Isis and Sheril are having a discussion about some of Sheril's email. She recently got a question from a reader, who queries:
Hi Sheril,
I read with interest your article on 'women and science' in the Discover website. Can I ask, from your perspective, what you think of this study suggesting that men are smarter than women? Because there are a lot of men who agreed with the study, and even some women. The truth is, while I don't want to agree with it, I can't help but think that men are smarter than women. Or at least, made much advances in the field of science than women. Yes, women have been kept at a disadvantage for so long, but I wonder if men also push themselves more? Maybe they want it more? I really hate feeling this way, but deep down, I kind of believe that it may be true. Have you ever heard of Camile Paguila, btw? She's a 'feminist' but believes that, since most of the inventions we have in the modern world are created by men, if it were left up to women, we'd be living in grass huts.I know that the study is old, but if you could offer your perspective, that would be great.
Thank you.
So, the "article" referred to is an old Daily Mail piece reporting that
"It is research that is guaranteed to delight men - and infuriate the women in their lives. A controversial new study has claimed that men really are more intelligent than women.The study - carried out by a man - concluded that men's IQs are almost four points higher than women's.
British-born researcher John Philippe Rushton, who previously created a furore by suggesting intelligence is influenced by race, says the finding could explain why so few women make it to the top in the workplace."
Oh gawd. Again? I think I blogged about this article back in 2006. Why won't this psudoscience crap die?
Sheril, also tired of this lame argument that really should have been laid to rest by now, invites readers to weigh in, and Isis is gearing up for a response of her own. In the meantime, let me share a couple of brief thoughts. I will try to be civil.
I would start by point out to Sheril's reader that the IQ test was originally designed as a test for *children,* that something as complex as adult intelligence (and, actually, children's intelligence too) can clearly not be accurately nor precisely represented by a single number, that science that pruports to measure said intelligence with said number is bullshit, and to read The Mismeasure of Man by Stephen Jay Gould.
Then I would ask the reader (and the "researcher," actually), IF we somehow believed that IQ represented intelligence (which it doesn't, right? it's bullshit, already said that), how big the statistical effect size was of the study? I would then point them to Janet Hyde's paper on the Gender Similarities Hypothesis (pdf) which is so critical the APA doesn't have it hiding behind a firewall. Hyde argues that calculating effect sizes (i.e. how different the means are between two sets of data) is way more important than the fact that the means are different particularly because simply using very large data sets can bias the means to be statistically significantly different. She does a meta-analysis of many studies done over decades to point out that, for most of the studies, the effect size is nil, and for only a couple is the effect size not zero but they are still small. So research that looks at gender with big ns without looking at effect sizes is essentially crap science. Clearly the "researcher" of said study could benefit from some something we call "learning" from past science in the context of a good statistics course. It could have saved him all that work.
THEN I would tell the reader to stop reading such poorly done science reporting (let alone science) which plays on our socialized beliefs that "there must be BIG differences between women and men, right? After all, we're from different planets.." Instead, read this PhD comic. Or any of the other ones. They're all funny. And better written.
Finally, I would reassure the reader that, IF you could measure intelligence by IQ and IF the means were "really there" and not just an effect of a large sample size, and IF there was somehow a big effect size, that doesn't mean that a random individual man is smarter than a random individual woman. This is not an individual-level study -- it's a study about groups. So this doesn't say anything about individual relationships -- therefore one's dumb male inlaw/colleague/partner/etc. can still be dumber than you.
Not that you needed a study to know that.
- Log in to post comments
Consider a behavior example which in modern society works badly for males. There is no 'bad driving' gene, but there is, cross culturally and significantly a problem with young male drivers. Not intelligence, not reflexes, and not cultural (society spends a large effort to counter it), but an underlying (and temporary) lack of fear of personal injury.
By way of consolation, I remind you that there are lots of people who:
1. deny evolution
2. deny AGW
3. STILL believe in a link between Mr. Hussein and al-Qaeda
4. STILL think that Mr. Hussein had WMD
5. believe that vaccinations cause autism
6. think that 9/11 was an inside job
7. etc
8. etc
9. etc
I'm sure that I can think of many other examples. In any case, you shouldn't let the fact that the world is full of blockheads lead you to despair. I admit, gender bias will take a long, long time to go away. I had my head set right in 1973, when my new wife, who had wanted nothing more than to be a housewife and raise lots of kids, took a women's lib course and suddenly did a 180. That was fine with me -- I loved her, and if that's what she wanted, it was fine with me. Besides, gender equality just makes more sense. By the 1980s, I figured all this stuff was old hat, but I kept running into guys who simply didn't get it. At first I thought it was because maybe they were in odd little pockets of antiquity, but eventually I figured out that some guys just don't want to get it, They'll figure it out eventually, but we have to keep up the pressure. It will take a few generations. That's sad, I know. But we'll get there.
And then we can start working on guys from the Islamic cultures... that might take a while longer...
This is a great point. And even if the statistics were really convincing, gee, I can't imagine there could be any reason other than innate biology that a man might be more "motivated" than a woman or "want it more." It couldn't possibly have anything to do with social conditioning. That would just be silly!
Nicely put Alice!
Uffda. Rushton is the bane of all that is good and progressive in science. Journalists need to learn to treat him like bigfoot claims. Actually, bigfoot has a better chance of being real than even the most benign-seeming of Rushton's claims.
Yea, I'm going with bigfoot on this one.
The important thing to know about IQ is that it predicts economic and educational incomes better than any other variable in psychology. Whether you want to call it "intelligence", "cognitive ability", or "bullshit", that fact remains.
For readers interested in a balanced introduction to the subject, here is a good introduction to the IQ debate from the American Psychological Association.
As a counterpoint to the Hyde article, here's a recent study which came out arguing that the gaps are significantly larger than those referenced in the literature:
Current research practices lead to inadequate assessment of the overall psychological differences between males and females; relatively small univariate differences are taken at face value, without properly aggregating them into
multivariate ES indices. When differences are measured on multidimensional constructs, multivariate indices will almost invariably produce larger estimates of the statistical distance between the sexes.
Even assuming there weren't a multitude of problems with the study, such a small yet statistically significant difference does not translate into meaning anything in the real world. Who honestly thinks 4 points makes any difference at all? That's not even one standard deviation. Utterly meaningless, even if it were a real effect.
Field Notes,
4 points makes a big difference on the tails of the curves.
There are so many problems with the linked article that someone ought to make a bingo game out of it. Some additional points that leapt out at me:
- One study doesn't prove anything by itself, especially not if its results are not supported by previous studies.
- A 4 point difference in IQ doesn't mean anything unless we know more about the distribution. Is four points significant? What happens at the high and low ends of the sample?
- What's the correlation between IQ and job advancement? If the proposed gender difference were real, would this effect be large enough to explain the glass ceiling? Off hand I doubt that it would be, and no information has been given to make me think it was plausible.
- Never mind all the well-documented instances of gender discrimination in the workplace. I guess that in Rushton's world women earn less than men for doing the same work because their bosses have been secretly conducting IQ tests all these years.
- What's this... men outperformed women on a test of "verbal reasoning". Which test would that be? More information please. We already know that women and men perform differently on certain specific IQ-related tasks; cut out the tasks in which women outperform men and you've got a nice scientific-looking sample of men outperforming women.
- Most of the original computer programmers were women. I realize this has nothing to do with the article, but it's a piece of information that more people should know.
ben: you are correct in general, but the tails of the curves are not well behaved in this kind of test, so really, it is meaningless. The fact that this particular author is known to have fudged his data (regarding 'race') again and again, however, is probably even more important.
Recently stated on WNYC: it's not because a study shows some kind of gender disparity that you can predicted for one individual how they will score. I.e.: if a study 'concluded that men's IQs are almost four points higher than women's', it doesn't follow that because I'm a woman my IQ is automatically 4 points lower than the IQ of any random man I encounter.
Time for more education about causation and correlation!
It's totally bogus--but if I had to reply to the idiot who emailed her, I would say that if men are to be measured by their inventions and actions, given that women have been held back--then the state of the world today, with war, poverty, inequity, ecological disaster and all that entails--would show that men are stupid.
Ah, Rushton. Always trying to bring 21st century science back to the 1920s.
What an ass.
greg,
Rushton and Lynn fudge a good amount of data (probably both as a result of conscious bias and of sloppiness), but they also present some solid data and arguments. I'm skeptical of the idea that there is a mean difference in IQ between the genders, as it contradicts a pretty well supported consensus from earlier work.
Then again, it is possible, and if true very meaningful. A general ignorance on the subject of IQ is being shown on this post and in the comments. IQ is a test of cognitive abilities that predicts educational and economic outcomes better than any other variable. A critique of its use therefore requires a better argument than calling it "bullshit" and waving the Mismeasure of Man in the air.
Remembering that article that suggested that male scientists tend to make their big discoveries at the time of life when men are trying to attract mates generally...you know, that time when if women are going to have children, they have to do it then...
Even if there were a statistical difference (which is highly doubtful), it would mean little in individual circumstances.
If one were to look for any underlying difference, from an evolutionary perspective, it would be behavioral rather than intelligence. The question is what kind of behaviors would evolution select for.
First a critical point:
There is NO behavior which is exclusively male or exclusively female. Anything that exists as a general rule does not necessarily apply to any particular individual.
Consider a behavior example which in modern society works badly for males. There is no 'bad driving' gene, but there is, cross culturally and significantly a problem with young male drivers. Not intelligence, not reflexes, and not cultural (society spends a large effort to counter it), but an underlying (and temporary) lack of fear of personal injury. By and large, older males, and females of all ages don't go through this transition but young males do.
Consider that historically young males who successfully challenged the alphas became the most successful breeders for the next generation, and lack of personal fear during that critical time of life paid of genetically. Males that were afraid to challenge, or males that challenged and lost were out, those that challenged and won are our ancestors.
Aggression paid off less for our female ancestors. The death of a mother is far more detrimental to her offspring than death of a father (especially a father who as had multiple mates), so a more conservative strategy is selected for. Females are less likely to make some of the 'mistakes' that males do for that reason.
[side point regarding the male preponderance in computers: It's NOT that females don't have the intelligence, but a significant number of computer geeks got their start by spending obsessive numbers of hours under fluorescent lights playing/hacking video games and tearing apart the machines. This mind numbing obsessiveness is almost exclusively adolescent male behavior; you will rarely encounter a teenage female who would not quickly become bored with that lifestyle]
The original email to Sheril sounds like classic concern troll. "Oh, I don't necessarily believe this, but some other people do, and don't you think we should take it seriously?"
I think you are missing the more important factor. This isn't about just race or gender - it is about science done for popularity - Pinker & Chomsky being great examples to add to all of this. You get to be a famous scientist far more quickly if your research tells people what they want to believe. I have met many people familiar with Pinker and Chomsky, and Rushton that have to be reminded who Mendel is. This need for scientists to be popular above being a scientists seems very pervasive in many levels of science right now, (such as these examples above, as well as lab dynamics, constructive feedback, etc.)
ben: Yes, you are correct that a detailed and scientific rebuttal is the only appropriate ... rebuttal ... of a scientific finding. And that should and will happen and if PR happens to be right this one time, then ... well, I'll personally nominate him for the Nobel Prize.
I have read virtually every paper and page of print produced by him over the years and it simply is the case that there is so much chicanery and bullshit in his work that it is perfectly reasonable to write it off.
I would be shocked, by the way, if there was NOT a measurable IQ difference between men and women as populations, what with all the hormonal brain damage during puberty and through the early 20s. But as usual, Rushton has it backwards.
Sorry but reality is not going to conform to your wishes, and science is about finding out about reality and not about constructing fairy tales. It's obvious that the OP knows close to zilch about intelligence research, and just flays about blindly attempting to find some reason to criticize research whose results do not agree with her own prejudices.
In the first place, brain imaging research clearly shows differences in brain structure between males and females, and also differences in developmental trajectories between boys and girls, AND there may be differences in the relationship between brain structure and cognitive function between males and females. (I know, because I'm first author on some of those studies.)
Sorry, but THAT claim is bullshit. A single number in fact does very well represent human intelligence to a large degree, it's called g, which you find when you do factor analysis on all tests attempting to measure intelligence.
This is bullshit too. If there is a nonzero effect size, then there's a nonzero effect size, which means the means are in fact different, and it will be found statistically significant if the study has sufficient power. There is NO bias whatsoever.
No, it isn't. If statistically significant result is found then a statistically significant result is found. The fact that the effect size may be small does not mean the result is "crap science".
Actually you could use a good statistics course yourself. It's evident you don't even understand the basics.
Again results are not suspect merely because they were obtained from a study with a large sample size. There's no such thing as an "artifact" of a large sample size.
Do you really think that intelligence researchers are too stupid to know this?
Alice, that was me that wrote that stupid letter. I apologize for making myself appear as an ass to your readers. No, I'm not a concern troll..it's just that I hear and have read these comments enough; that it..never. mind.
Actually anonymous, I don't want to get involved in too much, because I am more interested in the responses, but aren't there many, many problems associated with the G factor?