A while back, Chauncey Tinker wrote a post on Intelligence, Concentration, and IQ Tests:
I do not believe that IQ tests measure intelligence. Rather I believe that they measure a combination of intelligence, learning and concentration at a particular point in time. …
You may wish to read the whole thing there.
The short response is that I basically agree with the bit quoted, and I suspect that virtually everyone who takes IQ tests seriously does as well. We all know that if you come into an IQ test hungover, sick, and desperately needing to pee, you’ll do worse than if you’re well-rested, well-fed, and feeling fine.
That time I fell asleep during finals?
Not so good.
Folks who study IQ for a living, like the famous Flynn, believe that environmental effects like the elimination of leaded gasoline and general improvements in nutrition have raised average IQ scores over the past century or two. (Which I agree seems pretty likely.)
The ability to sit still and concentrate is especially variable in small children–little boys are especially notorious for preferring to run and play instead of sit at a desk and solve problems. And while real IQ tests (as opposed to the SAT) have been designed not to hinge on whether or not a student has learned a particular word or fact, the effects of environmental “enrichment” such as better schools or high-IQ adoptive parents do show up in children’s test scores–but fade away as children grow up.
There’s a very sensible reason for this. I am reminded here of an experiment I read about some years ago: infants (probably about one year old) were divided into two groups, and one group was taught how to climb the stairs. Six months later, the special-instruction group was still better at stair-climbing than the no-instruction group. But two years later, both groups of children were equally skilled at stair-climbing.
There is only so good anyone will ever get at stair-climbing, after all, and after two years of practice, everyone is about equally talented.
The sensible conclusion is that we should never evaluate an entire person based on just one IQ test result (especially in childhood.)
The mistake some people (not Chuancey Tinker) make is to jump from “IQ tests are not 100% reliable” to “IQ tests are meaningless.” Life is complicated, and people like to sort it into neat little packages. Friend or foe, right or wrong. And while single IQ test is insufficient to judge an entire person, the results of multiple IQ tests are fairly reliable–and if we aggregate our results over multiple people, we get even better results.
As with all data, more tests + more people => random incorrect data matters less.
I think the “IQ tests are meaningless” crowd is operating under the assumption that IQ scholars are actually dumb enough to blindly judge an entire person based on a single childhood test. (Dealing with this strawman becomes endlessly annoying.)
Like all data, the more the merrier:
So this complicated looking graph shows us the effects of different factors on IQ scores over time, using several different data sets (mostly twins studies.)
At 5 years old, “genetic” factors, (the diamond and thick lines) are less important than “shared environment.” Shared environment=parenting and teachers.
That is, at the age of 5, a pair of identical twins who were adopted by two different families will have IQ scores that look more like their adoptive parents’ IQ scores than their genetic relatives’ IQ scores. Like the babies taught to climb stairs before their peers, the kids whose parents have been working hard to teach them their ABCs score better than kids whose parents haven’t.
By the age of 7, however, this parenting effect has become less important than genetics. This means that those adopted kids are now starting to have IQ scores more similar to their biological relatives than to their adoptive relatives. Like the kids from the stair-climbing experiment, their scores are now more based on their genetic abilities (some kids have better balance and coordination, resulting in better stair-climbing) than on whatever their parents are doing with them.
By the age of 12, the effects of parenting drop to around 0. At this point, it’s all up to the kid.
Of course, adoption studies are not perfect–adoptive parents are not randomly selected and have to go through various hoops to prove that they will be decent parents, and so tend not to be the kinds of people who lock their children in closets or refuse to feed them. I am sure this kind of parenting does terrible things to IQ, but there is no ethical way to design a randomized study to test them. Thankfully, the % of children subject to such abysmal parenting is very low. Within the normal range of parenting practices, parenting doesn’t appear to have much (if any) effect on adult IQ.
The point of all this is that what I think Chauncey means by “learning,” that is, advantages some students have over others because they’ve learned a particular fact or method before the others do, does appear to have an effect on childhood IQ scores, but this effect fades with age.
I think Pumpkin Person is fond of saying that life is the ultimate IQ test.
While we can probably all attest to a friend who is “smart but lazy,” or smart but interested in a field that doesn’t pay very well, like art or parenting, the correlation between IQ and life outcomes (eg, money) are amazingly solid:
The correlation even holds internationally:
There’s a simple reason why this correlation holds despite lazy and non-money-oriented smart people: there are also lazy and non-money-oriented dumb people, and lazy smart people tend to make more money and make better long-term financial decisions than lazy dumb people.
Note that none of these graphs are the result of a single test. A single test would, indeed, be useless.
To be continued…