August 26, 2005

Female IQ findings: O vast incorrectness!

"The Times" of London is a highly recommendable paper for its straight-down-the middle accounts of most things (unlike a certain New York publication). The fact that it is a Rupert Murdoch publication (think Fox News) probably has something to do with that. But I think even "The Times" was a bit courageous to print the following story (excerpt):

"A study claims that the cleverest people are much more likely to be men than women. Men are more intelligent than women by about five IQ points on average, making them better suited for "tasks of high complexity", according to the authors of a paper due to be published in the British Journal of Psychology. Genetic differences in intelligence between the sexes helped to explain why many more men than women won Nobel Prizes or became chess grandmasters, the study by Paul Irwing and Professor Richard Lynn concluded. They showed that men outnumbered women in increasing numbers as intelligence levels rise. There were twice as many with IQ scores of 125, a level typical for people with first-class degrees. When scores rose to 155, a level associated with genius, there were 5.5 men for every woman. Dr Irwing, a senior lecturer in organisational psychology at Manchester University, said that he was uncomfortable with the findings. But he added that the evidence was clear despite the insistence of many academics that there were "no meaningful sex differences" in levels of intelligence".

Source

Before all my female readers delete their bookmarks to this site, let me explain. I am a psychometrician by trade so I do know a little about this. First, let me point out that British Journal of Psychology is Britain's top academic psychology journal. So it needs to be reckoned with. What it reports, however, has in fact been known to psychometricians for about 100 years. And that is that men and women have the same intelligence ON AVERAGE but the scatter of intelligence differs between the sexes. Female intelligence clusters much more closely around the average -- so there are fewer very dumb women and fewer very bright women. And that, I am afraid, is how the cookie crumbles. The geniuses tend to be men but so do the dummos. And I am sure most women have met plenty of the latter.

Some correspondence about the above

Let me start with one well-expressed email that I received:

"As a woman, I don't have a problem with the IQ findings. I tend to believe it, as my personal experience has shown that men tend to be more analytical than women. My issue is with the way IQ is measured. My opinion is that IQ tests place a lot of emphasis on analytical abilities, but not much on "other types" of intelligence, such as creativity, multi-tasking, musical genius or whatever. For example, women tend to be a lot more perceptive than men, especially when it comes to relationships. They also can have more agility of mind; that is, they can do more than one thing at the same time, and do it well (better known as multi-tasking). Everyone has his or her own strengths, which leads me to believe that IQ tests are mostly useless. Instead, instructors/employers should be trained to identify individuals' strengths and how to capitalize on them.

Most of what the lady says is right. There are ways in which women tend to do better than men -- and multi-tasking is certainly one of those ways. What the lady does not know is that the abilities measured in IQ tests are NOT just some arbitrary selection of puzzles. The whole notion of IQ arose from an OBSERVATION: the observation that people who tend to be good at solving one sort of puzzle also tend to be good at solving lots of other seemingly unrelated puzzles. In other words, what Binet discovered in the 19th century was that problem-solving is GENERAL. There is such a thing as general problem-solving ability (often abbreviated as 'g'). So over a hundred years have gone by since Binet's discovery and most people still don't know of it! So IQ tests are simply collections of different puzzles that do in fact go together. Success on one does tend to predict success on all the rest.

And what that means is that IQ tests are VERY useful. For instance, if you are hiring for a job that requires a lot of problem-solving, you can use an IQ test to predict which applicant will be best at that job -- no matter what the problems may be in the job you are hiring for. And IQ tests are also very predictive of educational success. If you have a high IQ it is much more worthwhile to spend up big on a university education than if you have a low IQ.

As an example of how ability generalizes, take mechanical aptitude: I am very good at all sorts of academic things so lots of people would think I must be hopeless at practical things like mechanics. And it is true that any time my car needs fixing I hand the job over to an expert. But I like fixing locks. I am an amateur locksmith. Locks are just another puzzle to me. So one day, I was at a small gathering where some ladies were having trouble with the deadlock on their front door. So they took it off and opened it up. And immediately, bits and pieces went "SPROING" everywhere. They were of course completely stumped by that and did not for a moment think to ask a hopeless academic like me to help. So I said: "Maybe I can help". They looked at me with great skepticism. But in ten minutes I had it back together and all working properly. I hope they learnt something about 'g' from that episode.

Now I have just used an example above to illustrate what I am saying. But the example is NOT the proof. The proof is the gazillion times researchers have found that problem-solving generalizes. One of my other readers of my post yesterday made that mistake. She said that men got all the Nobel prizes because good education has become available to women only fairly recently. But that was not the point at all. The researchers who wrote the article in The British Journal of Psychology that I referred to yesterday relied for their conclusions on hundreds of studies with IQ tests. The bit about Nobel prize-winners was only an illustration, much like my locksmithing illustration above. Examples prove nothing by themselves. They just help you to understand how generalizations work out in practice.

Incidentally, creativity is NOT like IQ. It does not generalize much. People who are highly creative in one field are usually pretty uncreative in other fields. For instance, I am extremely good at writing articles for scientific journals. And that is a highly creative field. In that field you are creating new knowledge and understanding about something. And I have had hundreds of such articles published. But I could not write a novel for nuts! So even in the single field of writing, there can be different types of totally unrelated creativity!

TECHNICAL NOTE: The fact that there are few females in the top range of IQ can be explained in two ways -- in terms of a narrower distribution of IQ among females or as reflecting a lower average IQ among females. The first explanation is the usual one and the one I have adopted above. Irwing and Lynn reject that explanation and adopt the second explanation. They say that females are dumber all-round. That explanation is normally avoided and most tests are in fact constructed to show equal male and female averages

So Irwing and Lyn are saying that that is artificial and if the test is constructed solely of the highest loading items, a higher male score will emerge. That may be so but it depends a lot on the initial item pool used. So the point is rather moot. How the test constructors selected their initial item pool becomes the issue.

Since Irwing and Lynn used university students as their subjects, the generalizability of their findings is also unknown



Go to John Ray's Main academic menu
Go to Menu of longer writings
Go to John Ray's basic home page
Go to John Ray's pictorial Home Page
Go to Selected pictures from John Ray's blogs