing orientation is almost certainly an underestimate.
Prepare your hand and sink it deep in your face as you read page six of the paper:
If gender differences are highly consistent across cultures – i.e., if they are impervious to variations in gender ideologies and the strength of gender roles – then biologic accounts of gender differences gain in plausibility.
(But again, to play devil’s advocate: Social role theorists propose that virtually all existing societies are patriarchal, albeit to varying degrees. Thus, they argue that cross-culturally consistent gender differences reflect pervasive partriarchy [see Eagly & Wood, 1999, 2005].
At the same time, social role theorists predict that cross-cultural variations in gender differences should systematically covary with cross-cultural variations in the strength of gender roles – a prediction that is contradicted when gender differences are extremely consistent across cultures).
The size of gender differences, considered across cultures, also constitutes a kind of evidence that can be used to infer possible causes of gender differences. On the one hand, gender differences that are small and variable seem more likely to reflect the vicissitudes of social and cultural pressures (Hyde, 2005). On the other hand, gender differences that are large and consistent over time and across cultures seem more likely to reflect underlying biologic ‘presses’ or ‘biases.’
Isn’t this enough for you? The biological hypothesis predicts a pattern. The sociocultural hypothesis predicts another pattern. The observed pattern is the first one. How on Earth does the sociocultural hypothesis explains that the differences survive turning the dials of sexism from the minimum to the maximum? Only if you consider that, say Sweden and India are equally regressive in terms of how they view and treat women you could still maintain that it is culture. To people who think that, I just say sapienti sat.
A 2000 review of 10 studies related to gender differences in empathy also suggests men and women don’t have innate differences in this area. The researchers found that such distinctions were only present in situations where the subjects were “aware that they are being evaluated on an empathy-relevant dimension” or in which “empathy-relevant gender-role expectations or obligations are made salient.”
As I am not as familiar with this particular literature, I’ll pass on this paragraph. So far I have shown that all the arguments against the manifesto fail. Not all the claims in there have to be true for the whole to be broadly true. So I will accept, arguendo, that this is correct, and move on.
“The data on occupational interests do reveal strong male preferences for working with things and strong female preferences for working with people,” Grant wrote in a LinkedIn essay responding to Damore’s claims.
Again, see the SSC post.
In the memo, Damore suggested that women are biologically prone to express their extraversion as gregariousness instead of assertiveness, and to be more agreeable than men.
That difference, he claims, “leads to women generally having a harder time negotiating salary, asking for raises, speaking up, and leading.”
Again, Damore didn’t cite any evidence for this part of his argument. A 2005 analysis of 46 meta-analyses of gender differences suggests it’s false.
That analysis is Hyde’s again. That particular claim, that there are differences in negotiation outcomes, is minuscule (d=0.09), as are claims about gregariousnes (d=-0.07), but not for assertiveness (d=0.51). I don’t know why Hyde reported that.
Still, if one goes to the paper cited for gregariousness, Feingold (1994), the effect size cited there is d=-.15. Still small, no doubt, but positive, and far from zero.
However, the effect size of -. 15 for gregariousness, although very small, indicated greater female gregariousness
And a meta-analysis of leadership effectiveness published in 2014 suggests that when it comes to others’ evaluations of leaders (as opposed to the leader’s own perception), “women are rated as significantly more effective than men.”
When looking at self-ratings, however, “men rate themselves as significantly more effective than women rate themselves.”
That suggests that context and learned expectations are responsible for some observed gender disparities.
This is orthogonal: That evidence, if anything would account for differences in the proportions of leaders and non-leaders. Not differences between fields of study or work.
Damore also suggested that women are biologically prone to feel higher levels of stress and anxiety, and posited that difference might contribute “to the lower number of women in high stress jobs.”
The only source he gave for this information is Wikipedia. However, the misconception might have stemmed from analyses of the Revised NEO Personality inventory (the prominent personality test mentioned above).
On the test, according to a 2001 secondary analysis, women reported themselves to be higher in neuroticism.
But those responses are based purely on self-perception (which is heavily influenced by social and cultural factors) so it’d be problematic to consider that a biological difference.
How convenient of a critique! Yes, indeed, self-perception is not the best indicator there is for personality differences, that is totally correct. But that cuts both ways: many of the articles cited in the debate also use self-report. Alas, if one instead uses observer reports, the facts stay as they are: Women are, on average, more neurotic (d=.54 for adults) (In the sense of the Five Factor Model of personality), and thus feel more stress and anxiety. (McCrae et al., 2005)
In fact, using third party measures leads to increased effect sizes! Yes, self-report introduces bias, but it is bias in the direction of less differences, not more, in this case.
“Women on average look for more work-life balance while men have a higher drive for status on average,” Damore wrote.
As evidence for this, he cited a 2006 paper published in the British Journal of Guidance and Counseling.
That article highlights the fact that more women value a balance between their professional and home lives than men. It also suggests that men are more likely to make their careers their first priority.
However, nowhere does that paper suggest that these preferences come from biological or evolutionary differences between the sexes.
In fact, it makes this caveat: “They are differences of degree, with large overlaps between men and women.
They are not fundamental qualitative differences, as often argued in the past in order to entirely exclude women from ‘male’ occupations such as management, the military and the professions.”
Here she agrees with the claim, but disputes that the source is genetic. Okay, the study is only for the UK. One cannot draw a general conclusion for that. But if one has accepted everything else, one will also expect that this preference will replicate across the word. Unfortunately we don’t have those studies yet.
Thus ends my review of the first article.
Suzanne Sadedin, Ph.D response in Quora
Sadedin has a PhD in evolutionary biology, so she must surely have read the relevant literature. Let’s say what she says. Among the critiques I’ve seen it is one of the best, though in the end it still doesn’t succeed. Her summary. My comments in brackets
- argues for biologically determined sex differences in personality based on extremely weak evidence [As we have seen, the evidence is solid. More on this at the end of the document]
- completely fails to understand the current state of research on sex differences, which is based in neuroscience, epigenetics and developmental biology [Okay, let’s see what she has to explain]
- argues that cognitive sex differences influence performance in software engineering, but presents no supporting evidence. Available evidence does not support the claim. [THIS IS FALSE. Demore says that differences in traits influence the distribution of genders in tech (and Google), not that *within* tech and Google those differences matter. In fact, if we repeated the studies that we have been discussing with just googlers, we would find smaller differences. Because it is a self-selected sample.]
- fails to acknowledge ways in which sex differences violate the narrative of female inferiority; this shows intellectual dishonesty [?]
- assumes effective meritocracy in its argument, ignoring both a mountain of conflicting scientific literature and its own caveats (which I can only assume were introduced to placate readers, since their incompatibility with the core thesis is never resolved) [Let us study that, then]
- makes repugnant attacks on compassion and empathy [There are good arguments against compassion and empathy. In so far as we see them as emotions, they can bias us. Paul Bloom, cited by the manifesto, has a book where he discusses this. He is against empathy, for for cold, rational, compassion. Demore probably agrees with this]
- distorts and misuses moral foundations theory for rhetorical purposes [Maybe, let’s see!]
- contains hints of racism[ :O!]
- paradoxically insists that authoritarianism be treated as a valid moral dimension, whilst firmly rejecting any diversity-motivated strategy that might remotely approach it. [Nowhere he says that]
- ultimately advocates rejecting all morality insofar as it might compromise the interests of a group. [Oh, really? Where? ]
For the first claim about biology and gender differencs, she has this to say
We do have evidence for some of these to some extent for some gender differences in behavior. That, however, does not imply what the author thinks it does.
His implicit model is that cognitive traits must be either biological (i.e. innate, natural, and unchangeable) or non-biological (i.e., learned by a blank slate). This nature versus nurture dichotomy is completely outdated and nobody in the field takes it seriously. Rather, modern research is based on the much more biologically reasonable view that neurological traits develop over time under the simultaneous influence of epigenetic, genetic and environmental influences. Everything about humans involves both nature and nurture.
For an accessible introduction to sex differences in their developmental context, see:
One. Reference. Against a pile of research. So I’ll reply with just two. One, that the book has been heavily criticised and considered debunked. Second, that her more recent book also has important flaws. I mistook the book for Cordelia Fine’s for some reason. As soon as I can, I will edit this and address Pink Brain, Blue Brain. Apologies.
Do women and men differ in personality traits?
It’s true that women and men, on average, have been found in some studies to differ in empathizing/agreeableness, systematizing, gregariousness versus assertiveness; and neuroticism. There are also conflicting results, asdescribes.
That is the Varinsky article that I just criticised.
Now, to the meatier part. Saledin quotes the Google memo:
research suggests that “greater nation-level gender equality leads to psychological dissimilarity in men’s and women’s personality traits.” Because as “society becomes more prosperous and more egalitarian, innate dispositional differences between men and women have more space to develop and the gap that exists between men and women in their personality becomes wider.”
The internal quotes are from.
At first glance, this seems compelling. Let’s look more closely at this paper. Specifically:
“these changes appear to result from men’s cross-cultural personality variation. In more traditional and less developed cultures a man is, indeed, more like a woman”
And she offers this critique:
Hmm. That sounds a little different, now. In fact, Table 2 shows that, after controlling for human development index, the only gender equality-related factor that predicted gender differences was the ratio of female smokers. In other words, gender equality in general doesn’t change women’s personalities, or the difference between men and women. Rather, human development index changes men’s personalities much more than women’s.
That doesn’t support the claim that gender-liberal societies allow men and women to express innate differences more freely. If that interpretation were correct, women and men should diverge in gender-liberal societies independent of egalitarianism. Instead, men change personality in more egalitarian societies regardless of gender issues; women don’t.
How can we explain that? Maybe personality differences are mediated by power. It makes sense that relatively powerless individuals should be more agreeable and socially alert, less assertive, and more fearful/neurotic — that’s simply rational. How does it interplay with gender? In hierarchical societies, most men are (like most women) subordinate to a powerful minority, so the average man would act much like the average woman. In relatively egalitarian societies, men on average are less subject to oppression by other men, but women still remain low-powered on average relative to nearby men. We already know that.
That’s just one possible alternative interpretation. There are others. The point is, the study quoted by the author doesn’t come anywhere near demonstrating his claim.
This is a case of what Garett Jones calls the Everest regression. He says that controlling for height, the atmospheric pressure there is not low. Or as I say, controlling for latitude, the Sahara desert has good weather.
The error here is that HDI and gender equality are substantially linked. Controlling for HDI or GDP is like controlling for gender equality. As a general case, all good things are correlated: technology, moral progress, GDP, country IQ, industrialisation tend to be coupled. We don’t need power to explain those differences.
Do sex differences make women less suited to be software engineers?
I’m simply stating that the distribution of preferences and abilities of men and women differ in part due to biological causes and that these differences may explain why we don’t see equal representation of women in tech and leadership. Many of these differences are small and there’s significant overlap between men and women, so you can’t say anything about an individual given these population level distributions.
At what point did we jump from talking about personalities to abilities? It’s a massive leap to conclude that a slight difference in averagepersonality must undermine women’s professional abilities in software engineering.
Good point. Though there are differences in abilities between men and women too (See my previous two posts), Demore does not discuss ability. He surely must have had ability in mind, but maybe he decided not to write about ability to make the memo less controversial, but ended up slipping. Nonetheless, had he written about ability, he would probably have been right too.
Sex differences in cognitive abilitiesso it’s intriguing that Damore chooses to ignore this vast literature to focus on personality. The reason, however, quickly becomes clear when we look at the evidence: namely, there’s women should make worse programmers. On average, women score slightly worse on certain spatial reasoning problems and better on verbal tests. Their overall problem-solving abilities are equal. Women used to score worse on math, but inclusive environments negate that difference. Even the (relatively robust) difference in spatial reasoning can vanish when women are asked to . The only of coding competency by sex found that women were more likely than men to have their GitHub contributions accepted — but if they were project outsiders, this was true only if their gender was hidden.
The paper says that initially, mental rotation differences were moderaly large, d=.59, for men primed male and women rimed female. (p=0.01). For men and women both primed male, the effect was d=0.01. But what is the p-value or that? Well, p=0.94. Yes, 19 times larger than the standard 0.05 cutoff commonly accepted for statistical significance. For the whole set they report statistical significant results, but no effect size. We can also study statistical significance in the extreme case: female primed men and male primed women. If we plug in their data in a Welch’s t-test calculator, we get a p-value of 0.61. Again, not statistically significant.
The github paper is interesting. But is not at odds with the evidence so far: Again, the population in Github is self-selected. The average female Github user is not representative of the total population of females. Scott Alexander has a writeup about that paper here.
As Yonatan Zunger, empathy and collaboration are also central to competency, especially at senior levels. Published results confirm this: in a that attempted to identify the factors that influence software engineers’ success, the three most important attributes were: “team oriented”, “seeks help” and “helps others”. Neuroticism might hold women back from promotions, but there’s no evidence it makes them worse at their jobs.
Understanding others and playing along in teams is important, yes. But this is again compatible with the populations of men and women having differences in empathy and emotionality which would drive them to different careers. Within careers, we would observe less differences.
Thus, to say there’s “significant overlap” in male/female abilities is a massive understatement. There’s no evidence that any known sex differences make women worse at software engineering.
There is a significant overlap, yes. But if we look at the tails, as I’ve been stressing over and over, one can still see massive differences.
I tend to agree that the average women is as good as the average man at programming. That alone wouldn’t be sufficient against there being more men at the higher and lower ends of the ability distribution.
Abilities, anyway, are just one part, probably a secondary part, of the reason for different skills. Preferences are the main driver.
I have not read Halpern’s book. But a cursory googling reveals what Halpern thinks about gender differences:
In her preface to the first edition, Halpern wrote: “At the time, it seemed clear to me that any between-sex differences in thinking abilities were due to socialization practices, artifacts and mistakes in the research, and bias and prejudice. … After reviewing a pile of journal articles that stood several feet high and numerous books and book chapters that dwarfed the stack of journal articles … I changed my mind.”
Why? There was too much data pointing to the biological basis of sex-based cognitive differences to ignore, Halpern says. For one thing, the animal-research findings resonated with sex-based differences ascribed to people. These findings continue to accrue. In a study of 34 rhesus monkeys, for example, males strongly preferred toys with wheels over plush toys, whereas females found plush toys likable. It would be tough to argue that the monkeys’ parents bought them sex-typed toys or that simian society encourages its male offspring to play more with trucks. A much more recent study established that boys and girls 9 to 17 months old — an age when children show few if any signs of recognizing either their own or other children’s sex — nonetheless show marked differences in their preference for stereotypically male versus stereotypically female toys.
Halpern and others have cataloged plenty of human behavioral differences. “These findings have all been replicated,” she says. Women excel in several measures of verbal ability — pretty much all of them, except for verbal analogies. Women’s reading comprehension and writing ability consistently exceed that of men, on average. They outperform men in tests of fine-motor coordination and perceptual speed. They’re more adept at retrieving information from long-term memory.
Men, on average, can more easily juggle items in working memory. They have superior visuospatial skills: They’re better at visualizing what happens when a complicated two- or three-dimensional shape is rotated in space, at correctly determining angles from the horizontal, at tracking moving objects and at aiming projectiles.
Navigation studies in both humans and rats show that females of both species tend to rely on landmarks, while males more typically rely on “dead reckoning”: calculating one’s position by estimating the direction and distance traveled rather than using landmarks.
Short term memory sounds useful for programming, but the key take here is that Halpern actually lends support to the view that there are biologically based differences in cognitive ability.
How about preferences? It’s worth remembering that many of the first programmers were women, and that they made enormous contributions to developing the field of computer science. Female participation only declined when programming became a lucrative, gender-stereotyped career.
In my previous post I showed that such a thing is unlikely. Yes, many women were among the pioneers of computer science. But for the same reason there are so many women in CS in India. At the time, men were overrepresented in Mechanical engineering and other disciplines more. Potential earnings may also play a role, but that need not involve gender stereotypes. And as I showed in my other post, if one looks at India, the proportion of women who study CS is similar to that of the US.
All in all, we have no reason to think female software engineers should perform worse at software engineering based on female trait distributions. And there’s a huge amount of evidence that promoting diversity.
At least one of those studies has failed to replicate. The author replied (Herring, 2017), saying that his thesis holds… with a p-value<0.1 . Again, not statistically significant. The R² for the correlations are around 0.097 and 0.13. Plus the causal mechanisms for this are dubious, as there is no relation (Horwitz & Horwitz, 2009) between biodemographic diversity ( and team performance . What does improve outcomes is task-related diversity
Bio-demographic diversity represents innate member characteristics that are immediately observable and categorized (e.g., age, gender, and race/ethnicity) whereas task-related diversity is acquired individual attributes (e.g., functional expertise, education, and organizational tenure) that have been postulated to be more germane to accomplishing tasks than bio-demographic diversity.
Sadedin then talks about estereotype threat
We know that negative stereotypes. We know that influences our judgement of others’ competencies. Consequently, whenever there’s significant cultural prejudice against certain groups, as there is with female software engineers, we expect to see inequalities emerge. So it’s implausible to attribute these differences to biology alone. When we know that competent people are being held back by prejudice, for that via strategies that enhance diversity.
Sigh. Stereotype threat is empirically dubious (Wicherts & Flore, 2015). While a meta-analysis finds a small effect size, g=-0.22, strong pulication bias was found, so one should be cautious. The authors conclude,
To be more explicit, based on the small average effect size in our meta-analysis, which is most likely inflated due to publication bias, we would not feel confident to proclaim that stereotype threat manipulations will harm mathematical performance of girls in a systematic way or lead women to stay clear from occupations in the STEM domain. Of course, we do not challenge the fact that stereotypes might strongly influence a person’s life under unfortunate circumstances; however, we want to avoid the unjustifiable generalization that stereotype threat, based on the evidence at hand (i.e., the average small effect that stereotype threat manipulations have on instant test performance within this meta-analysis), generally leads to lower math grades and women leaving the STEM field
This is what the funnel plot looks like for this study:
It should be roughly symmetrical, but instead we find an skew to the left.
True, gender gaps don’t always imply sexism. But sexism always implies sexism. We don’t need to infer the existence of sexism from the gender gap in software engineering, we can see it in the countless expressions of misogyny we hear from software engineers. Hiding sexism behind a mask of pseudo-rational argument doesn’t make you any less sexist — as Damore’s document illustrates.
You can see sexism. But can you eyeball effect sizes easily? That there is sexism does not mean that sexism is causing anything relevant. It can be! But that has to be proven.
The next bit taks about moral values, etc. Here I am in agreement with her. Also, as a bonus thing, social conservatives are less intelligent on average, and there is limited evidence that progressives (d=0.31)and conservatives (d=0.46)are both less intelligent, on average, than libertarians (Haidt et al. 2012). I acknowledge that self selection into the test can be an issue. I welcome friendly critiques of this bit.
Again to the article:
I would like to highlight several paragraphs that I find extremely repugnant.
Demoralize diversity. As soon as we start to moralize an issue, we stop thinking about it in terms of costs and benefits, dismiss anyone that disagrees as immoral, and harshly punish those we see as villains to protect the “victims.”
By that logic, we should. It’s perfectly possible to have a civil, respectful discussion that includes morality; in fact, a mutual commitment to fairness and empathy usually makes for much more productive discussion.
I see Demore’s statement as a call for a cold and reasoned assessment of facts. Moralisation there means getting emotional and hysterical about issues. If he instead calls for a utilitarian morality, that is still does not escape moralisation: it is a system of ethics after all that focuses on a single locus of moralisation: happiness or preferences. I am not an utilitarian, but I see where Demore is coming from.
Another pernicious false dichotomy. Being emotionally disengaged in fact leads to. Whereas self-awareness — which involves understanding and acknowledging your emotions — does help you .
Now, see this quote from Demore
Philosophically, I don’t think we should do arbitrary social engineering of tech just to make it appealing to equal portions of both men and women. For each of these changes, we need principles reasons for why it helps Google; that is, we should be optimizing for Google—with Google’s diversity being a component of that. For example currently those trying to work extra hours or take extra stress will inevitably get ahead and if we try to change that too much, it may have disastrous consequences. Also, when considering the costs and benefits, we should keep in mind that Google’s funding is finite so its allocation is more zero-sum than is generally acknowledged.
What does Sadedin read here? Fascism
What these paragraphs together are advocating is. For all its mild tone, this is textbook fascism:
The core principle — what Paxton defined as fascism’s only definition of morality — is to make the nation stronger, more powerful, larger and more successful. Since fascists see national strength as the only thing that makes a nation “good,” fascists will use any means necessary to achieve that goal.
The guy is saying that Googlers should focus on making Google work well, not on diversity for the sake of diversity, and that is fascism!?
the Left tends to deny science concerning biological differences between people (e.g., IQ and sex differences).
The passing mention of IQ is interesting, since it has nothing to do with gender, which is the focus everywhere else. He’s presumably talking about race, but he doesn’t want to be branded a racist, so he keeps the reference subtle. So why risk doing it at all? It’s a dog-whistle to the alt-right.
Demore cites an article for that claim, this one that indeed touches on race and IQ. That is an interesting scientific question, albeit one that tends to elicit strong feelings in many people. In the past I have written a defense of a researcher that was accused of racism for investigating those issues. As this article of mine focuses on gender differences, I will pass on commenting more about this issue. The rest of the article is, to play with the author’s own words, “a poisonous fact-free violent mob rethoric routinely espoused by SJWs” and so here I end my review.
Once again, the arguments against the manifesto fail. These facts still stand:
- There are differences in key aspects of personality and cognition between and women across countries
- There are differences in preferences between men and women across countries
- Stereotype threat has not been proven to affect performance of women
- More equal countries, if anything, show greater differences
- The same gender distribution pattern is observed basically everywhere in the world
Now, this is the sort of evidence that backs a strong role for biology. What else do you want? If you say that this strong role is not biological, then what evidence could possibly convince you that biology (not only) is the main driver, especially in developed countries?
It cannot be western culture, as the pattern is also found in Asia.
It difficultly can be patriarchy, because one would expect that at least sexism had a negative effect on this findings, but it actually has a positive one.
It can difficultly be institutions: again, they happen under many countries around the world, each with their own.
It can difficultly be socialisation into a worldwide, equally affecting culture of patriarchy. There is a small literature that studies gender differences in toy use in babies, and they also find differences (samples sizes are small, to be honest) and there is even a small literature finding also gender differences in toy use in monkeys.
What else do you want? What alternative theory can explain all of this? And most importantly, what evidence is there for other explanations being stronger?