IQ tests are not like tape measures

Intelligence and IQ are never far away from the news, and debates about whether IQ is ‘real’ never seem to get resolved. Some people think that intelligence is nothing but a cultural construct. It’s almost clichéd to say that IQ is just what IQ tests measure. It’s certainly true that IQ tests have often been used as little more than scientific veneer for political oppression. On the other hand, the science is clear. Intelligence is genetically influenced, linked to specific measures in the brain, and IQ scores relate to a huge variety of important life outcomes, including life expectancy itself. These are some of the most robust findings in all psychology.

Yet discomfort remains. Exactly what IQ ‘is’ remains frustratingly ambiguous.

We usually talk about intelligence as if it were like, say, height; as if it were an objectively measurable part of our biology. But is it really? Height is objectively measurable, with units of measurement that are tied to the structure of the physical universe. (A metre is the length of the path travelled by light, in a vacuum, in 1/299,792,458 seconds.) IQ tests aren’t like this.

Let’s take this comparison further. What is true of height is also true of some other measurements you could take of the human body, like, say, mass, or the circumference of the waist. And if you were to collect the data and do the analysis, you would find that these measurements correlate with one another. The taller someone is, the heavier they are likely to be, and the bigger will be the circumference of their waist. The measurements aren’t perfectly aligned, of course — some people are tall and light, others are short and heavy — but there is nevertheless a high degree of correlation.

To explore a bit further, you could then do statistical tests to identify major patterns in the data. Intelligence researchers often do this. They look at performance across a range of tasks, and they find that a single component predicts a large amount of the variation. They call this component ‘g’, for ‘general intelligence’. If you did the same for height, mass, circumference, and other ways of measuring the body, you would, I’m sure, also find a single statistical component that predicts a great deal of the variation. Let’s call this component ‘size’.

You could then, if you wished, do science about size. You could study how size relates to life outcomes, or the extent to which size is inherited from one generation to the next. And what you would likely find out about size is many of the things we have found out about intelligence: that there is, for instance, a relationship between size and various life outcomes like income, education, and life expectancy, and also that size is strongly genetically influenced.

Now, here’s the thing. Show me someone’s ‘size’. Point to it. What is your size?

You can’t do this! You can’t directly measure size, because we don’t really have a single size, not strictly speaking. You have a height and a mass. Size, however, is only a statistical abstraction of these measurements. It’s not — not strictly — a physical thing that can be measured on a single scale.

That doesn’t mean however that size isn’t a coherent scientific idea. On the contrary, it’s a completely sensible topic of enquiry, and you can study it using all the normal tools of science. But it’s not the same sort of thing as height. In the jargon of philosophy, height and size have different ‘ontologies’. One is an objective measurement, tied to the physical structure of the universe; the other is a statistical abstraction of such measurements. They are different types of thing.

Let’s now go back to intelligence. What sort of thing is it? Is it like height and mass, or is it more like size? My strong bet is that it’s the latter. It’s descriptive and statistical. It’s not something directly measurable itself. Unfortunately both the scientists and the skeptics talk about intelligence the other way, as if it is like height. That’s why they keep talking past each other.

Bottom line. Intelligence is real, just as much as size is — but that doesn’t mean that IQ tests are like tape measures. Some people are ‘big’, and some people are ‘intelligent’, but both these words are vague and loose, and they lose meaning without context.

So the truly interesting questions to ask about intelligence aren’t about whether it’s ‘real’ or not. Of course it’s real, just like size is real, and saying that shouldn’t be controversial. But saying that it’s real doesn’t mean we know exactly what it consists of.

To advance our understanding further, we need more discussion and more research about the underlying factors. What are the natural, physically measurable factors that underpin intelligence, in the way that height, circumference and mass all underpin size? We don’t have answers to that question at the moment, but if we could get some we should be able to dilute some of the discomfort that exists around the idea of intelligence, despite scientific reality.