(This post was co-written with John Macintyre)
Summary: In this post, I am going to look at how well differences in socio-economic status can explain racial IQ differences. First, we will see that affluent Blacks score lower than poorer Whites and Asians on tests of cognitive ability. Second, it will be seen that the Black/White IQ gap is highest among the rich and lowest among the poor. Third, it will be shown that the Black/White SES gap has fallen dramatically over the last 100 years while the Black/White IQ gap has not. And finally, it will be seen that most of the Black/White IQ gap remains when holding SES constant. Over all, it will be concluded that SES probably explains only a small portion of the Black/White IQ gap.
- Poor(er) Non-Blacks vs Rich(er) Blacks
The simplest way to demonstrate the failure of SES to explain racial IQ differences is to compare black populations to poorer white populations.
Internationally, we can do this by comparing African Americans to various non-Black nations which are less wealthy than African Americans are. Using GDP Per Capita data from the World Bank and the Census, and IQ data from Lynn and Vanhanen 2012, we find that the following non-Black countries are poorer and smarter than African Americans:
The most extreme examine in this data-set is probably China, which has a mean IQ of 105.8 despite the fact that the average Chinese is only about one third as rich as the average African American.
A similar result is found in America. The SAT (or Standardized Assessment Test) is a highly g-loaded test that serves as a proxy for conventional IQ tests. One study found that if average SAT scores of high school students are broken down by their parent’s SES (for both Blacks and Whites), Whites from families who earn less than $20,000 a year do just as well on the SAT as Blacks from families that earn more than $200,000 a year. Other studies have found that Whites from severely impoverished families who earn under $10,000 a year tend to have higher SAT scores than Blacks who earn over $100,000 a year.
Less consistent are the results of Loehlin, Lindzey, and Spuhler 1975 who analyzed 13 studies and found that 8, or 61%, showed low SES Whites outperforming high SES Blacks on IQ tests. On the whole though, the evidence is clear: poor Whites often out-preform even rich Blacks on intelligence tests.
- Racial IQ Gaps Across SES Levels
On top of this, there is some good reason to think that Blacks may be less able than Whites to take advantage of the cognitive opportunities that a high SES environment provides. Specifically, the richer people you compare, the larger that Black-White IQ gap becomes.
This was shown by Shuey 1966 (page 519) who analyzed 32 studies and found a mean Black-White IQ gap of 12 points between people of low SES and 20.3 points between people of high SES. Similarly, Loehlin, Lndsey and Spuhler 1975 analyzed 7 studies not included in Shuey 1966 and of them 5, or 71%, showed a larger Black-White IQ gap among high SES people than among low SES people. Herrnstein and Murray 1994 (page 288) found the same thing in their analysis of NLSY data. The same type of finding is reported in Jensen 1998 (page 358). Finally, the previously referenced Human Varieties paper found a positive interaction between SES and the Black-White IQ gap across several large and nationally representative samples ranging from the 1970’s to 2009. Thus, across 48 studies 46, or 96%, found that the Black-White IQ gap rises with SES. The totality of the evidence is clear then, Whites and Blacks who grow up in high SES homes differ more in IQ than Blacks and Whites who grow up in low SES homes.
Source: Herrnstein and Murray 1994 page 288
Source: Jensen 1998 page 358
That Blacks benefit less from a high SES environment is also suggested by the fact that SES correlates with IQ less in Black samples than in White samples. This was found in the Lohelin, Lindzer, and Spuhler analysis. Of the 13 studies, 11 found that the IQ/SES correlation was higher for Whites than Blacks, 1 found no difference, and 1 found that it was higher for Blacks than Whites. Thus, higher SES seems to lead to less cognitive improvement among blacks.
- Racial Difference in Income and IQ Over Time
Another way to test the relationship between SES and racial differences in IQ is to look at how the two have changed over the time. The U.S. Census began collecting information on individuals income in 1939. Since then, Black wages relative to White wages have risen from 44% to 67% in 1989 (Maloney). Similarly, according to data from the US Census Bureau, between the 1950’s and 2014, there was a dramatic reduction in the size of the Black White income gap. The median income of Whites went from 2.4 times larger than the median Black income during the 1950’s, to only around 1.3 times larger in 2014. Rates of post-secondary educational attainment revealed the exact same pattern. In 1940, Whites were nearly 4 time more likely than Blacks to have a Bachelor’s degree or higher. However, as of 2014, Whites were only 1.5 times more likely than Blacks to have a BA or higher. All and all, the historical trends reveal huge reductions in the education and income gaps between Blacks and Whites (due to Affirmative action).
During this time, the Black-White IQ gap hasn’t budged much. IQ testing during the first world war gave Blacks a mean IQ of 83 (Shuey 1966 page 317). As reviewed on this sites page on U.S. racial IQ gaps, today Black Americans have a mean IQ of 85. Thus, the Black-White IQ gap has been reduced by 2 points, or 11%, during a time period in which the Black-White socio-economic gap has been reduced by at least 50%.
- The B/W IQ Gap Holding SES Constant
Finally, the most obvious way to test the effect of SES on racial IQ differences is to compare the IQ scores of members of different races who have the same parental SES. Now, doing this requires a certain amount of prudence in interpretation. If IQ variation causes SES variation then it may be that Blacks from rich/poor families have parents with higher/lower IQs than average and have inherited said IQ from them. Thus, when we compare Blacks and Whites with the same SES we are, to some degree, comparing Blacks and Whites with genotypic intelligence held constant. With that said:
Shuey 1966 (page 518-519) reviews 42 studies carried out between 1913 and 1966 in which Blacks and Whites of equal SES were compared on IQ tests. Of the 42, 40, or 95%, found that Whites still had higher IQs than blacks. 2 found that no significant racial difference persisted after SES was controlled for and none found that Blacks were more intelligent than Whites. 33 of these studies reported their sample size and mean IQ score by race. The N-weighted mean IQ difference between Blacks and Whites across these 33 studies was 11 points after holding SES constant.
Loehlin, Lindzey, and Spuhler 1975 (page 172) reviewed 7 studies done between Shuey’s work and 1973 and all 7 showed that Whites had higher IQs than blacks even when comparing groups of the same SES. (The average difference is not given.)
Jensen 1974 reported the same finding.
This result was replicated again by Herrnstein and Murray 1994 who found that controlling for SES reduced the black/white IQ gap from 18 points to 12 points. An author at Human Varieties replicated this result yet again using 11 data sets collected between the early 1970’s and 2009. An analysis of NAEP data collected between 2002 and 2007 showed the same thing once again (Vanhanneman et al 2009)
To test this finding at the national level, I utilized IQ data from Lynn and Vanhanen 2012 and 2014 GDP per capita data from the World Bank. I created a dummy variable to represent whether 116 nations were mostly populated by Whites or Blacks. (Non-Black and non-White nations were excluded from the analysis.) The un-adjusted IQ gap between White and Black nations was 22 points. Controlling for GDP per capita reduced the gap to 18 points, a 4 point reduction. Whether or not a country was Black or White was also found to be a better predictor of its IQ than its GDP per capita was.
In total then, we have 63 studies carried out between 1913 and 2014 (101 years), and nearly all, 61, or 97%, show that Blacks have lower IQs then Whites even after controlling for SES.
The fact that these studies controlled for SES in a variety of ways (some compared real groups of people who were matched for parental income, or education, or neighborhood, .etc, while others statistically adjusted for SES related differences) and measured SES in different ways (some used parental education, income, occupation status in a combination, measures of home quality, etc.) affords us strong confidence in these findings. The gap that remains after controlling for SES seems to be 3 to 4 points less than the un-adjusted gap, suggesting that, at most, roughly a fifth to a quarter of racial IQ gaps can be casually explained by SES. The facts concerning genetics, IQ, and SES, suggest that the actual casual impact is substantially less than that.