◄►Bookmark◄❌►▲ ▼Toggle AllToCAdd to LibraryRemove from Library • BShow CommentNext New CommentNext New Reply
My impression of Damore’s Google Memo is that it is a thoughtful and well-considered personal opinion about workplace differences in abilities and attitudes. The tone is reserved, measured, and reasonable, avoiding sweeping claims. For example, it restricts its scope to the particular office in which he worked, and not Google as a whole. It is clearly written, with minimal jargon. Damore is at pains to explain himself, to clear up misunderstandings and to support his opinions with references. That does not mean that he gets everything right, nor is it intended to be a complete review of the literature of the sort one might find in a journal, but he makes a reasoned case for biological factors in sex differences. He makes helpful suggestions for Google to implement. I cannot comment directly on Google because I do not know about it first-hand, but I can look at the key assertions in this memo which can be tested by reference to psychological research. These have relevance for all workplaces.
I should make clear right away that I find most of his opinions perfectly reasonable and well supported by the research literature, though on some others I doubt their relevance or that they settle the issue. Of course, there is debate about all these matters, partly because there is debate about many important issues, but even more so because social science is often reluctant to consider biological causes for people’s abilities and attitudes. So, while on each of these topics there are references which favour either biological or cultural positions, there will be a majority which follow cultural interpretations. Damore is correct to say that there is a Left bias in social science, and unfortunately it has affected the interpretations placed on human behaviour, to the detriment of alternative biological explanations. Nonetheless, to associate a set of arguments with a particular point of view, Left or Right, is not to refute them. That must be done by a fair and balanced assessment of the evidence. In my view, there are many strong arguments to support the points Damore makes, though I can see that many people will not find them conclusive. Indeed, the research literature is as prone to culture wars as government regulated workplaces. Rival factions arrange their evidence in battle lines. As Buz Hunt used to say: “they are lawyerly rather than scholarly”. References from different perspectives rarely overlap, and often run in parallel, a case of perpetual confirmation bias. Even when a finding would seem to strengthen or weaken a particular position (I think that neonate visual preferences are in this category) the main flow of argument continues unabated. Are we swayed by evidence? Only sometimes, it would appear.
I find the ferocity of some of the replies to Damore extreme. The vehemence of the opposition is coruscating, and absolute. These issues should be matters of scholarly debate, in which the findings matter, and different interpretations contend against each other. Expressing different opinions should be a cue for debate, not outrage. We are far from having definite proofs about these matters, though personally I think we can see the direction of travel of the debate, which is that the case for genetics being a part cause of individual differences is gaining ground. It is only doing so because it can increasingly account for some of variance. A decade ago it was not possible to associate the genetic code with intelligent behaviour. Now studies which link snippets of code to intelligence are being published every few months. The pace of discovery is extraordinary. “Nature” and other science journals report frequently on new genetic correlations with important human behaviours, notably mental ability and mental illness and health generally.
Curiously, many people have reacted to Damore’s memo by assuming that cultural explanations for sexual and racial differences must be right, by definition. It is assumed that a culture-only explanation has triumphed because of the weight of the evidence. Damore’s view that both biological and cultural factors are involved in human differences seems to have been interpreted as him saying that only biological factors are involved. Damore has clearly argued for culture and biology being involved.
Worse, it has been assumed by some commentators that the consideration of biological explanations for sex differences is of itself reprehensible. As Jim Flynn said wryly to me of himself: “I know that I am on the side of the angels”. Whether they come from purported angels or devils, all hypotheses should be tested. As Jensen pointed out long ago, the Culture Only faction have to make a more fundamentalist case than the Heredity and Culture faction. The latter can concede at least 50% of the variance on a case by case basis, the former have to posit cultural explanations for all observations: a more demanding requirement.
The furore surrounding this memo seems to be based on sex differences, not racial differences, so I will concentrate on the former, though it has implications for the latter.
By way of background, it would be good to put the technology business into context, because many occupations, if not most, do not have a 50% representation of the sexes. An occupation is not an opinion poll: occupations represent competence, not opinions. In my view, there should be no requirement that a workplace be a mirror of society. Not every man wants to be a clinical psychologist. Personally, I have no objection to my technical computer assistance coming largely from Indian and Tamil men. I am interested in getting 5-star advice. As regards nurses and doctors, my main requirement is that they should be kind, and give me evidence-based and compassionate care.
Here are some issues raised in the memo on which I feel I can make some comments.
1 Differences in distributions of traits between men and women may in part explain why we don’t have 50% representation of women in tech and leadership.
Of course. The phrase “may in part explain” is general and not contentious. We need to look at the distribution of traits between and women to see if there are any such differences in traits, and if so, how big they are, and whether they “may in part explain” the representation of women in technology and leadership. We do not actually have to show what causes them, but it would help to show how easily they could be changed, and over what time span and with what amount of effort. For example, if women can be trained to be better at three dimensional tasks that is a good thing, but probably not of immediate comfort to a potential employer.
There are numerous sources, because the literature on sex differences is so extensive. I have made some suggestions in a previous post “Google Sex”. Here are a few I forgot to add to that list:
I would add that meta-analyses are not sacrosanct. They certainly provide larger sample sizes, but there can be idiosyncratic judgements about what is included and how the results are depicted. There will always be strategic differences between “lumpers” and “splitters”. Hyde (2014) is a case in point.
Hyde has done a very large scale meta-analysis, which is often quoted. Like most others, she uses Cohen’s d and rightly says that it assumes normal distributions and equal variances. This is a problem with sex differences, because males tend to have wider standard deviations than females. She considers this matter towards the end of her paper, and provides comparative statistics. In general, I think she tends toward lumping together. For example, she looks at many behaviours which are not considered likely to show sex differences. Of course, there is very much in common between the sexes. That should not be forgotten, but is not at issue. The questions are whether the sex differences which people notice are really there, how big they are, and whether they matter.
Here is a post on whether sex differences are underplayed or exaggerated:
On mathematical ability Hyde summarised the field thus:
Overall, then, it appears that girls have reached parity with boys in mathematics performance, at least in the United States. The gender difference in complex problem solving in high school is smaller than it was in the 1990 meta-analysis and has even disappeared in one analysis of National Assessment of Educational Progress data (Hyde et al. 2008a).
Of course, from an employment point of view we need to look at young men and women, either when they leave school at 17/18 or when they graduate at 21/23. For a highly desired job, such as Google, one should concentrate at the 2 sigma levels, that is, the top 2% or so. You will see from the research I list in Google Sex that men predominate at those levels.
On spatial ability, usually considered a male forte, Hyde argues thus:
According to an early meta-analysis, the gender difference in 3D mental rotation is large, favoring males, d=0.73 (Linn & Petersen 1985). In a later meta-analysis, the gender difference was moderate in magnitude, d=0.56 (Voyer et al. 1995). These overall effect sizes, however, mask some complexities. Gender researchers have suspected for some time that, for mathematical and spatial performance, tightly timed tests—which measure speed as much as skill—are advantageous to males, whereas untimed tests or tests with ample time provide more opportunity for females to display their skills. One meta-analysis found that, indeed, with short time limits, the gender difference in mental rotation was large (d=1.03), whereas in tests with no time limits the effect size was only moderate (d=0.51) (Voyer 2011; see also Maeda & Yoon 2013)
This argument makes me smile. Hyde seems to take as granted that males have an advantage on “tightly timed tests for mathematical and spatial tasks”. Is it simply my male point of view that to do well on any test, in the sense of getting things right, and doing so quickly, would be considered a double advantage? Why regard speedy thinking as a complexity of interpretation? Why is speed in correctly completing a task judged to be “speed as much as skill”? Absurdly, the prompt and correct completion of a task seems to be cast as mere male impetuosity. Furthermore, any employer reading this argument would be justified in thinking “On difficult tasks involving maths and spatial analysis, women need more time” so, given a chance, it might be better not to employ them.
Hyde mentions neonate preferences:
There is some evidence from individual studies that a male advantage in mental rotation emerges as early as infancy (Moore & Johnson 2008, Quinn & Liben 2008). However, it is a bit too early to tell exactly what these findings mean.
Well, in my view these findings mean that there is a male advantage, until someone does a study which shows otherwise.
Hyde does not mention one of the most striking results on neonate social preferences: Connellan, J., Baron-Cohen, S., Wheelwright, S., Batki, A., & Ahluwalia, J. (2000). Sex differences in human neonatal social perception. Infant Behavior and Development, 23, 113–118. Newborn boys are more likely to look at things than people.
This was also found at 12 months of age by Lutchmaya and Baron-Cohen 2002.
Publications like this strongly suggest that the “things versus people” dimension is an innate sex difference. Prof Simon Baron-Cohen confirmed to me today that he has not come across a replication study, and says it may be due to the requirement to test 100 neonates. Such a replication study might have called into question this male advantage. This is a very telling, because the neonate study currently stands as a way of showing an important sex difference which could not be due to cultural factors. You would think it would be a good target for an attempted refutation, yet no one has done so. As per usual, I offer a bottle of fine French wine to the first researcher to carry out this obvious test of innate sex preferences. As it stands, newborn males show a preference for things. (Not all of them. It is an average difference.) However, using a different method of choices on 48 infants, Escudero et al. (2013) do not find sex preferences in 4-5 month old infants. We should keep gathering findings for yet another meta-analysis.
I think that the debate about sex differences could be put into context by looking at sex ratios in intelligence under two assumptions: equal intelligence with a 1-point higher male variability; and a 4-point male advantage with the same 1-point higher male variability. The first assumption is generally held among intelligence researchers; the second is held by a minority, though with growing supportive evidence.
Assuming no sex differences in average intelligence, and a slightly narrower standard deviation for women, at IQ 130 there will be 59% men and at IQ 145 67% men. So, in a very intellectually demanding occupation, simply by appointing people according to ability, we might expect to find that two thirds are men.
On the minority view (Lynn and Irwing, 2004) that there is a 4-point male advantage in intelligence, such that it is reasonable to consider male intelligence at 102 and female intelligence at 98, then at IQ 130 there will be 74% men and at IQ 145 84% men.
If you have not already seen it, you will find relevant studies here:
By the way, I am still not sure that the 4-point male advantage is proved, in the sense of being shown in current population studies, so I am tentative about it, but the brain scanning work comes up with that figure, and it will be interesting to see if that male advantage is retained as the sample size increases from n=900 to the recently released n=1200.
I do not know what intelligence levels are required at Google, nor how levels differ between different roles in that organization, so these figures are for illustration only. However, I would assume that they can recruit very bright people, as many of them as they need, so the upper levels may be the most appropriate for the key roles in that organization.
On the general matter of sex differences in preferences it is useful to look some summaries
Richard Lippa (2010) Gender Differences in Personality and Interests: When, Where, and Why? Lippa concludes:
Results show that gender differences in Big Five personality traits are ‘small’ to ‘moderate,’ with the largest differences occurring for agreeableness and neuroticism (respective ds = 0.40 and 0.34; women higher than men). In contrast, gender differences on the people–things dimension of interests are ‘very large’ (d = 1.18 ), with women more people-oriented and less thing-oriented than men.
Gender differences in personality tend to be larger in gender-egalitarian societies than in gender-inegalitarian societies, a finding that contradicts social role theory but is consistent with evolutionary, attributional, and social comparison theories. In contrast, gender differences in interests appear to be consistent across cultures and over time, a finding that suggests possible biologic influences.
In the Sternberg and Kaufman Handbook of Intelligence Halpern et al. Chapter 13 say “on average, women and men live systematically different lives”. At more length (page 263) they go through the evidence, saying:
There are substantial differences in the values, attitudes and interests of contemporary males and females, which may help to explain cognitive sex differences.
They tend to a social and cultural explanation of these differences, saying (page 266) that biology cannot explain “the vast improvement of female performance on certain measures such as the increasing numbers of females at the highest end on the SAT Math test (Blackburn, 2004).”
You should balance this view against the more recent publications described in “Maths is a man thing*”
By the way, I am late in posting about this, but in conference proceedings this May, Guy Madison of the University of Umea, Sweden, email@example.com described an interesting technique to detect bias against women in academia. Sweden is of particular relevance because it has done so much to try to reduce sex differences.
Sex differences among higher academics in Sweden
ABSTRACT: There are significantly fewer women than men among the highest academic ranks in most countries. Many different explanations have been proposed, from sex discrimination to interests and priorities and cognitive ability in the high range. Any of these explanations could be consistent with the widely varying sex distribution among professors across disciplines (10-30% in the technical and natural sciences and 40-95 percent in the Humanities, Social sciences, and in Medicine and Veterinary science). In Sweden, several measures have been taken to increase the overall proportion of women to at least 40 percent (Utbildningsdepartementet, 1994, p. 37), but after more than a decade of these policies, women constitute only 24 percent of professors across all disciplines (Sandström & Wold, 2016). Here, we test the hypothesis that this difference is due to discrimination, according to which women should have higher merits than men at the point in their career when they are appointed to the position of professor. Specifically, preferential hiring of men at the expense of equally or more qualified women should be reflected in the latter having published more scientific papers and having had greater scientific impact in terms of more citations, and publishing in journals with higher impact factors.
From the total population of 1,345 professors appointed in 2009 to 2014 at the five largest universities in Sweden, we drew two random samples of about 100 persons from each sex, and compiled all their publications from the Web of Science. Differences across disciplines were relatively small within the social sciences, with a mean of 4.5 publications and 30 citations, but professors in medicine had a mean of 30 publications and 500 citations. Contrary to the hypothesis, male professors had about 80% more publications and 40% more citations than female in the social sciences, and in medicine the males had about 60% more publications and 200% more citations. There were no significant sex differences in impact factor.
2 Political orientation is actually a result of deep moral preferences and thus biases. Considering that the overwhelming majority of the social sciences, media, and Google lean left, we should critically examine these prejudices:
There are two issues here:
A) Are political orientations based on deep moral preferences?
B) Do the overwhelming majority of the social sciences, media and Google lean left?
On A) I am unsure of whether political orientation is based on moral preferences, but it is certainly based on deep differences in outlook, and the table of biases (preferences) seems a reasonable summary. Here are some posts on the link between intelligence and political attitudes.
On B) Social Sciences in the US lean heavily to the Left, perhaps because they are brighter.
The media also seem to lean Left.
I do not know about political orientations within Google.
However, Left wing orientation does not of itself prove or disprove the strength of the arguments. It may show that much of the literature considers environmental and social variables to the detriment of genetic ones, but once again we have to consider the strength of the arguments, and not just count the numbers of publications.
3) At Google, we’re regularly told that implicit (unconscious) and explicit biases are holding women back in tech and leadership. Of course, men and women experience bias, tech, and the workplace differently and we should be cognizant of this, but it’s far from the whole story.
On average, men and women biologically differ in many ways. These differences aren’t just socially constructed because:
• They’re universal across human cultures
• They often have clear biological causes and links to prenatal testosterone
• Biological males that were castrated at birth and raised as females often still identify and act like males
• The underlying traits are highly heritable
• They’re exactly what we would predict from an evolutionary psychology perspective
Note, I’m not saying that all men differ from women in the following ways or that these differences are “just.” 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.
Comment: implicit unconscious biases. That has certainly been widely asserted, but more recent work suggests that the implicit association method is not a reliable predictor of real life attitudes.
Universal across cultures. Largely so, though not entirely. Biological explanations are not invalidated by exceptions, but biological factors seem to cover the main findings pretty well, in that there is a general picture of sex related preferences across virtually all social groups. However, exceptions to the rule are instructive, and there can be great social changes for cultural reasons such as religious prohibitions and legal changes, so both biology and culture can be influences on sex differences.
Clear biological/testosterone. Well, there is certainly a strong case that males and females differ biologically, and it is very probably the case that hormones are involved, but the mechanism is not crucial. What is crucial is how well one can establish sex differences, and whether they can be linked to biological causes in general.
Biological males castrated at birth. I don’t know, and cannot find good papers on this.
Heritable. Yes, most human behaviours are heritable, and that is true of mental ability. If anyone is really denying that, they are wrong. If, however, anyone argues that such behaviours are heritable it means that they are not subject to environmental influences, then I think that they would be wrong. Damore does not make that mistake in his memo. Here are Turkheimer’s (2000) three laws of genetics:
“First Law: All human behavioural traits are heritable.
Second Law: The effect of being raised in the same family is smaller than the effect of the genes.
Third Law: A substantial portion of the variation in complex human behavioural traits is not accounted for by the effects of genes or families.”
Evolutionary psychology. Although it is true that evolutionary psychology certainly deals with these issues, and has much supportive evidence, it still needs to be demonstrated what that supportive evidence is. The best approach is to read the literature. On the question of sex differences, journals like Intelligence, and Personality and Individual Differences are good sources of research findings. My links only cover a few papers. For a text book summary look at the evolution of intelligence written in 2011 see Gaborra and Russon’s evolutionary history of intelligence, which is Chapter 17 in Sternberg and Kaufman “The Cambridge Handbook of Intelligence” 2011.
The Note makes it very clear that men and women “differ in part due to biological causes”, that many such differences are small, with significant overlaps, and that you cannot say anything about an individual on the basis of population level distributions.
The last point is not strictly true. In fact, you can say some things about individuals given population level distributions of male and female traits if there is a sizeable difference in mean, and/or standard deviations. For example, you can estimate that 90% of men have more upper body strength than 90% of women. Many of these estimates would be error-prone, but would still be better than chance. The bigger the mean differences the better the predictive power. In my view it would be a good guess that a man selected at random would be better at visualizing the rotation of three dimensional shapes than a woman selected at random. Interestingly, if people have been properly selected by Google, none of these observations matter. All Google employees will be good at algorithms, and will have the necessary intellectual power to carry out the correct, accurate and fair classification of the world’s knowledge. Making allowances for their sex and race would be redundant.
My main reaction to the Damore memo is that much research can be found to back up his claims, even though the culture-only narrative is predominant in social research. There is much contention on many of the points he makes. I think he is well supported, but not all the findings run his way, and that is the general rule if you are evidence-based. Faced with contrary results you try to give a summary, and plot out the main trends, but there will always be contrary findings.
Others have given their opinions about these matters, and as I have said, I am struck by the intolerance of many of Damore’s critics. Their negative comments, often extremely vehement and dismissive, have an absolute quality, usually claiming that he is totally wrong and that his views have no merit whatsoever. No one can really pretend to be neutral on all matters of opinion, though they do not have to have an opinion on everything, but finding out which version of reality contains the fewest errors should be a cooperative procedure. I know that is hard to achieve, but I think it worth trying.
I would have been both briefer and more comprehensive in my coverage, but I wanted to comment whilst this topic is still a matter of public debate. As always, I am open to further reading (and have several tabs open on specific studies so as to be able to add more later).
For those who are curious and open-minded there is much to read and discuss, and I hope some of the references shown here will encourage you to look into the topic yourself.