Why "Smart" People Are "Stupid"

[drwex]

Today's xkcd (http://xkcd.com/c242.html) is funny, but it also illustrates a truism: smart people are more likely than less-smart (in the IQ sense) to make the same mistake twice or even three times.

It's a combination of curiousity asking questions that normal people really don't care about (is that a random thing? is it repeatable?) with the arrogance of the smart (that couldn't possibly happen to me... again).

Contextually, this is part of the discussion about why conventional (IQ-like) measures of intelligence are outmoded or just flat-out wrong. It's part of the science of why intelligence and emotion are really inseparable (and thus why Meyers-Briggs is a load of horsepuckey) and makes me want to get back into reading that body of literature.

Comments

[chienne-folle.livejournal.com]

What a lot of people don't know is that the original purpose of intelligence testing has been turned on its head.

Waaay back when the first one was made, it was because Alfred Binet was assigned the task of figuring out which kids were underperforming. Teachers noticed that some kids weren't doing very well in school, but when extra time was spent with these kids, some kids benefitted greatly and others didn't. The original purpose of IQ tests was to try to find a measure of intellectual potential that was free of socioeconomic bias, so that kids who were smart but poor could be located and helped to live up to their full potential.

I'm sure the poor guy is spinning in his grave at what they've done to his test.

I was taught to give IQ tests while I was in grad school, and the teacher stressed that the final number was *not* the most important part of the test. Most IQ tests consist of a dozen subtests, and the idea behind testing is supposed to be that the pattern of differences among the scores on the various subtests helps teachers figure out where a kid needs extra help or what a kid excells at. IQ tests are what helped educators figure out that many children who can't read are of normal intelligence and spawned our understanding of what dyslexia is and how to treat it, as opposed to dumping dyslexic kids on the garbage heap.

So although they call it "intelligence," what the IQ test is supposed to measure is the ability to perform well in a conventional school setting. That's all. Nothing else. And it does that reasonably well.

It's true that most conventional IQ tests don't measure a heap of important things, but they were only ever designed to measure the much narrower domain of what-is-taught-in-schools. The tests have been perverted beyond their original purpose by people who don't understand them, as have most personality tests. The MMPI, for example, is supposed to help identify which form of psychopathology a person has when it is known that s/he has something -- it's NOT supposed to be used on random job applicants.

The people who make these tests could almost certainly make up tests of emotional intelligence or suitability for various jobs, but the tests currently in use are not those tests. And the tests currently in use are *not* supposed to be used by anybody other than trained psychologists, who know about the limitations of the tests. But greedy test-printing companies will sell them to anybody with the funds to buy them, even though they claimed that they wouldn't do this when the tests were turned over to them.

It's not the fault of the tests or their makers that they've been perverted to uses they were never designed for by people who are lazy, unscrupulous, stupid, or ignorant.

[chienne-folle.livejournal.com]

P. S. --

The current misuse of psychological tests is kinda like using the CD drive of your computer for a cup holder and then complaining that it isn't strong enough to hold a ceramic mug. And we'd all say, "It was never designed to do that! It's not *supposed* to do that. Stop using your CD drive as a cup holder, you moron!"

Well, psychological tests were never designed to be used in the way that stupid people are currently using them. The next time you encounter a psych test being used badly, just think "cup holder." :-)

[drwex]

While I agree with nearly everything you say in the longer comment (see below for one question) I disagree about the cup holder thing. I use the "cd drive as cup holder" example in my HCI class as a way of helping students think through the eyes of computer users. The behavior is perfectly reasonable once you understand its context.

Likewise, you can't really blame people or call them stupid if you hand them something called a ruler and they then go around trying to measure things with it. People who understand will know that you don't measure edge molding with a seamstress's tape, but you need at least some level of specialized knowledge to understand why that tape and a tape measure differ. Likewise, I *do* blame the people who promote M-B because I think they're a vile form of snake oil salesmen who ought to know better.

Now to jump back a comment: I haven't used MMPI since undergrad days but I thought it was useful as a general measure for tracking a person's mood, outlook, and approaches to life over time. I don't recall it being a specific diagnostic tool for pathologies. Or am I confusing it with something else?

[chienne-folle.livejournal.com]

I haven't used MMPI since undergrad days but I thought it was useful as a general measure for tracking a person's mood, outlook, and approaches to life over time. I don't recall it being a specific diagnostic tool for pathologies. Or am I confusing it with something else?

I think you're confusing it with something else. The MMPI -- Minnesota Multi-phasic Personality Inventory -- was the first well-known test that was constructed empirically, rather than being theory-driven. The researchers took a bunch of people who were supposed to be of ordinary mental health and also a bunch of people who were hospitalized mental patients, whose diagnoses were known. The researchers gave these subjects an enormous number of questions. Any question that reliably distinguished between ordinary folks and mental patients made the cut; the rest of the questions were thrown out. Then these remaining questions were examined to see if they could reliably distinguish a mental patient with one sort of diagnosis from a patient with a different diagnosis. Presto: A bunch of different scales, for the different diagnoses they looked at. So a person who takes the resulting test gets scores for such things as tendency towards hysteria, tendency towards schizophrenia, tendency towards psychopathy, and so forth. This test has absolutely no business being used with any sort of normal population -- none.

You mentioned the Myers-Briggs earlier, whose abbreviation is the MBTI (Myers-Briggs Type Indicator) -- is that what you thought I meant by MMPI?

you can't really blame people or call them stupid if you hand them something called a ruler and they then go around trying to measure things with it.

Quite true. But most of these tests are only supposed to be sold to licensed professionals. So nobody *handed* these folks a ruler; it's more like they snuck in and stole something they thought was a ruler, only it's actually something else. They keep trying to measure stuff with it, not realizing that it's calibrated for a completely different measurement.

One of my favorite psychological testing stories involves the army, many decades ago now. They were trying to figure out who should be picked for missions to extreme climates, like the arctic or the equator. Not that they were all that concerned about the comfort of their personnel; it was more that soldiers who did not do well in these climates tended to fail at the missions.

So they hired a bunch of psychologists, and the army and its psychometricians tried a whole bunch of different things to try to figure out who just could not stand the arctic and who would be okay there. They measured body temperature in cold conditions and return of the body to normal temperature after extended exposure and a whole bunch of other things. After extensive testing they found the best way: Ask the soldier. Yep, it turns out that if you ask people whether they like hot or cold weather better, they can tell you. And if you ask them how much extremes of their preferred weather bother them, they can tell you. Duh, just ask. That NEVER occurred to the army! :-)

Cory, who knows she will do MUCH better in the arctic than at the equator