> IQ doesn't picture everything about a human's cababilities
Straw man, no one claims that IQ measures everything, only that it's important.
> In that regard, they are very useful, but to try to apply them
> to the upper range of scores is sort of meaningless.
From
http://infoproc.blogspot.com/2010/09/more-on-psychometrics.html -- "everyone thinks that test validity drops off for scores higher than their own IQ"
> Yes creativity is key, and is not measurable by an IQ test,
> that's a given.
Creativity is measurable, and it does correlate with IQ. Dec 2010: "Contemporary creativity research views intelligence and creativity as essentially unrelated abilities, and many studies have found only modest correlations between them. The present research, based on improved approaches to creativity assessment and latent variable modeling, proposes that fluid and executive cognition is in fact central to creative thought" --
www.citeulike.org/article/8465858
And re: "that's a given" -- what does that even mean?
> I mean it's anecdotal but let's face it so are IQ tests
The Feynman anomaly is anecdotal (and has been widely discussed on the www for years), but IQ tests as a whole are anything but anecdotal, as they are the result of hundreds of millions of individual test scores. Unless you consider that anecdotal.
> nowadays, with the work of people like paul torrance on creative
> intelligence, there are also other types of tests
Sure, lots of tests devised by Goleman, Sternberg, Gardner and others measure e.g. "emotional intelligence" or "intrapersonal intelligence", which is fine, except for the fact that the predictive value of these tests is basically zero. "Conventional" intelligence tests, on the other hand, are moderately to highly correlated with things like income, school/job performance, health, atheism, etc., and have been for decades.
> IQ tests are not like math tests with tasks and answers that are
> well defined
Except that many components of IQ test *are* like math tests, and *do* have answers that are well defined. A random one from the www: "John bought three books for five dollars each, and paid ten percent sales tax. How much did he pay all together?" -- I would say $16.50 is the correct answer, but you might go out of your way to devise alternative answers -- maybe John is actually a diplomat who is entitled to tax-free shopping in certain zones of the country where he's stationed, and thus pays $15.00 -- does that answer make you smarter than the test-makers? Or does it only indicate that you like to be seen as contrarian?
> scores are "normalized" to make a certain number of people
> come out with a certain score, as there is no intrinsic menaing to
> getting 25 out of 26 word analogies "correct" on a test.
Of course, that's their whole point, to rank people (however rank that sounds). Also, there is a lot of work being done to develop bio/neuro-based IQ tests with a cardinal scale, rather than a ratio scale (cf. Jensen's 2007 "Clocking the Mind").
> it used to be said that IQ was computed by dividing ones
> mental age by ones actual age.
Correct, with "used to" meaning "like a hundred years ago". In that sense "intelligence quotient" is an unfortunate misnomer, but it's part of the language now and would be hard to change (cf. Dawkins' "bright" campaign to re-brand atheism; well intentioned but so far not so effective).
> If brilliant scientists have low "IQ's", then the IQ test being used
> is not measuring anything interesting.
Another straw man. What "brilliant scientists" allegedly have low IQs? Unless, as some claim above, you consider 120 or 130 to be low.
> Do people still take them then? Why what is the point unless
> as MIH says it's to highlight potential difficulties in education?
Well, although they've largely been banned from various jobs because the results are found to systematically discriminate against some population groups, the military has been exempt from these PC-driven concerns because governments can't risk having extremey low-IQ people on the battlefield, however much that might hurt a potential recruit's feelings. And they're obviously still used as entrance exams to various schools (the SAT and LSAT, for example, are essentially IQ tests), because, lacking complete information, admissions boards have to make tough choices, and IQ/SAT/LSAT tests have demonstrably proven over the decades to be valid predictors of success, unlike touchy-feely "EQ" tests. Which is a good thing, unless of course you would prefer doctors and engineers to be "nice" and "empathetic" rather than "smart" and "effective".
> IQ tests measure people's itellectual potential and makes no
> pretension to predict if that potential will ever be realized.
That's definitely true, although IQ does correlate (moderately) with conscientiousness, i.e. elbow grease.
> High IQ is not a requirement to becoming a scientist.
Hmm nice notion, but sadly untrue. Try training someone with an IQ of 90 to become a theoretical physicist -- it won't happen. But books like Gladwell's "Outliers" -- which posits that anyone can achieve anything if they put in 10,000 hours of practice and have some good luck -- probably do have a positive knock-on effect for society in that they encourage people to "accomplish something", which is generally a desirable (if ultimately delusional) aim.
> I stopped caring about my IQ results when they started giving
> me lower results...
Haha they do drop a bit with age, nothing wrong with that.
> If IQ really measures some "intrinsic ability", then, at the
> absolute minimum, the scores you get on a test should
> stay relatively constant over time.
IQ tests, unlike "EQ" tests or personality tests, are notoriously difficult to game. You might be able to convince a test-giver that you're more extroverted or empathetic than you really are, but you can't score "better" on an IQ test just because you want to give the impression of being smart (though malingerers can make themselves seem more stupid than they really are, e.g. in order to be classified as retarded and thereby avoid the death penalty, but even this is more difficult than it might seem).
You might be able to add a few IQ points by assiduously practicing e.g. Raven's matrices, but even months of Kaplan-like SAT training can only only add maybe 40-50 SAT points (see e.g. the 2009 WSJ article "SAT Coaching Found to Boost Scores -- Barely",
http://online.wsj.com/article/SB124278685697537839.html)
> Your score shouldn't increase as you learn new things. This is
> not the case with actual tests.
As I mentioned above, IQ tests are a ranking. So of course the average 15-year-old has accrued more knowledge than the average 10-year-old, but his relative ranking among other 15-year-olds is probably about the same as it was when he was 10. That's what IQ tests are for.
> Look at virtually all Nobel winners and the quality of their work...
> pretty much all of them score above 120 and are closer to 140.
Exactly. And the top 1/4 of the top 1% of IQ scorers have 2x-3x as many Nobels and patents as the lower 1/4 of that top 1%; in other words, there's a significant difference in scientific success between an IQ 130 and IQ 145.
> I've heard there's a much larger scope for disorders and depression
> for those with an incredibly high IQ than there is for success
> relative to those of just above average IQ.
I think the evidence there is mixed. Though some disorders like schizophrenia and ADHD tend to skew low-IQ, depression and autism tend to skew high-IQ.
> Why is IQ such a secretive and elitist number?
Maybe that's a job for Wikileaks...
> IMHO intelligence is not something that is fixed, it's
> always changing
That's true, but it doesn't vary wildly, maybe a 10-point range. And it's largely fixed by age 11 or 12.
> For that matter, what is genius, anyway? Top 1%?
Well, by that standard (which equates in the US to an IQ of about 130), there would be about 3 million "geniuses" in the US alone, which kind of dilutes the value of the distinction.
