[CLIP]
[from zaimoni@....]
> >It would take some extreme education to increase intelligence. Most
> >education doesn't affect intelligence, just the knowledge base.
>
> I disagree with this.
>
> Define intelligence in a testable way. How does the Princeton Review help
> people increase SAT/GRE/LSAT/MCAT scores...they teach you how to psyche out
> the test: provide you with the proper memes, huh? They have a pretty high
> success rate in increasing peoples "intelligence"...if you don't think
> standardized tests measure intelligence (and I don't, either) what should
> we use in place
I do not consider the SAT/GRE/LSAT/MCAT etc. intelligence tests. Having
taken the SAT, I know first-hand that the SAT is a very general
knowledge-base test. It has inadequate resources for testing for raw
speed, and for generic short-term memory capacity. I have no reason to
believe the others are any different. [as you might guess, these are key
to my definition of intelligence.]
As such, it is usually possible to learn how to 'psych out' a
knowledge-based test, by being able to spot obviously wrong answers. I
recall a test to try to evaluate US national performance in math, where I
applied modulo ten arithmetic combined with order of magnitude estimates
to savage the arithmetic section in less than a sixth of the alloted
time--in the 6th grade. Did I psych out the test, or merely apply
evaluation methods that were highly unconventional to a multiple-choice test?
Any multiple-choice test in physics or chemistry is also vulnerable to
this [possibly requiring college algebra]. I mention this as one of the
things an A in College Algebra is good for, to my College Algebra
students on the first day of class. ["If you flatten this course,
you can still get an A on a low-level physics or chemistry test without a
calculator."]
=====
An intelligence test really should be a psychometric test, covering at least:
Generic short-term memory capacity, before applying mnemonics.
Raw speed of thought. Depending on the psychological model used,
more than one mode may have to be tested.
The paradox [as you mentioned below, in a point I agree with (and
clipped)] is that you need some content in order to test the above two
traits.
As such, any so-called intelligence test, if honest, should have a
qualifier as to what culture it is valid for, and any physical
disabilities that interfere with its effectiveness.
For instance, the Stanford-Binet IQ test would be a first-order
approximation for WASP American subculture. Its effectiveness is
suspect in other American subcultures, even. Of course, the arithmetic
subsection is vulnerable to mind-set effects.
Both of the traits I mentioned as minimal coverage for an intelligence
test CAN be changed by education, but they won't change just from
attending class and doing the homework. It is quite possible to get a
Bachelor's degree in most intellectually demanding fields without causing
significant shifts in these traits.
[CLIP}
//////////////////////////////////////////////////////////////////////////
/ Towards the conversion of data into information....
/
/ Kenneth Boyd
//////////////////////////////////////////////////////////////////////////