science and racism? [Archive]

bpesta22

23rd July 2004, 06:47 PM

If you insist on citing crappy mismeasure, then please also include the good science in the area that completely contradicts Gould's most embarassing work. See, e.g.,

Hunter, J., & Schmidt, F. (2000). Racial and gender bias in ability and achievement tests: Resolving the apparent paradox. Psychology, Public Policy, and Law, 6, 151-158.

Jensen, A. (2000). Testing: The dilemma of group differences. Psychology, Public Policy, and Law, 6, 121-127.

Neisser, U., Boodoo, G., Bouchard, T., Boykin, W., Brody, N., Ceci, S., Halpern, D., Loehlin, J., Perloff, R., Sternberg, R., & Urbina, S. (1996). Intelligence: Knowns and Unknowns. American Psychologist, 51, 77-101.

Reynolds, C. (2000). Why is psychometric research on bias in mental testing so often ignored? Psychology, Public Policy, and Law, 6, 144-150.

Schmidt, F., & Hunter, J. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124, 262-274.

Anything recent by Jensen-- the G factor-- would also be a good source of what real scientists in this area have concluded re race differences in iq.

Dancing David

23rd July 2004, 08:29 PM

Except for the fact that they don't really control for the differences in population.

Sorry Bpresta , I still respectfully disagree. They have not matched the demographics enough to prove that IQ is really different.

Is there some new evidence that IQ really has a high coorelation to anything. Like a 60% coorelation?

Anticipation and respect.

(PS How would you do on the !Kung IQ test?)

bpesta22

23rd July 2004, 10:19 PM

Originally posted by Dancing David
Except for the fact that they don't really control for the differences in population.

Sorry Bpresta , I still respectfully disagree. They have not matched the demographics enough to prove that IQ is really different.

Is there some new evidence that IQ really has a high coorelation to anything. Like a 60% coorelation?

Anticipation and respect.

(PS How would you do on the !Kung IQ test?)

Hey DD.

The differences have been replicated since WW I, and the effect size is about 1.0 for black white differences.

Effect size is what you're asking about, since you seem to want large correlations before assuming they're meaningful.

There's a famous statistician named Cohen who has guidelines for how to interpret effect sizes. Note, he does no research on IQ, these are guidelines for effects in any area:

.2 = small
.4 = medium
.6 = large.

He doesn't even have a label for 1.0, which must therefore be HUGE!

At any rate, it's true that within group variability is larger than between race differences. However, an effect size of 1.0 has serious and profound implications for blacks and whites with regard to various outcomes at the group level.

You'd be hard pressed to find a more replicated effect in all of experimental psychology then the fact that blacks and whites score differently on unbiased tests of intelligence.

B

Dancing David

24th July 2004, 06:41 AM

I understand the test but I would claim a cultural bias to the testing. (How did you do on the !Kung IQ test?)

My question is this:

How can you control for the single confounding factor of a single enviromental toxin, lead?
As the forties and fifties progressed in America there is a shift in population, white populations to the suburbs and AA populations to the inner city. It has been shown repeatedly that inner city enviroments are very high in lead dust and other forms of enviromental lead.

So who do you control for the fact that it might be exposure to lead which creates or contributes to the difference?

Jeff Corey

24th July 2004, 07:17 AM

...You'd be hard pressed to find a more replicated effect in all of experimental psychology then the fact that blacks and whites score differently on unbiased tests of intelligence. [/B]
But we are still faced with the problem that this is not experimental evidence. No independent variable was manipulated, so all these studies boil down to correlations.

Rob Lister

24th July 2004, 07:20 AM

Originally posted by Dancing David
I understand the test but I would claim a cultural bias to the testing. (How did you do on the !Kung IQ test?)

Must ask!!!!!!!!!!!!!!!!!!!!!!!!!!!!
What is the !kung iq test?

Jeff Corey

24th July 2004, 08:23 AM

You are dropped into the middle of the Kalahari Desert and are forced to find water and forage for berries and grubs.

wittgenst3in

24th July 2004, 08:42 AM

Originally posted by Jeff Corey
You are dropped into the middle of the Kalahari Desert and are forced to find water and forage for berries and grubs.

Pass.

I think this without a doubt proves my intelligence.

pupdog

24th July 2004, 08:56 AM

It has been shown repeatedly that inner city enviroments are very high in lead dust and other forms of enviromental lead.
And effects of lead on IQ were noted at least as early as 1943 (reported in Time magazine, December, based on a report published in American Journal of Diseases of Children).

bpesta22

24th July 2004, 11:38 AM

I'm not arguing for the cause of the differences, just that they exist.

I dunno what causes them-- it could be lead.

I don't think anyone knows what causes the difference, but it's not cultural bias.

The tests are race blind in that they predict just as well for minorities as they do for whites.

I'd submit there's a whole bunch of IQ tests I'd score really poorly on. But, calling it an iq test doesn't make it one. There are issues like reliability and validity that determine if the test indeed measures iq.

As far as I know, unless the test measures "g", you can call it iq, but it wont predict any important social outcome. On the other hand, if the test measures g, it predicts every important social outcome, and unfortunately produces large race differences.

Jeff Corey

24th July 2004, 12:07 PM

Originally posted by bpesta22
... On the other hand, if the test measures g, it predicts every important social outcome, and unfortunately produces large race differences.
Every important social outcome? They correlate about .2 to .3 with success in a number of different jobs. That's not accounting for much variance.
Wagner, R. K. (1997) Intelligence, training and employment. American Psychologist, 52, 1059-1069

bpesta22

24th July 2004, 03:14 PM

Dancing David

24th July 2004, 03:28 PM

Originally posted by Jeff Corey
But we are still faced with the problem that this is not experimental evidence. No independent variable was manipulated, so all these studies boil down to correlations.

Well stated!
And a lack of demographic controls.

Dancing David

24th July 2004, 03:30 PM

Originally posted by Rob Lister
Must ask!!!!!!!!!!!!!!!!!!!!!!!!!!!!
What is the !kung iq test?

The !Kung are also called Kalaharui Bushmen, I assume that all intelligence tests are culturaly biased and that thiers would be very different from the Eurocentric one.

wittgenst3in

25th July 2004, 03:03 AM

Originally posted by Jeff Corey
You are dropped into the middle of the Kalahari Desert and are forced to find water and forage for berries and grubs.

I'm taking this way to seriously but...

Just realised, but this probably isn't a true intelligence test. If I remember correctly the tests are not supposed to test your education level, but more your general aptitude for problem solving etc.

Meaning that you wouldn't drop a adult Kalahari bushman that had been specifically trained by the 'tribe'(?) out in the desert, because he's been trained for this specific situation.

On the other hand comparing a Kalahari childs responses compared to another childs responses would be a valid test of KIQ.

Jeff Corey

25th July 2004, 07:19 AM

Originally posted by wittgenst3in
I'm taking this way to seriously but...

Just realised, but this probably isn't a true intelligence test. If I remember correctly the tests are not supposed to test your education level, but more your general aptitude for problem solving etc.

Meaning that you wouldn't drop a adult Kalahari bushman that had been specifically trained by the 'tribe'(?) out in the desert, because he's been trained for this specific situation.

On the other hand comparing a Kalahari childs responses compared to another childs responses would be a valid test of KIQ.
No, they don't test your general aptitude for problem solving. They measure your probable performance in school or similar settings. That was the original purpose and Simon and Binet (1905) accomplished that task pretty well.
A number of people, such as Howard Gardner and Robert Sternberg, present a strong argument that there is no unified g, but rather a number of different factors involved.
So here's a question.
Washington is to Lincoln as:
a. 1 is to 5
b. 1 is to 16
c. 10 is to 7
d. All of the above

wittgenst3in

25th July 2004, 07:40 AM

Originally posted by Jeff Corey
Washington is to Lincoln as:
a. 1 is to 5
b. 1 is to 16
c. 10 is to 7
d. All of the above

A) George is on the $1 bill and Abe is on the $5.
B) George is the 1st prez and Abe is the 16th
C) George's surname takes 10 letters and Abe's takes 7

Therefore D is correct.

/Not even american
(minor googling was required to confirm my suspicians of B and A).

Jeff Corey

25th July 2004, 07:47 AM

Originally posted by wittgenst3in
A) George is on the $1 bill and Abe is on the $5.
B) George is the 1st prez and Abe is the 16th
C) George's surname takes 10 letters and Abe's takes 7

Therefore D is correct.

/Not even american
(minor googling was required to confirm my suspicians of B and A).
But you're not allowed to google when you take the test. And it's -25 points for spelling "suspicions" wrong.

Bikewer

25th July 2004, 08:01 AM

Although Gould-bashing seems to be in vogue currently, one of the more important points of the above-mentioned book was that human intelligence is not something easily classified by a single number.
As I recall, the standard IQ test was developed simply as a diagnostic tool, rather than some sort of defining quantification of intelligence.

From my readings in neuroscience, I believe that the current thinking is that "intelligence" can be subdivided into a number of different skill-sets and abilities, with most folks excelling in some and deficient in others.

bpesta22

25th July 2004, 08:09 AM

Originally posted by Bikewer
Although Gould-bashing seems to be in vogue currently, one of the more important points of the above-mentioned book was that human intelligence is not something easily classified by a single number.
As I recall, the standard IQ test was developed simply as a diagnostic tool, rather than some sort of defining quantification of intelligence.

From my readings in neuroscience, I believe that the current thinking is that "intelligence" can be subdivided into a number of different skill-sets and abilities, with most folks excelling in some and deficient in others. .

I'd be willing to buy this, but:

1) why does the single number representing one's IQ predict so many important variables. And, why does the prediction depend critically on how much the single number is a good measure of g.

2) Why is the positive manifold (positive correlations among tests) always found when a group of diverse mental abilities tests is given

I don't claim that g is the only type of intelligence, just that it's the only one that matters when predicting things.

I have a lot of respect for Gould as an evolutionary biologist, but he really should have stuck to what he knows.

bpesta22

25th July 2004, 08:35 AM

Originally posted by Jeff Corey
Every important social outcome? They correlate about .2 to .3 with success in a number of different jobs. That's not accounting for much variance.
Wagner, R. K. (1997) Intelligence, training and employment. American Psychologist, 52, 1059-1069

Ok, did some research on this.

First, Wagner's a colleague of Sternberg's at Yale. Go figure, he gets his American Psychologist pub when Sternberg is editor of the journal.

You want to see some crap science, read any Sternberg AP paper on iq.

Anyway, from the article you cite:
Beginning in the 1970s, two developments dramati-
cally changed this picture. First, the development of meta-
analysis, arguably one of the most influential method-
ological developments in recent decades, made it possible
to cumulate quantitatively the results of large numbers
of small-scale studies, resulting in a quasi-massive-scale
study. Second, the results of large-scale studies of mili-
tary personnel and others also became available. The re-
suits of both kinds of studies provided strong evidence
of remarkably general validity for cognitive ability tests
for selection across a broad range of jobs (Hunter &
Schmidt, 1983; McHenry, Hough, Toquam, Hanson, &
Ashworth, 1990).

Then, Wagner says:

The average observed validity coefficient or correla-
tion between cognitive ability test scores and job perfor-
mance is between .20 and .30 (Hartigan & Wigdor, 1989;
Wigdor & Garner, 1982).

Just two studies cited, neither is a review article, neither is a meta-analysis.

But, then he qualifies:

When observed validity coeffi-
cients are corrected for measurement artifacts such as
restriction of range and measurement error, the average
validity coefficient increases to about the .50 level
(Hunter & Hunter, 1984; Ree & Earles, 1992; Schmidt &
Hunter, in press). A validity coefficient at the .50 level
is an impressive accomplishment.

Then he tries to argue this is an overstatement due to multicolinearity. He cites an in-press article by schmidt and hunter which sorta makes his point.

One year after this article, though, we get

Schmidt, F., & Hunter, J. (1998). The validity and utility of selection methods in personnel psychology: Practical and theoretical implications of 85 years of research findings. Psychological Bulletin, 124, 262-274.

They have a table-- I will go through the trouble of posting it here if someone asks-- showing the validity of various selection methods, controlling for other methods. The table also shows how much other methods add to validity over and above IQ.

It's impressive support for using IQ tests for selection.

Further, if you add the cost of an IQ test ($1.72 if bought in bulk) and how long it takes to administer (8 minutes!), hands down IQ is the most useful selection method in existence.

See:

http://www.wonderlic.com/

Bikewer

25th July 2004, 08:44 AM

I'm not a psychologist, just a police officer with a rather limited time to devote to my "studies".

Still, I'd wonder:

What might be the definition in these studies of "black"?

As has been noted, there is essentially no genetic difference between groups of human beings, who after all comprise but a singe species. Genetic variation within "racial" groups being greater than the differences between them.

"Black" is a term applied to extremely diverse groups of people, and no population on earth could claim to be "pure" in this regard.

Certainly this applies even more so to black people living in the Americas, all of whom would no doubt have in their ethnic background a wide variety of other influences.

So, are we talking about the stereotypical "inner-city" people, the product of poverty, deficient schools, poor nutrition, lead poisoning, etc. etc., or a broader mixture of peoples from a variety of backgrounds and cultural settings?

Could variables in these criteria include culture? I seem to recall that children of Asian descent generally score better than most other groups in a variety of testing. Would we assume a genetic superiority for these children, or perhaps a culture that prizes scholarship and achievment from an early age? (Not to mention culturally-reinforced "filial piety")

Testing human beings is fraught with difficulty, as it's so difficult to control life experiences. Has there been any attempts to study children of caucasian descent living lives of poverty, neglect, poor nutrition, etc? The Appalachias spring to mind, and I believe that there have been a number of studies done, though I couldn't cite sources.

Virgil

25th July 2004, 08:48 AM

we are all of the human race.

V

Dancing David

25th July 2004, 11:43 AM

Originally posted by bpesta22
I'm not arguing for the cause of the differences, just that they exist.

You used the word effect in your priot post that would imply that the difference of IQ scores was an effect of the racial difference, perhaps I misread that but I think you used the word effect.

If the question is about race then it would be important to weed out the socio econmic factors associated with poverty.

I dunno what causes them-- it could be lead.

I don't think anyone knows what causes the difference, but it's not cultural bias.

That is the question is it not Hamlet?

A lot of people do believe that because of thier construction IQ tests are culturaly biased.

The tests are race blind in that they predict just as well for minorities as they do for whites.

The last set of correlation I saw were so low that i can't see them being predictors of much, except ability to take IQ test.

I'd submit there's a whole bunch of IQ tests I'd score really poorly on. But, calling it an iq test doesn't make it one. There are issues like reliability and validity that determine if the test indeed measures iq.

Which is why I have argues in the past that the tests need to be redesigned, say for various and targeted skills, work, social, quality of life, the methoid that was used to construct the MMPi could produce a series of quotient test that would probably have a 90% correlation with the target area.

As far as I know, unless the test measures "g", you can call it iq, but it wont predict any important social outcome. On the other hand, if the test measures g, it predicts every important social outcome, and unfortunately produces large race differences.

And is g correlated above the 60% level?

Dancing David

25th July 2004, 11:48 AM

Originally posted by bpesta22
.

I'd be willing to buy this, but:

1) why does the single number representing one's IQ predict so many important variables. And, why does the prediction depend critically on how much the single number is a good measure of g.

Is that correlation in the high sixties?
Are the predictors subject to the same demographic and cultural bias of the tests?

2) Why is the positive manifold (positive correlations among tests) always found when a group of diverse mental abilities tests is given

How high are these correlations?

I don't claim that g is the only type of intelligence, just that it's the only one that matters when predicting things.

I have a lot of respect for Gould as an evolutionary biologist, but he really should have stuck to what he knows.

If you read Gould you know that he is talking about the Victorian bias of race and it is in context of that, then there is the later stuff about modern mismeasurement.

What are you specificaly taking Gould to task for, have you ever studied anthroplogy/ I suppose you condem anthroplogists for not sharing your views as well?

Eos of the Eons

25th July 2004, 02:03 PM

I saw a test on college students where they looked for some genetic material from a very old specimen (I think one like lucy). They found some caucasion students had more genetic history with an american indian classmate than to each other. They were more related to the indian than to others that "looked" more like themselves.

When you get such a mixture as we are getting in North America, you certainly cannot tell as much with what genes are expressed for looks (eye colore, sking color), than those that you actually carry.

Each individual does not know their whole genetic history either. My son has a "black" great great grandparent. He has blonde hair and hazel eyes. If his dad and I were a one night stand or something like that, I wouldn't know that.

Regionally you may share similar IQs, but that doesn't show your actual dna history either. You can also get a 'black' with a higher IQ than a random Asian.

So no matter what is expressed (intelligence, hair color, foot size), you still have to look at the actual dna history, and that will show we are quite related in ways you'd never think.

Thus, to group people by races or intelligence types, or whatever is missing the point.

You may have more dna in common with the complete stranger 'black' or 'white' or 'indian' guy on the bus than with your "same race" friend. This is because wayyyyy back you could be descended from one common person (Lucy), or two sisters (your direct ancestor being Lucy, and the stranger's direct ancestor being Lucy's sister), or ...and not have a single other direct/indirect ancestor that is related to your friend.

I hope this makes sense.

bpesta22

25th July 2004, 02:15 PM

Originally posted by Dancing David
If you read Gould you know that he is talking about the Victorian bias of race and it is in context of that, then there is the later stuff about modern mismeasurement.

What are you specificaly taking Gould to task for, have you ever studied anthroplogy/ I suppose you condem anthroplogists for not sharing your views as well?

No, I condemn Gould for setting up a huge strawman, and presenting decades old research as if it were mainstream science, and for not understanding factor analysis, and then blaming those who use it.

Search Arthur Jensen replies to S. J. Gould-- it's an interesting read.

Why do you insist a correlation must be in the high .60 to have practical value.

search utility, IQ and job selection -- you'd be surprised the $$$ that even very low validities could save a company for use in selection.

The MMPI empirical keying method might work at identifying high IQ people. But since g is already reliabily and validly measured, I bet the MMPI scale would create adverse impact too.

You can devise all these domain specific tests, and if they're done right, they will be valid predictors of whatever specific skill or ability it is you want tested. However, are these tests worth the time and expense (you need one test for each skill you measure) when $1.76 and 8 minutes gives you premium validity?

Also, of all domain specific abilities ever measured, none predict success as well as (most even at all, when controlling for) g.

bpesta22

25th July 2004, 02:18 PM

Originally posted by Eos of the Eons
I saw a test on college students where they looked for some genetic material from a very old specimen (I think one like lucy). They found some caucasion students had more genetic history with an american indian classmate than to each other. They were more related to the indian than to others that "looked" more like themselves.

When you get such a mixture as we are getting in North America, you certainly cannot tell as much with what genes are expressed for looks (eye colore, sking color), than those that you actually carry.

Each individual does not know their whole genetic history either. My son has a "black" great great grandparent. He has blonde hair and hazel eyes. If his dad and I were a one night stand or something like that, I wouldn't know that.

Regionally you may share similar IQs, but that doesn't show your actual dna history either. You can also get a 'black' with a higher IQ than a random Asian.

So no matter what is expressed (intelligence, hair color, foot size), you still have to look at the actual dna history, and that will show we are quite related in ways you'd never think.

Thus, to group people by races or intelligence types, or whatever is missing the point.

You may have more dna in common with the complete stranger 'black' or 'white' or 'indian' guy on the bus than with your "same race" friend. This is because wayyyyy back you could be descended from one common person (Lucy), or two sisters (your direct ancestor being Lucy, and the stranger's direct ancestor being Lucy's sister), or ...and not have a single other direct/indirect ancestor that is related to your friend.

I hope this makes sense.

Line me up 10000 randomly selected americans. I will group them by eyeball according to race. Give them all a g loaded IQ test, and you will get these "race" differences we're talking about here.

I'd bet my life on it!

Dancing David

25th July 2004, 02:24 PM

Originally posted by bpesta22
No, I condemn Gould for setting up a huge strawman, and presenting decades old research as if it were mainstream science, and for not understanding factor analysis, and then blaming those who use it.

Search Arthur Jensen replies to S. J. Gould-- it's an interesting read.

Why do you insist a correlation must be in the high .60 to have practical value.

search utility, IQ and job selection -- you'd be surprised the $$$ that even very low validities could save a company for use in selection.

The MMPI empirical keying method might work at identifying high IQ people. But since g is already reliabily and validly measured, I bet the MMPI scale would create adverse impact too.

You can devise all these domain specific tests, and if they're done right, they will be valid predictors of whatever specific skill or ability it is you want tested. However, are these tests worth the time and expense (you need one test for each skill you measure) when $1.76 and 8 minutes gives you premium validity?

Also, of all domain specific abilities ever measured, none predict success as well as (most even at all, when controlling for) g.

Well I suppose we will agree to disagree then for I find correlations below that level to be unindicative and the fact that the IQ test exists does not mean a better one could not be devised.

I was just using the MMPi as an example of how a useful tool could be contructed post hoc, but I rant even more about how useless the MMPi is that I do the IQ score.

Best wishes and fare thee well.

Dancing David

25th July 2004, 02:25 PM

The Khe San in southern africa are 'caucasian' in morphology yet most ceratinly 'black' in heritage, how does one classify them then?

I thought I read recently that rural white students preform as poorly as urban black students in standardized tests, hmm what would that indicate?

Eos of the Eons

25th July 2004, 03:45 PM

Originally posted by bpesta22
Line me up 10000 randomly selected americans. I will group them by eyeball according to race. Give them all a g loaded IQ test, and you will get these "race" differences we're talking about here.

I'd bet my life on it!

Yes, but we're missing the point again. Certain traits from geological clustering does not change the fundamental story our dna can tell us about ourselves and where we came from, and how we are actually related.

It doesn't matter what "race" we are, it doesn't change the answer to the original question posed by Ixabert.

Ixabert needs to show that race doesn't matter, and it doesn't. DNA is not the only reason for how intelligent a person ends up looking on an IQ test. We all know that.

American

25th July 2004, 05:02 PM

Originally posted by Ixabert
I am searching for info disproving anti-black racism.

That's how I do science. I don't investigate reality- I set out to prove things, no matter how ridiculous or misleading, and with total disregard for social consequences.

Eos of the Eons

25th July 2004, 07:07 PM

Originally posted by American
That's how I do science. I don't investigate reality- I set out to prove things, no matter how ridiculous or misleading, and with total disregard for social consequences.

Showing that you should not discriminate against others just because they have a different skin color is not ridiculous, misleading, or having a total disregard for social consequences.

Reality is that we can be more related to someone with a different skin color than someone with the same skin color in many cases.

bpesta22

25th July 2004, 07:41 PM

Originally posted by Dancing David
Well I suppose we will agree to disagree then for I find correlations below that level to be unindicative and the fact that the IQ test exists does not mean a better one could not be devised.

I was just using the MMPi as an example of how a useful tool could be contructed post hoc, but I rant even more about how useless the MMPi is that I do the IQ score.

Best wishes and fare thee well.

Well, surely a better iq test can be devised. The validity is only .5 (for job performance). So far, no one's been able to do it.

I like the MMPI, but maybe we should save that for another thread....lol

bpesta22

25th July 2004, 07:49 PM

Originally posted by Eos of the Eons
Yes, but we're missing the point again. Certain traits from geological clustering does not change the fundamental story our dna can tell us about ourselves and where we came from, and how we are actually related.

It doesn't matter what "race" we are, it doesn't change the answer to the original question posed by Ixabert.

Ixabert needs to show that race doesn't matter, and it doesn't. DNA is not the only reason for how intelligent a person ends up looking on an IQ test. We all know that.

But, it does matter-- we just dont know why.

For some reason, races score differently on iq tests-- tests which predict at the group level important social outcomes.

ASAIK, so many studies have been done controlling for this that and the other thing in the environment, only to find the race difference is still there.

I'll have to dig up the cite, but when you match blacks and whites on income levels and parent's education, the race difference is actually largest at the wealthy end of the comparison!

I honestly don't believe i am a racist. But just ignoring 80 years of data by saying the tests are biased, or it's the "envionment," seems intellectually dishonest.

These data piss people off, and they should. But, if the science werent there to back it up, we wouldn't be using iq tests today-- for sure, there's no way the courts would allow companies to use iq tests in selection-- when those tests exclude proportionately more minorities from being hired-- if they weren't well validated and cultually fair.

We all have differing levels of knowledge and expertise in this area, but ask yourself: Would our federal court system, and the teeth of the EEOC (see that 24 million dollar sexual harassment verdict in the news lately) ever allow a firm to use an IQ test for selection-- thereby legally discriminating against minorities, if 1) the test wasn't valid, or 2) the test showed intercept or slope bias against minorities.

Skeptic

25th July 2004, 08:20 PM

There's a famous statistician named Cohen who has guidelines for how to interpret effect sizes. Note, he does no research on IQ, these are guidelines for effects in any area:

.2 = small
.4 = medium
.6 = large.

He doesn't even have a label for 1.0, which must therefore be HUGE!

(sigh)

This is one of the "factoids" that are often presented in favor of a "huge" difference between blacks and whites in IQ, but it shows more of the ignorance of those who promote it.

When a statistician says that a certain difference is "large" as opposed to "small", he means something like "a difference this size between populations is easily detected even with a relatively small sample from the two populations." It does NOT mean that the effect is absolutely "large" in terms of actual units (as opposed to SDs), let alone that whatever difference exists is of any practical importance.

For instance, suppose that every single white person had an IQ of exactly 100, and every single black person had an IQ of exactly 99. Mathematically speaking, the difference between the two is not 1, or 10, but literally an infinite number of standard deviations (since the SD of both populations is exactly 0). This is, indeed, a "huge" difference for a statistician, in the sense that there is an absolute, 1:1, correlation between race and IQ in this case: one's race totally determines what one's IQ is. So? It would still be still a tiny difference both absolutely (one IQ point) and practically (in terms of differences in "real life" performance).

The one SD difference in IQ, even if real--which is doubtful--might be "huge" in Cohen's statistical sense and still "tiny" or "insignificant" in any practical sense, in the same way that an EXTRA LARGE coke bottle can fit in a SMALL refrigirator. (mis)quoting Cohen in this way is merely using words from one context in another context, with the clear intent of making the IQ difference between blacks and white look as "important" as possible.

American

25th July 2004, 09:13 PM

Originally posted by Eos of the Eons
Showing that you should not discriminate against others just because they have a different skin color is not ridiculous, misleading, or having a total disregard for social consequences.

There's no should or should not about it, unless you define your motivations, which we haven't. The things you mention are a matter of opinion and individual values, not the realm of science.

All of this pre-supposes that today's psychology or sociology are valid sciences, which I say that they aren't, rather it's all political BS. So our arguing can go nowhere, and I will let you have the last word if you wish.

Eos of the Eons

25th July 2004, 11:03 PM

Oh bother, I was hoping you misunderstood, but not really. Thing is, we are using science, and science has shown that outter traits matter less when it comes down to who you are actually related to.

A really dumb person can be related to a really smart one. A black person can be related to a white one. When you hate one race or another, you are hating people you are/could be directly related to.

The dna test showed the caucasion in the class was more related to an american indian than another caucasion. That is science, not opinion.

Intelligence and skin color can cluster in areas because of the environmental stresses that push evolution one way or another. I have agreed with Pesta, showing why. Thing is, it doesn't matter. You can group people by skin color, and IQ test scores, etc. That doesn't make you any less related.

When you group people by actual inheritance, you find a mosaic, and you find that it is essential to have different traits in order to ensure survival.

So some folks don't do as well on IQ tests, but they'll whip your butt in a foot race. Or they will have less instances of skin cancer.

You are lumping people by traits. When it comes to genetics though, it's the group of traits that matters and who you are related to. The group of traits will matter when it comes to survival. Who you are related to doesn't really matter, but it shows you never know who you are more related to.

Point being, there is value to differences, and value to recognizing those differences are minor in the "big picture".

There is a story that I don't know is true. There was this racist who found out he had a disease. Turns out that only "black" people got the disease, so it turned out he was closely descended enough to a black person to have the trait expressed in his genome. He didn't look black, but he was racist against his own relatives.

Thing is, we do know that humanity started in Africa, and we ALL have common ancestors. Heck, we even have a common ancestor that also evolved into the great apes, and before that...

So a racist is an idiot if he thinks one 'race' isn't as good as another because of certain expressed traits.

Did you know that if someone with darker skin moves from a hot area to a cooler one, they can suffer because they don't get as much sun exposure? White folks can handle less sun better.

So we get adapted to each region we live in since people with certain traits will survive better in certain regions and pass those traits along. You can be a smart white guy, and still die of skin cancer before passing on your genes.

So again, I will say it does not matter if more black guys don't do as well on an IQ test as the white guys. They have other advantages. Those advantages are just as valuable.

I'm not saying anyone is wrong here, I've even agreed, but the original question is answered. Race doesn't matter when it comes to survival. We have to value differences. At any certain time those difference may be what keeps us from going extinct. One 'race' (or set of traits) is not better than another overall.

American

25th July 2004, 11:19 PM

Originally posted by Eos of the Eons
Oh bother, I was hoping you misunderstood, but not really. Thing is, we are using science, and science has shown that outter traits matter less when it comes down to who you are actually related to.

A really dumb person can be related to a really smart one. A black person can be related to a white one. When you hate one race or another, you are hating people you are/could be directly related to.

The dna test showed the caucasion in the class was more related to an american indian than another caucasion. That is science, not opinion.

Intelligence and skin color can cluster in areas because of the environmental stresses that push evolution one way or another. I have agreed with Pesta, showing why. Thing is, it doesn't matter. You can group people by skin color, and IQ test scores, etc. That doesn't make you any less related.

When you group people by actual inheritance, you find a mosaic, and you find that it is essential to have different traits in order to ensure survival.

So some folks don't do as well on IQ tests, but they'll whip your butt in a foot race. Or they will have less instances of skin cancer.

You are lumping people by traits. When it comes to genetics though, it's the group of traits that matters and who you are related to. The group of traits will matter when it comes to survival. Who you are related to doesn't really matter, but it shows you never know who you are more related to.

Point being, there is value to differences, and value to recognizing those differences are minor in the "big picture".

There is a story that I don't know is true. There was this racist who found out he had a disease. Turns out that only "black" people got the disease, so it turned out he was closely descended enough to a black person to have the trait expressed in his genome. He didn't look black, but he was racist against his own relatives.

Thing is, we do know that humanity started in Africa, and we ALL have common ancestors. Heck, we even have a common ancestor that also evolved into the great apes, and before that...

So a racist is an idiot if he thinks one 'race' isn't as good as another because of certain expressed traits.

Did you know that if someone with darker skin moves from a hot area to a cooler one, they can suffer because they don't get as much sun exposure? White folks can handle less sun better.

So we get adapted to each region we live in since people with certain traits will survive better in certain regions and pass those traits along. You can be a smart white guy, and still die of skin cancer before passing on your genes.

So again, I will say it does not matter if more black guys don't do as well on an IQ test as the white guys. They have other advantages. Those advantages are just as valuable.

I'm not saying anyone is wrong here, I've even agreed, but the original question is answered. Race doesn't matter when it comes to survival. We have to value differences. At any certain time those difference may be what keeps us from going extinct. One 'race' (or set of traits) is not better than another overall.

That's like saying I must like Vanilla Coke even though it's not REAL Coke, and if I don't then I'm a racist for preferring real regular Coke which is almost the same thing as the Vanilla flavor.

Look Jack- I don't have to like what I drink, and I ain't a racist because of it.

Eos of the Eons

25th July 2004, 11:30 PM

American, nobody is accusing you of being racist. I think you kinda get my point in a backwards way. Pop is pop, and people are gonna like different flavors. Doesn't make one flavor the best, and therefore 'superior'. Right? Variety is the spice of life, afterall. Just don't hate ginger ale for being so pale! My watergate cake wouldn't be the same without it!

American

25th July 2004, 11:36 PM

Originally posted by Eos of the Eons
American, nobody is accusing you of being racist. I think you kinda get my point in a backwards way. Pop is pop, and people are gonna like different flavors. Doesn't make one flavor the best, and therefore 'superior'. Right? Variety is the spice of life, afterall. Just don't hate ginger ale for being so pale! My watergate cake wouldn't be the same without it!

I'm just messing with ya.. I like all of those drinks. Ginger Ale's good too.

bpesta22

26th July 2004, 09:19 AM

Originally posted by Skeptic

(sigh)

This is one of the "factoids" that are often presented in favor of a "huge" difference between blacks and whites in IQ, but it shows more of the ignorance of those who promote it.

When a statistician says that a certain difference is "large" as opposed to "small", he means something like "a difference this size between populations is easily detected even with a relatively small sample from the two populations." It does NOT mean that the effect is absolutely "large" in terms of actual units (as opposed to SDs), let alone that whatever difference exists is of any practical importance.

For instance, suppose that every single white person had an IQ of exactly 100, and every single black person had an IQ of exactly 99. Mathematically speaking, the difference between the two is not 1, or 10, but literally an infinite number of standard deviations (since the SD of both populations is exactly 0). This is, indeed, a "huge" difference for a statistician, in the sense that there is an absolute, 1:1, correlation between race and IQ in this case: one's race totally determines what one's IQ is. So? It would still be still a tiny difference both absolutely (one IQ point) and practically (in terms of differences in "real life" performance).

The one SD difference in IQ, even if real--which is doubtful--might be "huge" in Cohen's statistical sense and still "tiny" or "insignificant" in any practical sense, in the same way that an EXTRA LARGE coke bottle can fit in a SMALL refrigirator. (mis)quoting Cohen in this way is merely using words from one context in another context, with the clear intent of making the IQ difference between blacks and white look as "important" as possible. [/B]

Skeptic.

I believe cohen's intent in suggesting those guidelines was to do precisely what you claim is ignorant-- to have some guideline for determining when an effect size has practical value / is large in the "sense" I've been talking about.

In fact, the effect size removes one big problem with p values-- they correlate with sample size. So, like your example above, you could have "highly significant" results, but the effect size is so small the results have NO practical value.

See cohen's example on the "significant" relationship between height and IQ in kids. This correlation really exists! But it has no practical value-- raising a kid's iq 30 points would require we increase his/her height by something like 12 feet!

Here are the facts (I'll back them with cites if you require) you tell me whether the black / white mean IQ difference has any practical significance:

1) Replicated since WW I, in literally 100's (if not 1000s) of studies, the black mean is around 85, the white mean is 100.

2) Since the standard deviation of an IQ test is 15, the effect size (d) is 1.0

Based on the effect size of 1.0, we know:

3) About 84% of blacks score below the white mean (100) on an IQ test. Only about 2% of blacks score 115 or higher on an IQ test.

4) Given the above effect size, pick any important variable that correlates with IQ. E.g., suppose we pick job performance, which correlates .5 with IQ.

Suppose we set the cutoff score for selection at 115, one standard deviation above the IQ grand mean.

We'd hire about 16% of the white applicants, but only about 2% of the black applicants. Extreme adverse impact.

Would it be legal? Yes, because the test is a valid predictor of job performance for both whites and blacks.

So, here you have a situation where the effect you imply is tiny (and that I am ignorant for suggesting otherwise) results in a bias such that 8 times as many whites (proportionally) are hired over blacks.

Yeah, this is a trivial effect.

Now, substitute job performance for graduate GPA (another correlation around .5). So, why is it we need affirmative action for minorities if we want them fairly represented in higher education?

Then substitute job performance for income (a smaller correlation of .30). 10% of the variance in how much money you make is determined by your IQ. Factor in the race difference and tell me an effect size of 1.0 is trivial.

Dancing David

26th July 2004, 10:37 AM

Originally posted by American
There's no should or should not about it, unless you define your motivations, which we haven't. The things you mention are a matter of opinion and individual values, not the realm of science.

All of this pre-supposes that today's psychology or sociology are valid sciences, which I say that they aren't, rather it's all political BS. So our arguing can go nowhere, and I will let you have the last word if you wish.

Some is science some is not, it's a lot like econmics and finance, some is research some is unsupported theory.

Beausoleil

26th July 2004, 04:07 PM

Originally posted by bpesta22
Line me up 10000 randomly selected americans. I will group them by eyeball according to race. Give them all a g loaded IQ test, and you will get these "race" differences we're talking about here.

I'd bet my life on it!

Given the other evidence in this thread about how mixed up genetic heritage is, all this would demonstrate is that the effect is not genetic.

You comment about being a predictor within minorities therefore differences between minorities are valid is a well known fallacy, isn't it?

If I organise two foot races, one normal and one with each participant carrying a sack of coal, physical fitness will be a good predictor of who wins each race, but of no use in explaining the average difference in times between the two events.

bpesta22

26th July 2004, 05:16 PM

Originally posted by Beausoleil
Given the other evidence in this thread about how mixed up genetic heritage is, all this would demonstrate is that the effect is not genetic.

You comment about being a predictor within minorities therefore differences between minorities are valid is a well known fallacy, isn't it?

If I organise two foot races, one normal and one with each participant carrying a sack of coal, physical fitness will be a good predictor of who wins each race, but of no use in explaining the average difference in times between the two events.

I dont quarrel with most of what you say here.....perhaps I wasnt clear.

I meant: if you look at just a sample of blacks, and use IQ to predict success, you get the same validity coefficient as in the white group.

There is no underprediction for blacks. The tests predict just as well for blacks as they do whites. So much so, the US courts ok use of the tests for selection, even though they discriminate against blacks.

Everyone has an opinion on this one way or another, which is fine-- but, i've asked this question a few times without any answer:

If IQ tests are such evil worthless instruments of racism, why does our court system let them be used for selection?

Even worse-- why has the concept of validity generalization for IQ tests been established in case law (the idea that the evidence for the validity of IQ is so overwhelming that a company today does NOT have to waste money validating it's IQ test)?

**
Re: the fallacy.

If I implied that because within groups, IQ is largely genetic, that therefore the between group difference is too, I didn't mean too.

If instead, you're arguing that the difference doesn't even exist, I'd recommend doing some research. It's not an opinion really. The difference is real. What causes it is what's being debated.

I personally believe the difference is biological. Whether genetic or environmentallly caused, I dunno.

But, that's my opinion, and we know what them are worth.

Eos of the Eons

26th July 2004, 06:36 PM

Sure they're biological, I think my posts have been arguing that all along.

The initial question was about grouping people based on certain traits and then making even further generalizations from there, that racists do. Racists don't have a leg to stand on. Sure, one group will show more of one trait, but racists don't realize how much the same we all are when it comes to where we came from in the first place. Not one trait is worth more when it comes to survival of the species. Racists argue the opposite. They also mix in paranoia and a few other disgusting ideas into it.

Intelligence is a trait of potential, and your offspring can evolve more intelligence, or less, depending on the environmental stressors. If things are easy, you'll find people with less intelligence surviving. Then if something happens that requires more intelligence to survive, you'll see a group getting more intelligent over time...a long time.

We also know intelligence can be affected by environment when a person is developing. A child exposed to alcohol in the womb can end up with FAS. Or malnutrition can affect it. Or maybe a child wasn't stimulated enough in certain ways when growing up. Disease can even affect the brain (encephalitus).

You also have to keep the individual in mind. There is that odd one that stands out, and we can then see why you can end up with a whole group standing out if those traits are desirable or beneficial. The odd one can be one that survives a disease, or has better vision, or does have more intelligence due to his dna. This all would be due to dna.

Racists don't appreciate that we need genetic diversity as a species. They are uneducated and close minded.

I think we've lost the thread starter here! I'm wondering if anything in this thread has helped him.