Friday, December 09, 2011

Skin color and occupational prestige among blacks

As a follow-up to the last post, I looked to see if skin color among blacks is related to occupational prestige. Interviewers rated skin color from very dark brown to very light brown. Here are the mean prestige scores (GSS data, sample size = 454):


Mean occupational prestige scores

Very dark brown 28.3
Dark brown 28.4
Medium brown 35.3*
Light brown 33.2
Very light brown 42.7*

*significantly higher than the very dark brown group

Lighter groups tend to have better jobs than darker groups. This is consistent with the common observation that elite blacks often look like they have more European ancestry.

7 comments:

bgc said...

I wonder if there is any evidence of a greater-than expected income/ job status/ educational attainment among those of mixed race?

The prediction arises from the fact that such individuals would have all the benefits of affirmative action, plus the benefits of the racial mixture (e.g. in 'g').

The advantage given by AA seems to be of the order of 1 standard deviation above the educational attainment (e.g. a student with an IQ of 115 gets into a college where the usual requirement is 130).

Those of mixed African and European parentage might expect to have an average IQ of - say - half an SD below European average - yet would get an AA advantage of a full SD compared with Europeans. Net advantage should be about half an SD - equivalent to about IQ 108

This might predict hat those of mixed race should actually do *better* than Europeans on average where AA is operating.

Chuck11 said...

”I wonder if there is any evidence of a greater-than expected income/ job status/ educational attainment among those of mixed race?“

bgc,

There's plenty. And the explanation offered is "colorism." Look up: Hochschild and Weaver (2008) "The Skin Color Paradox and the American Racial Order"; Harris (2008) "From color line to color chart?: Racism and colorism in the new century." "Colorism" is often cited as proof that disparities are due to discrimination -- since cultural explanations can't account well for the results.

Ron,

Could you provide SDs? With them we can see how much IQ theoretically could account for. In the African AMerican populations the color-ancestry correlation is only 0.45. And there will be range restriction, maybe Blacks range from .9 to 0.3 % African. So if there is a 1.2 SD difference in g between hypothetically pure AAs and pure Whites -- the maximum g difference would be 1.2 x 0.45 x .06. The maximum occupational difference explainable would be, 0.7 x that since that's the g occupation correlation. So maybe .23 SD in prestige between the upper and lower end could be accounted for by individual racial differences in IQ. It's more complex because there's a population level effect that no one ever discusses, except R. Gordon. More African Blacks are more likely to live in Black areas, e.g. Detroit, which amplify IQ effects --

Chuck11 said...

Ron,

So as indexed by skin color the genetic difference in white admixture between the most and least caucasian Blacks is maybe .6 x .45. or 30%. If the difference in genotypic IQ between Whites and Blacks, who on average are 25% caucasian, is 1 SD, the difference between pure African Americans -- which is not to say Africans because the former did not exactly hail from a representative sample of the latter -- and Whites would be 1.3 SD. The difference between the most and least admixed blacks as indexed by color, then, would then be .4SD. The predicted difference in "income/ job status/ educational " would then be .4 x whatever the correlation is.

bgc

"Net advantage should be about half an SD - equivalent to about IQ 108"

Do a google scholar search for "net Black advantage." This, along with the mechanism by which "colorism" works, is another one of those unsolved sociological mysteries.

Chuck11 said...

In the NLSF data, there's both skin color and SAT data. You can get the data online. In an unpublished paper "The Effect of Skin Color __ on Wages and __ Employment for Black Males," the authors argue that IQ can only explain 25% of the intrarace "color" disparity.

SFG said...

Hmmm...any way control for IQ here? Lighter skinned blacks are going to have more white admixture and hence, likely, higher IQs.

chuck11 said...

"Hmmm...any way control for IQ here? Lighter skinned blacks are going to have more white admixture and hence, likely, higher IQs."

I just posted on the IQ-color association in the Add Health data. The upper and lower ends differ by .5 SD.

Annempli said...

Ron, So as indexed by skin color the genetic difference in white admixture between the most and least caucasian Blacks is maybe .6 x .45. or 30%. If the difference in genotypic IQ between Whites and Blacks, who on average are 25% caucasian, is 1 SD, the difference between pure African Americans -- which is not to say Africans because the former did not exactly hail from a representative sample of the latter -- and Whites would be 1.3 SD. The difference between the most and least admixed blacks as indexed by color, then, would then be .4SD. The predicted difference in "income/ job status/ educational " would then be .4 x whatever the correlation is. bgc "Net advantage should be about half an SD - equivalent to about IQ 108" Do a google scholar search for "net Black advantage." This, along with the mechanism by which "colorism" works, is another one of those unsolved sociological mysteries.