From: Olie Lamb (neomorphy@gmail.com)
Date: Thu Aug 24 2006 - 21:49:38 MDT
On 8/25/06, Michael Anissimov <michaelanissimov@gmail.com> wrote:
>
> G is a very interesting value which has been defined as what IQ tests
> are supposed to measure. The term "intelligence" is somewhat
> ill-defined, and as Michael Vassar mentioned, sometimes when you call
> somewhat "smart" you actually mean "rich", so there is a lot of
> ambiguity in these things. IQ, however, is much more precisely
> defined. It doesn't really matter is g is "THE" mechanism of
> intelligence or not, a strong predictor of behavior is a strong
> predictor of behavior.
>
Whoah, hang on there. You're admitting that "What IQ tests are
_supposed_ to measure" and "The mechanism of intelligence" are not
identical?
In other words - the "general intelligence factor" in humans is not
necessarily the same as the "thingamy" (mechanism) that is
intelligence.
So, you've pretty much admitted that IQ tests aren't necessarily
supposed to measure general-intelligence.
(From other arguments not raised, it's unambiguous that IQ tests don't
directly measure general-intelligence, they just measure things that
correspond to particular intelligence-related skills, and therefore 1.
indicate nothing about whether there is a "general intelligence" in
humans ("Reification"), and 2. if there is "general-intelligence" in
humans, they only measure it indirectly.)
Now, L Gottfredson, Sandberg and others are claiming that the
coefficient* of a number of different tests correlate, and are a good
indicator of, job performance in most jobs, and the likelihood of a
human to "do well" in society by a number of criteria. (Note that
some of the criteria are heavily culturally biased - I'd personally
view {children born in wedlock} as as much of a failing as {children
born out of wedlock}, as to me, wedlock is usually a symbol of the
triumph of hope over reason.)
*coefficient is not quite the right word, but nor is quotient.
Gottfredson admits that although "g" - by which I /presume/ she means
"IQ test scores" - has a strong correlation with job performance, that
other metrics are better indicators of job-specific job performance
(eg: manual dexterity in certain manual jobs).
Is it surprising that problem-solving tests are a good indicator of
job performance where problem-solving is a job component? Would it be
surprising that ability-to-speak tests were a good indicator of
performance in jobs where verbal interaction was necessary?
Even abilities that aren't directly used in a job will have certain
correspondence with job-performance metrics. Ability to juggle will
generally indicate lack of brain damage, completeness of the hands,
good eyesight etc. Consequently, I would expect a positive-yet-weak
correspondence between chemistry-lab-tech performance and juggling
ability.
...
It would be possible to take a set of tests derived from job-specific
skills, run them through a progressive matrix, and you'd get a test
that looked in some ways similar to an IQ test.
I'd bet good money that this "Job-Quotient" test would have a much
stronger correlation to job performance than an IQ test. Furthermore,
with this "JQ" test, you wouldn't have people complaining about
cultural bias etc, because any bias would only reflect the biases of
the world of employment.
...
Anyway, as I've said before (perhaps not here): There may be such a
thing as general-intelligence, but IQ-tests as they stand are only
loosely correlated with it.
General-Intelligence does not require symbol-use.
General-Intelligence does not require language use. A high level of
General-Intelligence does not imply knowing which spoon one uses for
dessert, and which for soup.
Although General-Intelligence can be used to figure out spatial
problems, a complex problem like chucking a stick at a moving target
in an environment of obstacles would require a fair chunk of GI to
tackle successfully. Humans DON'T use General-Intelligence for such
tasks, we use specialised visual modules to deal with them.
Now, the functionality of one "brain module" is going to be correlated
with the functionality of another. Malnutrition will damage auditory
function as it damages visual function. Brain health is also
vicarious.
--- IQ tests are "a Load of Crap" not because they are inconsistent (they aren't), not because they don't measure something (they do), not because they are unimportant (they are important), not because they are useless (they are useful). IQ tests are a Load of Crap because they don't measure what they claim to. -- Olie
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:57 MDT