[DaniFong]

The problem may be that our predictors are terrible. There are a lot of Nobel prize winners who later discover their 'dismal' performance on IQ tests. I know grad students at Princeton who've confided sub-100 IQ scores, bad SAT's, horrible performance on one math test or another -- it's like this big, shameful secret for otherwise brilliant people.

Everyone carries around the absurd burdens of judgments and measurements, but they don't always mean what people think they mean.

[byrneseyeview]

Are there any other quantifiable variables with such high predictive power? I'd be really interested if there were some other test that could better estimate the odds that someone would get a degree, earn lots of money, stay out of jail, vote often, delay having kids, etc.

It would be interesting to know more details, which I'm sure you can't divulge without violating someone's privacy. But it would be neat to find out if those people had other skills that correlate strongly with IQ, like the various digit-memorization/recitation tests, or reaction time. Were these people autistic? Were they taking a test in a language they didn't know too well?

[DaniFong]

They weren't autistic, and there didn't seem to be any strange behaviors socially, and they were quite inquisitive and swift on the uptake.

I once heard IQ defined as "ability to navigate bureaucracy, getting the answers others think correct in a manner testwriters imagined, and color in the lines" (paraphrased). This may have something to do with the phenomenon here.

The odds that someone would 'get a degree, earn lots of money, stay out of jail, vote often, delay having kids, etc.' seem to have more to do with successfully conforming to certain values of society's upper-middle class.

I think I'm capable of inventive thought, but I don't particularly want to get another degree, or earn much money, or delay having kids, or vote often, and the sort of things that one has to do to stay out of jail, are, honestly, quite often absurd, and I often rail against them.

And if you start to observe, closely, just how these things are tested, you'll start to get the impression that maybe bright people will find their ways through the cracks more than anticipated. There's an enormous weight given to quick answers -- time directly influences scoring. Linear answers are expected, and alternative interpretations are docked. There's often insufficient information in the questions, or it's based on a model of the world that's wrong. Domain knowledge like mathematics or vocabulary is brought into it. Analogies are made to hone in on one relationship from the many that could exist. Scorers of essays give insufficent weight to substance and too much to form -- despite their lipservice, they are indeed swayed by big words.

Even tests like GRE physics are bad. They claim to be testing rapid physical intuition, but in practice what divides good from poor scores is prior experience with 100 simple systems and the ability to get the factors of 2 and pi right with three minutes per question.

[sethg]

I will never forget one of the questions from the IQ tests that I took in fifth grade.

The tester asked me, "Who discovered America?"

Ah, I thought, a trick question. "The Indians!" I said.

Pause.

"Who is generally credited with discovering America?"

[byrneseyeview]

I think your quote comes from Cosma Shalizi. Thomas Sowell addressed that complaint years before by noting that most of the countries where you find that modern complexity are pretty nice places to live, whereas other places are hellholes. The correlation between IQ and income holds true for places outside of the US, too, so one might be tempted to think that a bureaucratic, find-objective-answers, fill-out-these-forms-in-triplicate society is better than most of the brutal alternatives.

"...seem to have more to do with successfully conforming to certain values of society's upper-middle class. I don't particularly want to get another degree, or earn much money, or delay having kids, or vote often, and the sort of things that one has to do to stay out of jail, are, honestly, quite often absurd, and I often rail against them."

My point is that, all else being equal, you could get a great job if you chose to, and that if you do illegal things, you apparently do them in such a way that you won't get caught. IQ seems to correlate with the ability to delay gratification, which itself seems to be a better predictor of success than IQ (sadly, it hasn't been tested in a rigorous, long-term way -- so it can't match the hundred or so years worth of data people have compiled on IQs).

One of the reasons that these tests emphasize time is that 'quickness' is a component of IQ. Francis Galton, the first guy to really study the subject, liked to think of things in that way -- and given that physical reaction time correlates so well with IQ, he had a point.

I agree that the essay part of standardized tests is messed up. Lots of the recent changes to tests seem to arise from political correctness. Test-makers found out that you can't design a test that has predictive value without getting politically incorrect results, like lots of men on the extremes, or lots of high-IQ Jews and Asians and low-IQ Mexicans and blacks. So they periodically adjust the scoring mechanism or the test to get bell curves closer to the same median and standard deviation (more focus on the median than the SD, since most of the people who complain about such things don't know what 'standard deviation' means). So keep that in mind when complaining about the essay, or the rebalanced scores, or the analogies (which were dropped from the SAT -- analogies happen to be more IQ-weighted than other categories of questions).

[DaniFong]

For the most part, life in latin america seems a lot happier than in more bureaucratically encumbered states. So I don't really buy that argument.

My main point is that truly excellent thought doesn't depend on the same skills that would allow scoring highly in an IQ test. In many cases such skills would inhibit it. In an IQ test it helps to rapidly adapt to the assumed constraints of the problem and come quickly to the closest, most linear answer. If you have a mind compelled to bring questions back to reality, to challenge assumptions or think of things from many different angles, you'll do, on the whole, more poorly than if you had not. But these behaviors are sometimes precisely what you'd want!

[byrneseyeview]

For the most part, life in latin america seems a lot happier than in more bureaucratically encumbered states. So I don't really buy that argument.

The millions of immigrants from there to here don't seem to buy your argument. Which Latin American countries do you mean?

"... truly excellent thought doesn't depend on the same skills that would allow scoring highly in an IQ test."

I happen to agree with you! One thing that's bugged me for years is that you can't design a test that distinguishes, in advance, between cleverness and original stupidity. So that means that whole lot of time-consuming activities can only be measured after the fact if at all. The difference in our views, I think, is that I would argue that IQ tests measure the kind of raw data-processing skills that are useful in any situation (the military tests people heavily, and for all tasks they've found that IQ correlates positively with results -- I think for some technician jobs, it explains about 60% of the variance in individual performance). There just aren't any studies I know of showing that people with low scores on IQ tests go on to succeed in any measurable way. It would be convenient, to say the least, if you argued that the success-deficit among low IQ people is more than compensated for by a success-surplus that happens to be impossible to measure. So I'd like to know if you can find some way to quantify your argument. It would change my thinking on a lot of subjects if I found that doing poorly on an IQ test predicted doing well at some other task.

[DaniFong]

In happiness surveys, Latin America does pretty well -- I think as a region it comes out on top.

This is contentious, but for most of the immigrants, they didn't move because the culture was more enjoyable here. Most of the Latin American immigrants had ther livelihood strip from them in two stages: one, from general industrialization pressures that have affected practically every country, forcing specialization into cash crops to compete, two, increasing competition from the USA and other countries, coupled with crop failures endemic to semi-arid areas.

The few areas in which modestly educated Latin Americans could reliably compete were in illegal cash crops (for which they had less competition in the USA), which continues to cause economic distruption and criminal activity, and manual labor. Those forced into either business might have better chances in the USA, but on the whole they're happy countries. When you're there, you can feel it.

As for success and low IQs, if I recall correctly, someone got the bright idea of testing Caltech professors while Feynman was there -- this may have been prompted after he won the nobel prize, and found out his score from highschool was 125. When they tested the professors, the scores were surprisingly low. I forget who it was, but someone got to make a big deal of his 105 score, since it was three points higher than Feynman's.

My point is that these measures can be stunningly irrelevant, and, if so, while they might have some utility for, say, selecting an undergraduate class, they are often best ignored by an individual when it comes time to decide what one is capable of.