John Knight wrote:
> "Mark D. Morin" <mdmpsyd at NOSPAMgwi.net> wrote in message news:<3D2F69E4.1020304 at NOSPAMgwi.net>...
>>>John Knight wrote:
>>>>>"Mark D. Morin" <mdmpsyd at NOSPAMgwi.net> wrote in message
>>>news:3D2EA20B.6040504 at NOSPAMgwi.net...>>>>>>>>>>John Knight wrote:
>>>>>>>>>>>>>"Mark D. Morin" <mdmpsyd at NOSPAMgwi.net> wrote in message
>>>>>>>>>>>>>>>>>>>>>>The mere mention of the differences between the sexes, in your mind, is
>>>>>>>"prejudice"!
>>>>>>>>>>>>>>Do we have that right?
>>>>>>>>>>>>No.
>>>>>>You are exercising circular reasoning.
>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>>Is this your final answer?
>>>>>>>>>>>>>>John Knight
>>>>>>>>>>>>>>>>>>>>>>>>>>>>ps--and now that you've concluded that "prejudice" is involved,
>>>>>>>>>>>>Prejudice has been involved for millenia. Prejudice exists despite
>>>>>>evidence and prejudice is what blinds people to evidence. Prejudice
>>>>>>evokes the circular reasoning that you have been so good at. Prejudice
>>>>>>is independant of data--show me the data fool.
>>>>>>>>>>>>>>>>>>>>>>>>>>>Well, it's pretty clear that you're just trolling now.
>>>>>>>>>>You haven't got the slightest idea if anyone's "prejudiced" regarding
>>>>>>>their
>>>>>>>>>>>opinion, whether it's their opinion of the data, or of the crooks who
>>>>>manipulated it. You didn't level that charge because you wanted to
>>>>>>>avoid or
>>>>>>>>>>>critique any possible prejudice. You levelled that charge only because
>>>>>>>you
>>>>>>>>>>>(and I) know that it's impossible for you to defend your position.
>>>>>>>>>>You keep repeating the mantra about "measurement errors" of one of the
>>>>>simplest things in the world to measure,
>>>>>>>>Who is trolling? What do you know about measuring brain volume or number
>>>>of neurons in a brain? Both are exceptionally unreliable.
>>>>>>>>>If it was this unreliable, then r-squared for brain size versus GRE scores
>>>would be nowhere NEAR 0.8795. The ONLY thing that could be gained by even
>>>more reliable measurements would be an increase in r-squared.
>>>>you don't understand the concept of reliability. It is scientificly
>>dishonest to take the measurement that yeilds the best correlation and
>>then throw out the rest of the data--as your links suggest. Your 0.8795
>>is a figment of your imagination.
>>>>> Your argument is that there are measurement errors. But if that were
> true, then correlation would be much lower than 0.8795.
and the authors reported r's less than 0.8. MUCH lower. And didn't
bother to say why they simply picked the higher.
And even IF
> there are measurement errors, they clearly aren't significant enough
> to interfere with the correlation.
Which correlations?
a correlation is simply a measure of association between two numbers.
you have yet to demonstrate that the correlation in question can be
reliably reproduced.
>> You can bet that Philippe Rushton didn't pick and choose the numbers
> that fit the curve, because he attempted to correlate brain size to
> "IQ scores" rather than to GRE scores. He was surprised at the low
> r-squared he got from doing this. The error was that IQ scores were
> intentionally manipulated when Wechsler et. al. threw out 94% of the
> IQ problems which are EXACTLY the most important problems.
The articles you cite in your sig line do in fact suggest that numbers
were picked and choosen.
>> They weren't removed from the GRE, nor the TIMSS.
Strange, the GRE research was not able to replicate the findings you cite.
>>>>>Again, the only outlier is the Black man, and if he's removed from the
>>>calculations, r-squared gets very close to 1.0. You *must* know that you
>>>cannot get better correlation than that.
>>>>>>Where are the measurement errors?
>>>>brain measurements
>>>>>>>>>>>>>>without ever pointing out
>>>>>specifically where you believe these measurement errors are. This was
>>>>>simply an attempt to discredit what you know (or at least sense) to be a
>>>>>FACT.
>>>>>>>>>>You're willing to go to great lengths to attempt to refute the
>>>>>>>correlation
>>>>>>>>>>>between brain size and GRE scores, no matter what the facts are.
>>>>>>>>The fact is, there is a robust correlation between age and brain volume.
>>>>If age is not controlled for, you can not say anything about between sex
>>>>differences.
>>>>>>>>>>>>>This is patently false, and since it's already been explained to you, it's
>>>not clear if you're just trolling, or if you really don't understand how
>>>inconsequential age and brain size are to the correlation we're discussing.
>>>>>>Let's try this a different way. Even IF "there is a robust correlation
>>>between age and brain volume", you cannot also ignore that there is also a
>>>"robust" correlation between brain volume and GRE scores,
>>>>a finding reported at a single site. not published, not reviewed, using
>>data that were sorted through to find optimal numbers. Point to a
>>published piece of literature, in a peer reviewed journal that supports
>>the above assertion.
>>>>> Guess what? I don't care WHAT's been published. Do you know why?
> Because what I've formed my own opinion and i don't care what the facts are.
>>>>>as brain volume increases, so do GRE scores. Any adjustments for
>>>measurement errors for increased brain volume would be cancelled out by the
>>>increase in GRE scores, and you'd be back to r-squared = .8795
>>>>>>But since you haven't even attempted to quantify this "robust correlation",
>>>it should be pointed out that after the age of 20, average brain volume
>>>remains virtually FIXED by race and sex.
>>>>heh? you are disregarding the past twenty years of research? The rate
>>of shrinkage by age is close to "fixed" but even then, there's a lot of
>>variation. LOOK at an MRI of a 90 year old brain and compare it to that
>>of a 20 year old. Even you should be able to see a difference.
>>>>> The point you keep ignoring, probably because you don't understand the
> futility of your argument, is that adjusting for any "shrinkage" in
> brain size is NOT necessary in the first place, because correlation is
> already remarkable.
Correlation will simply provide an estimate of explained variance.
Place gender and age in the regression equation and there will not be
any variance left over to be explained by gender.
The other point is that adjusting for "shrinkage"
> could not possibly reduce r-squared, because it's inevitible that if
> brain size and GRE scores are correlated (which they obviously are),
obvious only to you.
> and if there really is some significant "shrinkage" that needs to be
> considered, that r-squared would *increase*.
>> You can deny, deny, deny all you want. Your argument gets weaker the
> longer you fail to demonstrate how you think there are "measurement
> errors" in the fact of an r-squared of almost 0.9.
try pulling another number out of your ass.
>> John Knight
--
====================================================
You can't make someone love you, but you can let
yourself be loved by someone.
http://home.gwi.net/~mdmpsyd/index.htm
