In article <[EMAIL PROTECTED]>,
Richard Ulrich <[EMAIL PROTECTED]> wrote:
>On 23 Apr 2004 09:23:26 -0500, [EMAIL PROTECTED] (Herman
>Rubin) wrote:
>> In article <[EMAIL PROTECTED]>,
>> Richard Ulrich <[EMAIL PROTECTED]> wrote:
>> Good studies have demonstrated that the most important factor
>> is purely genetic. Identical twins reared apart are more
>> similar that fraternal twins reared together, and in some
>> cases than identical twins reared together.
>Of the effects that they measured, genes were important.
>There is that generational effect, where IQ increases by 3
>points per decade since 1950, which lies on top of that.
>It sounds to me like you never did read about the Flynn
>effect, which I pointed you toward during an exchange in
>December, 2003:
Or is this transitory, with exposure to better
nutrition, and test taking "training". Multiple
choice tests are not TOO bad if students rarely
are exposed to them.
>======== start of extract from sci.stat.edu, Dec 4:
>[ RU here is message prior to Dec 4. Dec 4 has no initials.]
>RU >
>> >Yes, IQ scales are periodically re-normed. That is
>> >the STRONG, world-wide evidence for the Flynn
>> >effect -- the fact that researchers discovered that
>> >everyone around the world has to make the tests
>> >tougher, by a few points every decade, to keep the
>> >averages for normals at 100.
See above. Except possibly for such things as awards,
all scales should be absolute. We do not redefine
20 degrees Celsius according to the temperature at
which people are comfortable.
>HR >
>> Are the tests tougher? Or has it become memorization
>> and regurgitation, with the practice on taking multiple
>> choice tests raising the levels? The test designers are
>> very poor at thinking themselves, and thus have eliminated
>> this part from the tests. Besides, it is very difficult
>> to test thinking at all on multiple choice tests.
>Herman, you are floundering so bad here, you must
>be totally out of touch with psychometric literature.
>" memorization ... multiple choice ... the test designers... ".
>The Flynn effect was first documented in the standardization
>of IQ tests. *YOU* have been willing to talk about
>IQ tests as if they measured something real.
See the above. There had to be some initial standardization
of IQ measurements, and the original ones were not done by
the use of multiple choice tests. When my son was tested
at age 6, a multiple choice test was not used; he could
have taken one quite easily. It was considered unreliable.
But if the original standardization is wrong, we do not
go out and make a willy-nilly restandardization. The
original values used by Fahrenheit were wrong; nobody
has made the attempt to recalibrate so that normal body
temperature is 100 degrees. When there have been changes
in the scales or units, to the extent possible they were
within measurement errors. If selective breeding results
in a mean IQ of 120, we should still keep the same scale.
>A decade or so ago, one obvious "explanation" was
>that kids were being crammed with facts, etc.,
>for the tests.
They are being drilled, not crammed. Memorization,
beyond a fairly small amount, seems deadly to
understanding. This is why students today do not
understand concepts, but demand to be taught "how
to solve the problems on the tests".
But further study shows that big surprise:
>improvement seems to be biggest on the tests like
>the Raven Progressive Matrix that are the best
>regarded measures of 'g' or fluid intelligence or
>the ability to think fast (as opposed to crystallized
>intelligence or mere knowledge).
To think fast, or to recall fast?
>I haven't been keeping up, so I googled
>on <"Flynn effect" FAQ> . The first hit led me to
> http://www.indiana.edu/~intell/flynneffect.html
>which seems like a competent overview.
>That pointed me to a Flynn summary published
>in the American Psychologist, 1999, vol. 54, 5-20,
>which was also good. And has further references.
>===== end of extract. Back to current.
>RU> >
>> >And then there is this complication: Our measured IQ strongly
>> >does reflect something cultural and educational, particularly for
>> >those tests of 'abstract reasoning' (Raven's progressive matrices)
>> >that still are (I think) regarded as having the strongest genetic
>> >loading.
>> I doubt that the psychologists can understand abstract
>> reasoning, especially that it is not incremental. The
>Here at Pitt, the Grad. school in psychology used to depend
>heavily on an Analogies test, with (I estimated) an entry level
>of at least IQ 130. Is that abstract in the right way?
I am inclined to doubt this. Reasoning by analogy may
involve the process of abstraction, but it is not
abstract reasoning. The words have some connection,
but it is not precise. An abstract concept MAY have
been found as the abstraction of more concrete ideas,
but the connection then becomes that the abstract idea
can be considered as an idealization and hence applied,
and not necessarily to that concrete situation. We do
not apply the abstract "Euclidean geometry" to the
measurement of the earth, which is what the word means.
Abstract reasoning is the ability to deal with fully
abstract ideas; these are found in mathematics and logic,
and to some extent in foundational endeavors. Arrow's
proof that there is no consistent means to combine the
preferences of individuals to get a "social welfare" is
abstract reasoning. The current educational strategy of
building up to a more general or more structured approach
is a very poor way to teach abstract ideas; an abstract
idea is purely mental. When the mental concept is clear,
THEN it can be tied into appropriate real world usage.
>> place where it is most clear is in mathematical concepts,
>> and few outside of abstract mathematicians even see these,
>> which can be understood without too much difficulty by
>> children, but apparently not by those in "education".
> ... and the above seems, mainly, invidious.
>> > - By the way, Herman --
>> >Do you have any reference for your description of the 'better
>> >scaling' for IQ, with bigger 'real' differences at the high end, or
>> >is that personal, subjective opinion? I do not remember that.
>> It is partly subjective, but all of the "old" IQ tests,
>> produced before the introduction of the normal distribution,
>> have IQs ranging well over 160, including estimates of more
>> than 200. It was admitted that it was hard to measure at
>> this level, but this does not mean that scores above 130
>> should not be reported.
>Pragmatically, the IQ should be 'linear' with something, if it
>is going to be useful as a measurement. Isn't this one of
>your own guidelines, or an immediate corollary?
I have never stated that it should be linear. What I stated
is that if there is a linear relationship, it is destroyed
by nonlinear transformations. Is the Van der Waals law
(P + a/V^2)(V - b) = RT
linear? Is the Weber-Fechner law, or the other uses of
logarithmic scales, linear? There is a type of linearity
pervading science, but it is not in gross measurements.
Of course, one usually has LOCAL linearity. One of the
largest uses of least squares was in refining the mapping
of the United States, where the relations are non-linear.
Local linearity can be, and is used; it is used in refining
orbits of asteroids, etc.
>The statistician who first mentored me in 1970 offered the
>observation that there was very little 'functionality' for IQs
>that have been measured above 135 or 140 -- it did not
>translate to academic success (and IQ is strongly related to
>previous academic preparedness).
What do you mean by academic success? If you mean grades,
you are mistaken. I have had far too much experience with
students getting lots of A's who understood little, and
with those getting lower grades who could put it all together.
Nor did extra-high IQ
>predict success in business.
I would not expect it to. One apparent aspect of giftedness
is the inability to tolerate routine for long.
So. I did not know that recent
>tests were truncating the top end, but I find it easy to believe
>that the truncation of the score would be for linearity. It may
>take an IQ of 130 or so to become a successful theoretical
>physicist or mathematician, as I recall reading, but the points
>above that are mostly wasted. (I joined Mensa when I was
>19; I was not impressed at the couple of meetings I went to,
>though I still imagine that the high IQ points were measuring
>*something*.)
>I have read something else that relates.
>Regression-towards-the-mean between parents and children
>does exhibit a bit of what you describe: Smart parents *do*
>tend to have children who score a bit higher than predicted from
>the scores, as scaled. Thus, your modification of the scale --
>stretching the top -- would correct that bias. However, the
>discrepancy was not huge, and I don't find the alternate
>explanation difficult to accept: Namely, those same parents
>are better off economically and do encourage their children,
>so the extra points arise from simple social advantage.
Or is it the positive correlation between intelligence
and economic success? My immigrant parents were quite
poor during much of my early life; I did get a start,
and my father was quite definitely up there in intelligence.
However, my siblings, with at least the academic and social
status which I had, never showed any signs of more than
ordinary intelligence. Nor did our daughter, probably of
Mensa quality, show anything like our son. BTW, she
definitely got higher grades.
--
This address is for information only. I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
[EMAIL PROTECTED] Phone: (765)494-6054 FAX: (765)494-0558
.
.
=================================================================
Instructions for joining and leaving this list, remarks about the
problem of INAPPROPRIATE MESSAGES, and archives are available at:
. http://jse.stat.ncsu.edu/ .
=================================================================
