• Quick note - the problem with Youtube videos not embedding on the forum appears to have been fixed, thanks to ZiprHead. If you do still see problems let me know.

Skin Color & Earning Studies=Woo?

Dave1001

Illuminator
D
Joined
Jul 21, 2006
Messages
3,704
I'm a bit skeptical about the scientific rigour of these studies. I think it's fair to say the study authors generally are expecting the answer that they find: that skin color correlates with earning power. I know of two studies thus far: this one on immigrants, and another one mentioned briefly in this article on african americans.

I don't know how the african american study measured skin tone, but this one claims that the original government research used an 11 point scale where researchers ranked the study participants' skin tone.

I'd like to know in more detail how this worked, because I see potential problems such as researchers being influenced by clothes study participants wore, their posture, or even more subtle things like time spent outdoors affecting the evaluations. I'd like to know how they controlled for all this.

Anyone know the link to the actual government study/survey explaining how they controlled for these factors?
 
Dave1001 said:
I'm a bit skeptical about the scientific rigour of these studies. I think it's fair to say the study authors generally are expecting the answer that they find: that skin color correlates with earning power. I know of two studies thus far: this one on immigrants, and another one mentioned briefly in this article on african americans.

I don't know how the african american study measured skin tone, but this one claims that the original government research used an 11 point scale where researchers ranked the study participants' skin tone.

I'd like to know in more detail how this worked, because I see potential problems such as researchers being influenced by clothes study participants wore, their posture, or even more subtle things like time spent outdoors affecting the evaluations. I'd like to know how they controlled for all this.

Anyone know the link to the actual government study/survey explaining how they controlled for these factors?
Click to expand...

what studies?
 
andyandy said:
what studies?
Click to expand...

I found this: http://www.msnbc.msn.com/id/16831909/

Joni Hersch, a law and economics professor at Vanderbilt University, looked at a government survey of 2,084 legal immigrants to the United States from around the world and found that those with the lightest skin earned an average of 8 percent to 15 percent more than similar immigrants with much darker skin.

{snip}

Hersch took into consideration other factors that could affect wages, such as English-language proficiency, education, occupation, race or country of origin, and found that skin tone still seemed to make a difference in earnings.

That means that if two similar immigrants from Bangladesh, for example, came to the United States at the same time, with the same occupation and ability to speak English, the lighter-skinned immigrant would make more money on average.

{snip}

"We estimate that dark- or medium-skinned blacks suffered a discriminatory penalty of anywhere from 10 percent to 15 percent relative to whites," he said. "This suggests people cue into appearance and draw inferences about capabilities and skills based on how they look."

Darity said it is not clear whether the bias is conscious or subconscious.

Hersch drew her data from a 2003 federal survey of nearly 8,600 new immigrants. The survey used an 11-point scale for measuring skin tone, in which 0 represents an absence of color and 10 the darkest possible skin tone.

From those nearly 8,600 participants, she focused on the more than 2,000 who were working and whose skin tone had been recorded during face-to-face interviews.

{snip}

Hersch said her findings, which will be presented at the annual meeting of the American Association for the Advancement of Science next month in San Francisco, could support discrimination lawsuits based not on race, but on color.

"There are very few color discrimination suits, but they are on the rise," she said. "But these suits can be hard to prove."
Click to expand...

These seem like pretty big conclusions based on a sampling of only 2000 immigrants. Of the 2000, we don't know how many were in each color group. I doubt they were evenly distributed.

As always, it's very hard to tell from a press release.
 
They didn't control for IQ, which correlates .33 with SES, and shows substantial "skin color" differences.

Seems like they're reifying "discrimination" as the causal mechanism for something clearly attributable to g.

I'm just waiting for them to report the same correlation to 6 decimal points two years from now...

:boxedin:
 
Indeed.

Arthur Jensen, 2005, Intelligence:

A large number of national and geographic population samples were used to test the hypothesis that the variation in mean values of skin color in the diverse populations are consistently correlated with the mean measured or estimated IQs of the various groups, as are some other physical variables, known as an ecological correlation. Straightforward statistical analyses clearly bear out the hypothesis, showing a significant positive ecological correlation between lightness of mean skin color and mean IQ across different populations.


*** Rushton (2000). Personality and Individual Differences:

We correlated mean IQ of 129 countries with per capita income, skin color, and winter and summer temperatures, conceptualizing skin color as a multigenerational reflection of climate. The highest correlations were - 0.92 (rho = - 0.91) for skin color, - 0.71 (rho = - 0.75) for mean high winter temperature, - 0.61 (rho = - 0.68) for mean low winter temperature, and 0.63 (rho = 0.74) for real gross domestic product per capita.
 
templer_22.gif
 
The above table uses aggregate data but it spans the whole range of skin color for the scale they used to measure it.

The correlation between skin color and IQ here (just looking at countries where IQ was actually calculated, versus estimated) is -.90!

At the aggregate level, IQ can be predicted very accurately just by knowing skin color using the following *linear* regression formula (I just calculated it):

Predicted IQ = 105.3 + -5.03(skin color).

IQ decreases linearly about 5 points for every change in skin color from pasty white to dark black.
 
bpesta22 said:
They didn't control for IQ, which correlates .33 with SES, and shows substantial "skin color" differences.

Seems like they're reifying "discrimination" as the causal mechanism for something clearly attributable to g.

I'm just waiting for them to report the same correlation to 6 decimal points two years from now...

:boxedin:
Click to expand...

I'm not sure that automatically applies for immigrant populations (for example Igbos in Taiwan) for obvious reasons. Also Hersh is claiming that people of the same immigrant ethnic groups, equally educated, have different earning outcomes based on skin color. Affirmative action wouldn't explain that, because it's based on ethnic group rather than skin color (a dark-skinned bangladeshi wouldn't have an affirmative action advantage over a light-skinned bangladeshi). I'm open to IQ explanations for social phenomena, but you seem to be jumping the gun here applying it so quickly to this case of skin color, immigration, and earnings.
 
Last edited:
Thanks for the info, bpesta22.

I'm sure I'll have more questions for you once I mull it over. ;)
 
Hmmm... I imagine any population's desk jockeys would spend considerably less time in the sun than their blue-collar countrymen. This alone could account for some or all of the pattern.

Or are they pulling their colour reading from where the sun don't shine?
 
bpesta22 said:
The above table uses aggregate data but it spans the whole range of skin color for the scale they used to measure it.

The correlation between skin color and IQ here (just looking at countries where IQ was actually calculated, versus estimated) is -.90!

At the aggregate level, IQ can be predicted very accurately just by knowing skin color using the following *linear* regression formula (I just calculated it):

Predicted IQ = 105.3 + -5.03(skin color).

IQ decreases linearly about 5 points for every change in skin color from pasty white to dark black.
Click to expand...

right, but there is a strong link between the education a child receives and IQ score attained - and so all the table shows is that the quality of universal education in countries in Africa is lower than the quality of universal education in Europe and the US. Which is hardly shocking.
 
andyandy said:
right, but there is a strong link between the education a child receives and IQ score attained - and so all the table shows is that the quality of universal education in countries in Africa is lower than the quality of universal education in Europe and the US. Which is hardly shocking.
Click to expand...

Why should the quality of education vary so nearly perfectly with skin color?

Plus, I guess it depends on your world view:

Education causes intelligence (Harvard makes people smart), or

Intelligence causes education (smart people go to harvard).

Just as two possibilities.

I could be jumping the gun, but given the strong link between skin color and IQ in the peer reviewed literature, not controlling for it (nor apparently even mentioning it as a possibility) is unforgivable, especially given how strongly they seem to argue it's discrimination causing the earnings difference.
 
bpesta22 said:
Why should the quality of education vary so nearly perfectly with skin color?

Plus, I guess it depends on your world view:

Education causes intelligence (Harvard makes people smart), or

Intelligence causes education (smart people go to harvard).

Just as two possibilities.

I could be jumping the gun, but given the strong link between skin color and IQ in the peer reviewed literature, not controlling for it (nor apparently even mentioning it as a possibility) is unforgivable, especially given how strongly they seem to argue it's discrimination causing the earnings difference.
Click to expand...

Please do this in a different thread, because you haven't answered how it plays into Hersh's specific claim about immigrants and intraethnic, intraeducation level earning spreads and skin color.
 
Dave

My intent wasn't to derail the thread-- sorry if you took it that way.

You asked if it was junk science. Not controlling for a variable that correlates .90 with skin color and has important implications for life success would seem to directly address the OP.

I'm not sure I understand the claim you want me to address re intraeducation. If it's important to you that I address it, could you rephrase. If not, feel free to ignore this.

B
 
bpesta22 said:
Dave

My intent wasn't to derail the thread-- sorry if you took it that way.

You asked if it was junk science. Not controlling for a variable that correlates .90 with skin color and has important implications for life success would seem to directly address the OP.

I'm not sure I understand the claim you want me to address re intraeducation. If it's important to you that I address it, could you rephrase. If not, feel free to ignore this.

B
Click to expand...

Hersh claimed that Bangladeshi immigrants with equal education levels had different earning levels: lighter skinned Bangladeshi immigrants earned more in America than darker skinned Bangladeshis. I though I have my doubts about the quality of this research and her conclusions from it, I don't see how what you posted about skin color and IQ is relevant to these specific claimed results.
 
andyandy said:
how's your testing going?

Still at the data collection stage?
Click to expand...

Please don't participate in a thread derail. We already have a thread or two for the general topics of IQ, heredity, subpopulations, etc. I see now that I had the misfortune of not naming this thread precisely enough, so as to enable BPesta to go on this tangent. :(
 
You must log in or register to reply here.

Back
Top Bottom