The Study of Racialism

kpauljohnson

http://www.dnatribes.com/dnatribes-digest-2008-08-29.pdf

I had asked before if any such comparative information had been published, and apparently this is the first time statistics like these have been available. Am fascinated to find that 6% NA admixture is typical of white Virginians, in light of my own results. As more people are tested, data like these will be more meaningful, but for now they provide some intriguing food for thought. See page five for a summary chart of the ten states reported.

fwsweet · Posted: Mon 29 Sep 2008 00:42 Post subject:

Outstanding! Thank you very much for posting this. I have archived the report as Who Are The Americans?, so that we have a permanent copy, even if the above link gets disconnected. The numbers reported confirm most of the estimates currently in use, especially the admixtures of African Americans and of Lumbees. There are two surprises, though:

Surprise 1. The average subsaharan admixture found in U.S. Whites (2 percent) is triple what Shriver and Kittles measured in their original study (0.7 percent). This is surprising, but it makes sense because the original study's samples were all from the same region (southern Pennsylvania), which might have been anomalously low.

Surprise 2. The 4.7 percent subsaharan admixture found in White folks in Minnesota is also suprising. It approaches the numbers measured for the Melungeons and Redbones (about 6 percent).

I make two cautionary observations for site members:

Caution 1. The findings have not been peer-reviewed and so I would take them with some skepticism, were it not for the fact that most of their numbers match those of prior studies.

Caution 2. It is a terrible shame that they did not reveal how many people were in each of their samples. As we have discussed in other threads, sample size is vital to computing confidence interval. If the numbers reflect a hundred individuals in each state, then you can take them to the bank. But if they reflect less than a dozen in each state, they are meaningless.

kpauljohnson · Posted: Mon 29 Sep 2008 19:43 Post subject: Weird indications of small samples

While the overall differences between states usually make sense, two things jump out at me as being completely implausible in a large sample. First is the very low Mediterranean component in New York whites compared to Virginia, a state that has had far less immigration from southern Europe. Second is the 22% Middle Eastern average for Michigan, which admittedly is home to the largest Arab-American population of any state. But those Michigan Arabs must be way overrepresented in the sample as their numbers are not large. It occurs to me that white Virginians taking the test are often motivated, like me, by oral traditions about Indians. So self-selection probably skews the Amerindian numbers here and in NC; people who have reason to suspect NA ancestors are more likely to take the test.

fwsweet · Posted: Wed 01 Oct 2008 16:28 Post subject:

I did not understand that their samples were simply people who had bought DNA tests. For some reason, I thought that the subjects were chosen randomly. Given the real situation, I can now understand why the results might be weird. People who have reason to suspect that they have interesting ancestry would be more likely to buy the test.