Sunday, April 25, 2010

Why IQ tests don't work

We've all heard people decrying the value of IQ tests: particularly that they are either culturally biased, or that they only exhibit one dimension of intelligence.

I suggest additional reasons for that value being limited.

I have always found IQ tests relatively easy - but I'm mathematically inclined.  Further, in my experience people that are clearly below an IQ of, say, 80, are clearly lacking some general capabilities.

My thinking is that  the upper end of IQ results reflect mathmatical/logical capability, but it doesn't measure the broad range of human capabilities.  Likewise, low scores are indicative of a disability.  Yet for those who score mid-range, it's hard to say anything useful about their intelligence.

There have been a number of alternatives suggested for the straight IQ measure, such as intelligence that is social, emotional, visual/artistic, musical, and so on.  I find myself in agreement that "IQ" measures only a limited range of a person's intellectual capabilities.

My suggestion is that those IQ measures that score mid-range are only demonstrating their mid-range logic capabilities, and that we have no sufficient measure of their capabilites in the broader aspects of human capabilities.

It was suggested to me that those with aptitude for classical music are likely to be pretty intelligent on typical IQ measures.  Yet my reading of the music industry more generally suggests that there are many musicians that are neither very logical in general, nor very capable of managing their own lives - even equalising for other factors such as self-medication.  Syd Barrett is typically held up for this measure in the music sphere; Van Gogh - and many others - are rightly or wrongly depicted as exemplary in the musical world.  I would be surprised if surrogate IQ tests didn't place them mid-range; however, I'm sure there are vast swathes of musicians that are highly intelligence in the IQ measure - it's just that such a measure is not directly relevant to their  particular expertise.

[What it has to do with brain function is an interesting question.  Recent findings have, for example, suggested that autism is much to do with a differential ability (or dis-)  of different regions of the brain to communicate with each other - and that high-functioning autism (so-called savant) may be an aspect of the same, that is, abnormality in the networking of different regions of the brain.]

Comments welcome.


bazza said...

Hello Stephen. I feel that IQ tests are an excellent measure of one's ability to take an IQ test.
I have always thought that Mensa is a pointless and self-congratulory organisation.
In society in general I am sure we would find that those who had a high IQ quotient would not exactly correlate with the 'successful' portion of the population - which ever definition of success one uses.

S Simmonds said...

Quite agree with you on all counts, Bazza.

I had idly pondered applying to Mensa in the past, but couldn't get past the speculation that it was full of people seeking refuge from bullying or some other form of under-appreciation.

That's probably quite unfair. But it is perfectly fair to question the relationship between IQ and success. The coupling, in general, must be somewhat loose.

Tom Eagerley said...

Hi Stephen. I had an invitation from Mensa once. They said I had acheived the impossible in terms of a low score and they invited me to never darken their door again. Rather harsh don't you think?

S Simmonds said...

Well Tom, look at it the Groucho Marx way: you wouldn't want to join any club that would have you as a member.

Bob said...

IQ is a crude measure of intelligence....but the best we have at the moment.

IQ is a poor measure of a person's potential for success. That being said a high score may give someone the confidence to aim higher.

S Simmonds said...


While maths/logic can figure high in some measures of success, it doesn't always trump.

There are plenty of people I admire for their other abilities, such as empathy (ability to work with others, and get people working together), their aesthetic sense, their groundedness, inventiveness, their solid work ethic. These and more are all largely independent of any maths/logic capability, yet they all contribute to humanity.

My contention is that with very poor skills on traditional 'IQ' measures, their are innate problems all around, yet high scores don't ipso facto translate to success in life or a chosen field.

Yes, high scores impart some confidence - but they only measure certain dimensions of ability. Other measures treated seriously would give other affirmations... yet the only thing we discuss is IQ measures.

Gretgor said...

Hey there. I completely agree with your point on IQ tests. They're a poor way of rating a person's capabilities for many, many reasons.

Another thing worth citing here is that, even though IQ tests were supposed to measure a very small portion of intelligence (that is, analytical reasoning and analytical reasoning only), they fail at that too, because the scores grow more and more biased according to how many said tests the person takes. The reason is simple: IQ tests tend to have the same puzzle patterns again and again, and said patterns repeat themselves in a way that people will most likely know naturally what to do if they went through many of them.

I had my IQ-obsessed stage myself, and I think it added nothing in the general sense of how good I am at what I study or how I study.

Since I study in a highly mathematical field, I see myself proving theorems in my mind very often, and it all narrows down to the fact that, even though I had high IQ scores, the ability to prove theorems only started to improve when I started practicing. If that means anything, is that IQ tests will never suffice in what concerns measuring a person's mental capabilities, since most mental activities depend on practice and effort to learn and master.

Still, I like to play around with mind puzzles, so maybe IQ tests would still be useful for "having a bit of fun" every now and then.