Friday, March 15, 2013

What is IQ?

Sandeep Gautam wrote in “Why IQ is a Myth”

To me, to be frank a score of 162 or whatever on a test means nothing, and I hardly care if the test is Cattels , WISC or stanford-binet. When a lay person sees a score of 100 or 160 he assumes that a) intelligence can be fully measured and quantified and b) IQ is that measure.”

It is very true that IQ scores divorced from their context are pretty meaningless. It's like saying, “Forty-two is The Answer to the Ultimate Question of Life, the Universe, and Everything” [Douglas Adams, The Hitchhiker's Guide to the Galaxy] Forty-two what? What was the question?

We can list a lot of things that an IQ score isn't:

It isn't The Answer to the Ultimate Question about a person.
It isn't an immutable number carved on your brain.
It isn't able to measure the full range of a person's abilities.
It isn't a measure of someone's worth.
It isn't a map of someone's destiny.
It isn't a limit (or an expectation!) placed on a person's potential.

But instead of underestimating the “lay person”, perhaps we should attempt to explain what an IQ score is. It is information about how a person performed certain tasks, under certain conditions, on a particular day. A full-scale IQ score makes sense only when we understand it in its full context, including: the type of test, its standard deviation, its ceilings, its error ranges, the other composite scores, and the subtest scores. A good tester also makes behavioral observations about how the person approaches problems and emotional factors that may have affected the results. To interpret the subtest scores, we need to understand the specific cognitive abilities being drawn upon by those tasks. A full-scale score is only the broadest summary of someone's performance ranked as a comparison to the scores of others in the same age group. It gives none of the details about strengths or weaknesses.

A typical full-scale IQ score (such as given by the
Wechsler tests), is called a standard score. This score corresponds to a percentile ranking of the sum of the scaled scores of selected individual subtests, which are themselves rankings of the raw scores for each task normed for each age group. It is essentially a measure of “unusualness”. The more unusual subtest scores someone has, the more that will be reflected in the sum, resulting in a more unusual full-scale score. The logic of this process can be confounded in situations where a person's scores in one area are unusually low and unusually high in another, giving the appearance of an average sum. This is why we need to look at the other composite scores (for example, Verbal or Working Memory) and the subtest scores in order to interpret the full-scale score.

An unusually high score, like Neha Ramu's, is possible only when all or nearly all of the subtest scores are unusually high. What does this tell us about a person? It means that in comparison to others her age, she excels in skills that are predictive of academic success, which is what most IQ tests have been designed to measure.

In its various incarnations, IQ has been linked to certain personal characteristics and learning outcomes—I'm not going to summarize those volumes of research here. The important thing to remember when reading about such research is that statistical correlations can tell you only about the likelihood of a given outcome. They cannot be applied to individuals. We cannot know ahead of time whether a person is part of the majority who will experience a given outcome or the minority who will not. The fact that there is such a minority does not invalidate the research, just as the fact that IQ scores may be misunderstood by some does not mean they are useless or absurd.

No psychologist trained to administer IQ tests would claim that intelligence can be fully quantified by IQ. I completely agree that it is sloppy journalism to report Neha's score as evidence that she is smarter than Albert Einstein, and equally sloppy to compare her score to an estimate (Einstein never took the test Neha took). But instead of dismissing IQ tests as nonsense, let's set the record straight.

No comments:

Post a Comment