Sexual pseudoscience from CNN

« previous post | next post »

The old oppositions between girls and boys — sugar and spice vs. snips and snails — continue to be reinforced and extended in the popular press by sexual pseudoscience. For example, Leonard Sax's false claims about boys' inferior hearing are front and center again in Paula Spencer, "Is it harder to raise boys or girls?", CNN.com/health, 6/17/2008:

Why don't boys seem to listen? Turns out their hearing is not as good as girls' right from birth, and this difference only gets greater as kids get older. Girls' hearing is more sensitive in the frequency range critical to speech discrimination, and the verbal centers in their brains develop more quickly. That means a girl is likely to respond better to discipline strategies such as praise or warnings like "Don't do that" or "Use your words."

I won't go over the scientific issues about hearing in detail, but you can get the details and the citations from these earlier posts:

"Leonard Sax on hearing", 8/22/2006
"Girls and boys and classroom noise", 9/9/2006
"Liberman on Sax on Liberman on Sax on hearing", 5/15/2008

Let me summarize the evidence in one simple sentence: There is no functionally significant difference between boys and girls in auditory sensitivity.

As for the business about "verbal centers in their brains", if we focus on vocabulary, it's true that girls are on average quite far ahead at 18 months. But they are less far ahead by 24 months — and they are actually behind, on average, during the age range 6-10 years, according to several studies cited in the last of the posts just listed.

And it's crucial to keep in mind that at every age, individual differences among boys and among girls are large compared to the average differences between the groups. (Sex differences in verbal development are discussed at greater length in "The main job of the girl brain", 9/2/2006.)

From Paula Spencer's CNN article again:

From birth, a girl baby tends to be more interested in looking at colors and textures, like those on the human face, while a boy baby is drawn more to movement, like a whirling mobile, says Dr. Sax. (These differences play out in the way kids draw: Girls tend to use a rainbow of hues to draw nouns, while boys lean toward blue, black, and silver for their more verblike pictures of vehicles crashing and wars.)

In a nutshell, girls are rigged to be people-oriented, boys to be action-oriented. Because girls study faces so intently, they're better at reading nonverbal signals, such as expression and tone of voice. Boys not only learn to talk later than girls and use more limited vocabularies, they also have more trouble connecting feelings with words.

For the business about connecting feelings with words, see "David Brooks, cognitive neuroscientist", 6/12/2006; and "Are men emotional children?", 6/24/2006.

The assertion that "Girls tend to use a rainbow of hues to draw nouns, while boys lean toward blue, black, and silver for their more verblike pictures" comes from pp. 21-22 of Leonard Sax's book Why Gender Matters (trade paperback edition, 2006), where he discusses the implications of his claim that male and female retinas are different in major ways:

Suppose you give crayons and a blank sheet of paper to young girls and young boys. Let them draw whatever they like. You'll find that girls will prefer colors like red, orange, green, and beige, because these are the colors that P cells are prewired to be most sensitive to. Boys prefer to simulate motion in their pictures. Boys prefer colors such as black, gray, silver, and blue because that's the way M cells are wired.

For a discussion of the  retinal physiology that Dr. Sax has cited in support of these claims, see "Of rats and (wo)men", 8/19/2006, and "Retinal sex and sexual rhetoric", 5/20/2008. In a nutshell: Leonard Sax's claims about large human sex differences in the distribution of retinal cell types are not supported by the papers that he cites, and are apparently false.

I haven't previously discussed Dr. Sax's claims about childhood sex differences in the ability to read nonverbal signals, and infant sex differences in interest in mobiles and the like. On pp. 18-19 of Why Gender Matters (trade paperback edition, 2006), he writes:

Most girls and women interpret facial expressions better than most boys and men can.20 Researchers at Cambridge University wondered whether female superiority in understanding facial expression was innate or whether it developed as a result of social factors … These researchers decided to study newborn babies on the day they were born.

Their plan was to give babies a choice between looking at a simple dangling mobine or at a young woman's face […] The boy babies were much more interested in the mobile than in the young woman's face. The girl babies were more likely to look at the face. […] The researchers concluded that they had proven "beyond reasonable doubt" that sex differences in social interest "are, in part, biological in origin."21

His citations for these claims are:

20: Judith Hall, Nonverbal Sex Differences, 1985; Erin McClure, "A Meta-Analytic Review of Sex Differences in Facial Expression Processing and Their Development in Infants, Children, and Adolescents", Psychological Bulletin 126:424-53, 2000.

21: Jennifer Connellan, Simon Baron-Cohen, et al., "Sex Differences in Human Neonatal Social Perception", Infant Behavior & Development, 23:113-18, 2000.

Let's examine Erin McClure's meta-analysis of facial expression processing, and leave Connellan et al. on infant mobile-watching for another morning.

Executive summary of McClure: From dozens of studies, she obtained an overall estimated effect size for tests of "FEP" (Facial Expression Processing) of d=0.13, meaning that the difference between group averages was about 13% of the within-group standard deviation. As she explains, this "suggests that on a typical measure of FEP, 53% of girls should perform above average, compared with only 46% of boys". A "reliability-corrected" estimate of the effect size was d=0.16, "indicating that 54% of girls should perform above average and 46% of boys should do so".

(For some discussion of what an "effect size" is, and what distributions for effect sizes in the 0.13-0.20 range look like, see "Gabby guys: the effect size".)

In any case, these small effect sizes are typical of the literature on sex differences in perceptual and cognitive abilities. I'll comment only that such effect sizes are smaller than those often found for socio-economic and ethnic divisions. Thus Martha J. Farah, et al., ("Childhood poverty: Specific associations with neurocognitive development", Brain Research 1110(1) 166-174, 2006), found an effect size of about 0.95 for differences on vocabulary and sentence-understanding tests between between middle SES and low SES African-American girls aged 10-13. (And with respect to Dr. Sax's ideas about retinal physiology, Huynh et al. 2006 found that the difference in average retinal thickness between white and asian schoolchildren was twice as great, in absolute magnitude as well as in effect size, as the difference between boys and girls — see the tables at the end of "Retinal sex and sexual rhetoric"  for the details.)

There are a lot of interesting facts and ideas in McClure's paper — especially the enormous range of effect sizes actually found in the studies that she surveyed, ranging from -0.59 (a fairly large advantage for boys) to 0.92. But I'll leave it there for now, and close with a brief semantic interlude: did Leonard Sax tell the truth when he wrote "Most girls and women interpret facial expressions better than most boys and men can"?

This obviously depends on what most means.

The American Heritage Dictionary says that most means "In the greatest number of instances: Most fish have fins." This is not a lot of help, as it just refers us to the problem of what "in the greatest number of instances" means.

Merriam-Webster says says that most means "the majority of <most people>". This provides a clear answer: if "most" just means "greater than 50%", then Sax's assertion is validated by McClure's paper, which tells us that on a given FEP test, we should expect about 53% of the girls to be above average, compared to about 47% of the boys.

But I wonder if this answer is the right one. It seems to me that people often take most to mean a large majority, large enough to be in some sense characteristic of the set quantified over, and not just barely more than half. And this helps explain why Paula Spencer, the author of the CNN article, took Sax to be making a much more categorical statement. Her version uses a plain plural to characterize the group difference: "Because girls study faces so intently, they're better at reading nonverbal signals, such as expression and tone of voice."

Indeed, in Why Gender Matters, Dr. Sax also usually dispenses with quantifiers of any kind, and reverts to plurals — for example:

p. 17: "… newborn baby girls really do hear better than newborn baby boys."

p. 19: "… girls are born prewired to be interested in faces while boys are prewired to be more interested in moving objects. The reason for that difference has to do with sex differences in the anatomy of the eye."

p. 23: "Girls draw nouns, boys draw verbs."

p. 25: "Let's look at another difference in how girls' and boys' brains work: geometry and navigation. Researchers have found that females and males use fundamentally different strategies for those tasks."

There's some experimental evidence that most people mostly interpret most in a way that lends itself to an easy transition to the generic plural. One piece of the puzzle: Tim Hunter, Justin Halberda, Jeff Lidz & Paul Pietroski, "Beyond Truth Conditions: The semantics of 'most'", SALT 18. They examined people's responses to statements like "Most of the dots are yellow", for displays in which the proportions of yellow and blue dots ranged between 1:1 and 2:1. They concluded that people interpret most in terms of a comparison of cardinalities mediated by the "Approximate Number System" (ANS), as discussed in Lisa Feigenson, Stanislas Dehaene and Elizabeth Spelke, "Core systems of number", Trends in Cognitive Sciences, 8(7): 307-314, 2004. Thus "participants' success rate … decreased as the ratio of the number of yellow dots to the number of nonyellow dots approached 1, closely matching the psychophysical function independently identified for the ANS".

According to Feigenson et al., there are  two core cognitive systems dealing with numbers: an approximate representation of numerical magnitude, and a precise representation of distinct individuals. Perhaps linguistic equivocation between these systems — as well as the inadequacy of either system to express even simple propositions about statistical distributions — helps to explain the general tendency to derive propositions about generic group characteristics from propositions about differences between group averages, even when these difference are small relative to within-group variation.

For some examples of this tendency in action, see e.g. last fall's posts on the gender happiness gap. [Update: or this one.]



26 Comments

  1. Mark P said,

    June 19, 2008 @ 9:28 am

    If the reporter had wanted to do real reporting, it would have required at least as much work as you put into your blog post this morning, and it's not like she gets paid for that … oh, wait, never mind.

  2. Greg said,

    June 19, 2008 @ 9:45 am

    It would help if reporters had any grounding in the sciences. I doubt that even reading the original research would have led the reporter to understand that Sax's work was incorrect.

  3. Craig Russell said,

    June 19, 2008 @ 11:00 am

    Hey, kids, get out the crayons. Let's draw some nouns!

    (Sorry, that phrase sounded funny to me. What was everyone else talking about?)

  4. Jean-Sébastien Girard said,

    June 19, 2008 @ 11:34 am

    In case it's any use, the Connellan et al paper is available here

  5. Virgil Ikari said,

    June 19, 2008 @ 12:08 pm

    It's insane to me that garbage like this gets continually reprinted, that people's sexist prejudices consistently get reinforced by what are apparently lies and half-truths that authors like Brizendine and Sax peddle for profit.

  6. Liz Ditz said,

    June 19, 2008 @ 12:56 pm

    At least Spencer didn't trot out the crockus.

  7. john riemann soong said,

    June 19, 2008 @ 1:20 pm

    Not to expect much from pseudoscience, but has Sax thought of any evolutionary pressures for why sexual dimorphism would occur in language?

  8. Mark Liberman said,

    June 19, 2008 @ 1:39 pm

    JR Soong: has Sax thought of any evolutionary pressures for why sexual dimorphism would occur in language?

    I haven't seen speculations of that sort from Sax, but Louann Brizendine suggested that the female brain is "[a] machine … built for connection. That's the main job of the girl brain, and that's what it drives a female to do from birth. This is the result of millennia of genetic and evolutionary hardwiring… [S]ince you're smaller, you probably need to band with other females to fend off attacks from a ticked-off caveman…"

    (See "The main job of the girl brain" for context and discussion.)

    But if there were actually large sex differences in verbal abilities, I'm confident that evolutionary psychologists could think of reasons for them.

    In fact, humans have substantial sexual dimorphism in the vocal apparatus that develops as a secondary sexual characteristic after puberty — about a 60% difference in vocal cord length, and about 15% difference in vocal tract length, compared to about an 8% difference in other length dimensions. I believe that this is unique among the apes — gorillas are much more sexually dimorphic in general, while chimps are more like humans, but neither has any particular sexual dimorphism in the vocal organs, as far as I have been able to determine.

    The human innovation is presumably related to sexual selection pressures associated with vocal displays; and the development of spoken language along the hominin line is probably not a coincidence.

  9. dr pepper said,

    June 19, 2008 @ 2:10 pm

    > Hey, kids, get out the crayons. Let's draw some nouns!

    The girls all do so. But the boys, with their inferior hearing, start drawing guns instead.

  10. John Laviolette said,

    June 19, 2008 @ 3:06 pm

    Boy, that CNN article is all kinds of fail.

    Not only did the reporter not do research to verify Sax's claims, but she didn't even get his claims right. "Girls tend to use a rainbow of hues to draw nouns, while boys lean toward blue, black, and silver for their more verblike pictures of vehicles crashing and wars." That seems to suggest that girls use a greater variety of colors than boys. But the Sax quote talks about color preferences, not range of color. How did that creep in?

  11. Jodi said,

    June 19, 2008 @ 4:39 pm

    Dr Pepper:

    Guns ARE nouns.

  12. john riemann soong said,

    June 19, 2008 @ 4:45 pm

    I worry about how many unwitting people have been converted to Sax's side because of that article. Can one take any legal action against news outlets for publishing pseudoscience?

    ""p. 23: "Girls draw nouns, boys draw verbs.""

    I didn't see this claim before. It's funny because I've seen some linguistic-determinists claim that Western culture perceive the world in the form of objects, whereas Native American cultures perceive the world in the form of processes. Or something like that; so apparently the hypermasculine Western culture is really feminine after all.

  13. HP said,

    June 19, 2008 @ 7:54 pm

    Okay, I can't claim any kind of scientific or statistical expertise here, but it seems to me [now there's a phrase that signifies!] that what all these gender studies do is a) pick some kind of metric to which they can assign an index, b) determine index values for a (hopefully random) sampling of men and women, c) plot normal distributions of the results, d) make a huge deal of the fuzzy fringe where the two normal distribution curves don't quite line up, and e) ignore the huge area that represents the intersection of the two curves.

    Even if the fuzz along the edges is statistically significant ("standard deviation," "margin of error," — I said I don't know statistics), the majority of both samples are still going to lie in the intersection of the two curves. When are we going to see a headline that says, "Men, Women, Mostly Alike, Experts Say"?

    Am I way off base in this assessment?

    My question is, why bother with gender differences? Why not pick something like blue jeans vs. khakis? Or fans of classic rock versus fans of album rock? You could probably wind up making the same kind of claims. ("Blue Jeans Wearers Use Fewer Adverbs, Experts Say.")

  14. John Atkinson said,

    June 19, 2008 @ 11:33 pm

    john riemann soong:

    " "Girls draw nouns, boys draw verbs." I didn't see this claim before. It's funny because I've seen some linguistic-determinists claim that Western culture perceive the world in the form of objects, whereas Native American cultures perceive the world in the form of processes. Or something like that;"

    Well, as far as the Native American vs Western claim is concerned, if you calculate the noun/verb ratio in typical hunks of text in, say, Mohawk and English, I bet you'll find the difference is _highly_ significant. Whether or not this actually provides evidence for the linguist-determinists' claims, it does raise the question: What's the noun/verb ratio of boy talk as opposed to girl talk?

  15. john riemann soong said,

    June 20, 2008 @ 12:43 am

    Does having a polysynthetic language for a native language, imply that you see the world in the form of processes?

    Wouldn't cognitive differences in language between the sexes slowly break a language into two along lines of sex? Yet today we cannot say, "Oh, this text is in the girl register, and this text is is the boy register." After all, if there were any significance to the cognitive differences, we should expect differences not only between noun and verb usage, but in fundamental grammatical structure, Wug test performance and diglossia of some sort. And apparently Sax appears to be drawing on English data only.

    But Sax doesn't appear to be *fascinated* with why such cognitive differences would exist. After all, truly significant differences in verb/noun use must reflect some more fundamental cause, enough for him to try to collaborate with researchers of other fields he's cherry-picking from (like say, a linguist). But apparently all he wants to do is push his single-sex school agenda, not actually discover the truth. The difference between a pseudoscientist and a scientist (as it appears to me), is that a scientist is always cautious about drawing premature conclusions from his findings; he is always open to better data. But Sax doesn't appear enthusiastic about the peer-review process. After all, if you were a genuine researcher who found genuine results that show real cognitive differences between the sexes, the pertinent and proper recommendation to make in your paper would be along the lines of, "These are interesting results — further research should investigate the underlying cause and effectiveness of gender-differentiated instruction of varying degrees." (As opposed to, "My findings show that everyone should be put in single-sex schools! NOW!!")

  16. Josh Millard said,

    June 20, 2008 @ 12:59 am

    The implication here is that Eskimos have a couple hundred words for snowing?

  17. möngke said,

    June 20, 2008 @ 5:42 am

    First comment from a dedicated reader (and future anthropologist):

    I am repeatedly saddened by the fact that what I essentially consider sexist drivel is considered by many to be 'hard scientific fact' of a sort. With regard to the drawing distinction: in my experience, at least (considering my own childhood drawings), it does not make any sense at all.

    What genuinely puzzles me is why proponents of gender differences don't concentrate on the [b]actual[/b] differences between girls and boys – in, for example, contrasting the gender percentages of (dedicated) players in MMORPGs, for example. Of course, this (as with all computer games, in fact) might be because males, in general, have more Free Time than females, their exemption from household chores being more or less institutionalised – however, it might also highlight more fundamental cognitive differences.

    In any case, I harbour a strong suspicion that most (biological) males are 'atypical' (social) males, if we take general gender stereotypes (under which the hearing opposition seems to be subsumed) as defining (social) males (or Men, as Shelly Errington would have it). If anything, the Sax case only seems to confirm this.

  18. Chris C. said,

    June 20, 2008 @ 11:22 am

    Guns don't kill people. Nouns kill people.

  19. M. Dalen said,

    June 20, 2008 @ 12:37 pm

    möngke : "What genuinely puzzles me is why proponents of gender differences don't concentrate on the [b]actual[/b] differences between girls and boys – in, for example, contrasting the gender percentages of (dedicated) players in MMORPGs, for example. Of course, this (as with all computer games, in fact) might be because males, in general, have more Free Time than females, their exemption from household chores being more or less institutionalised – however, it might also highlight more fundamental cognitive differences."

    In MMORPGs, I *suspect* that that is more of a cultural thing – computer usage started among a small group of antisocial young males, and as it expanded, the large majority of males kept women from getting in. I've seen the same thing in areas like non-computerized role playing games: they were originally targeted to young men, and as the genre matured, the community just wasn't very welcoming to women – who wants to hang around antisocial, often overweight men with poor social skills and hygeine?. (Much as I hate to say it, many of the "gamer" stereotypes do have a tendency to be true, at least in the dedicated minority that makes up the public face of the hobby.) It's getting *much* better recently, as women make inroads into gaming and RPGs and we get away from the scantily-clad, busty women. But there's still a heavy male majority in most areas.

  20. Ann said,

    June 20, 2008 @ 4:47 pm

    Josh Millard wins the thread!

  21. Paul Wilkins said,

    June 20, 2008 @ 5:11 pm

    So I read the article as a fluff piece by a columnist – a pseudo-journalist. As is pointed out, her assertions are not backed up. There's a reason – this was not that type of piece.

    If you haven't noticed, the CNN front page has been full of this type of stuff for a while now. It keeps us there poking around and looking at ads instead of working or surfing Parents.com or Wiki-how-to's article on proper Terrorist Fist Jabbing.

  22. TruePath said,

    June 22, 2008 @ 9:28 pm

    Thank you very much for publishing a thoughtful and balanced repudiation of this crap. The repeated attempts in the media to make statements suggesting dramatic innate gender differences is appalling, especially when it's just horrible old stereotypes repackaged.

    However, I'm equally appalled by some of the emotionally/ideologically driven counter-attacks/rebutals that try to pretend there is no evidence of any innate gender differences. Not only is this simply factually untrue but it turns every scientific study showing tiny innate differences into an argument for gender stereotypes and biased treatment. Thank you for offering a reasoned response based on the fact that studies don't show any differences large enough to be relevant to daily interactions.

  23. michael said,

    June 24, 2008 @ 11:57 am

    I agree with True Path. On the one hand, people keep trying to turn very small statistical differences into big gender differences. On the other hand, there are major differences, some of which are apparently so obvious that no one can be clever by citing them. A central one is that males are really physically stronger. In culture after culture, this superior physical strength is parleyed into male dominance in the most straightforward ways, always justified by some sort of rationalization (e.g., god created women as inferior, as in the Koran). You may be certain that this particular difference – which may or not be accompanied by innate average differences in temperament – is a difference "large enough to be relevant to daily interactions" throughout the third world – and elsewhere – in assymetries in beating, rape, and public social power. First worlder educated people, of course, find this sort of crude obvious difference not worthy of discussion.

  24. TruePath said,

    June 24, 2008 @ 11:25 pm

    To clarify my point I meant there were no innate mental/intellectual differences large enough to be relevant to daily interactions. Differences in physical strength are large enough to be relevant in daily interactions but only barely. Moreover, much of the reason gender differences in strength are relevant to daily interactions is because we no longer gather much data about the strength of our friends and associates in normal interaction. Hence we often don't have information about who can bench press what that would supercede gender based assumptions to use when we ask someone to open a jar for us or help us lift something.

    The point I was trying to make is that because any potential innate mental gender difference will be small compared to the variation inside the gender it doesn't make sense to use these kinds of studies to make deciscions in our daily lives. Even if it turns out that women are slightly worse at task X but better at task Y while vice versa for men it would take a very large effect to justify considering gender instead of merely looking at more direct indicators of ability/performance (tests, grades etc..).

    In fact as I have argued at length on my blog statistical facts about the general population tell you nothing (and even can be misleading) about the normal case we encounter of having partial information about a person's ability.

    For instance just suppose that there really was some innate factor that made women less interested in math/science professions than men. Now one might think that this would mean that a female applicant to grad school in these subjects would be less qualified than a man but it could easily be the opposite is true. Since fewer women do pursue these professions and peer pressure is a significant factor in what people choose to do it likely would take a greater level of interest as a woman to decide to apply for math/science grad school than as a man. Thus even though a random sample from the population might show women to be worse than men at math/science in this scenario once you have conditioned on applying to grad school the effect of gender might be reversed.

    Anyway that was of course entierly hypothetical but I just wanted to emphasize how silly it would be to directly apply the conclusions about small innate effects to real life deciscions. Of course most people on both sides of this debate don't realize this which makes it so personal and icky.

  25. Older said,

    June 25, 2008 @ 12:13 pm

    I'm sorry, Michael. The evidence is that there's pretty much the same situation with regard to size and strength as with regard to other variables sorted by gender. That is, there's a large overlap in the middle of the distribution and long tails where one sex is represented more than the other. Our "empirical" experience tends to disguise this fact with regard to size/strength, because nearly everyone thinks that women should choose men who are larger than themselves and vice-versa, so couples almost always present this situation. But there are large couples, and small couples, and there are plenty of women who are larger and/or stronger than "most" men. It's not as clear-cut as you suggest.

  26. I hate them said,

    June 24, 2010 @ 3:03 am

    "…But there are large couples, and small couples, and there are plenty of women who are larger and/or stronger than "most" men. It's not as clear-cut as you suggest."

    Ultimately, like everything else. This form of sexual science isn't about proving anything as much as it is a platform to degrade others based upon their gender. Haven't you noticed how its always men and boys who are referred to as 'inferior' and 'disadvantaged'? its always about how men are biologically, physically, sexually and intellectually inferior to women. It's always the same old story. In 1940 science and medical science were used to justify the inferiority of jews and other groups. Indeed it was 'proven' that such people were 'lacking' in many fundamental ways. There is an economic market and culture that revolves around male degradation. Research science these days is hugely influenced by money. Creating lies, half truths and other claims to satisfy the ego's and beliefs of feminists and or other ill minded people makes money. Hitler funded research that satisfied his claims and beliefs as well. Its all very simple y'know. However, the down side is that this form of sick science is taught to boys and is widely embraced by universities and schools. Yet, no one understands why boys feel so low these days and why there is a suicide rate of young men greater than the the casualties of some major wars. But then again, they are just boys/men and who ever cares about them? We are the inferior sex right? With our inferior sexuality when compared to women.

RSS feed for comments on this post