I'm annoyed this paper needed to be written. This debate just won't die, no matter how poor the proclaimed evidence is. I suspect this is because there's a concerted effort to maintain the debate in popular discourse. See: https://kevinabird.github.io/2019/12/18/The-Genetic-Hypothesis-and-Scientific-Racism.html …
-
-
Prikaži ovu nit
-
Here's a brief summary: Hereditarians make two major claims about Black and white IQ gap. 1. The gap in intelligence and educational attainment we see is substantially caused by genetic differences and 2. These genetic differences are the result of natural selection.
Prikaži ovu nit -
Historically these arguments have relied on indirect evidence based on heritability estimates (criticized here http://www.acteon.webs.upv.es/ARTICULOS/KEMPTHORNE-LOGICAL_EPISTEM-_1978-_BIOMETRICS.pdf …) life-history theory (criticized here http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.731.3826&rep=rep1&type=pdf …) or comparing national IQ scores (criticized here http://emilkirkegaard.dk/en/wp-content/uploads/Why-national-IQs-do-not-support-evolutionary-theories-of-intelligence.pdf …)
Prikaži ovu nit -
Recently, however, the wealth of genome-wide association studies led to some determined (and inexperienced) race scientists to use SNPs associated with educational attainment to prove genes caused the racial achievement gap. They got results like this and rejoicedpic.twitter.com/YcgfP0bYP2
Prikaži ovu nit -
Unfortunately for them, we quickly found out you can't just compare SNP frequencies between populations because systematic biases are introduced based on what population in which the SNPs were identified. (see https://www.nature.com/articles/s41467-019-11112-0 …)pic.twitter.com/7tNtkbCbL1
Prikaži ovu nit -
These biases have been shown to cause false signals of selection, even in traits we have spent a lot of time studying, like height (see https://elifesciences.org/articles/39725 & https://elifesciences.org/articles/39702 ) one solution is to use less biased effect sizes based on within-family studies.
Prikaži ovu nit -
These concerns were not taken into account by the people who wanted to show black people were genetically less intelligent, so I took it upon myself to show the results they found were caused not by actual evolutionary patterns but by systematic biases in GWAS data.
Prikaži ovu nit -
I used two approaches: first I compared the difference in polygenic scores of African & European populations from the 1000 genomes project to 10000 comparable randomly generated polygenic scores. When this method is applied to skin pigmentation it looks like thispic.twitter.com/GOfODnMzHS
Prikaži ovu nit -
When I used this method from the a GWAS on educational attainment using between family effect sizes I saw signals of divergent selection, like hereditarians claim, but when I used within-family effect sizes the signal completely disappearedpic.twitter.com/g2ob1tGIsQ
Prikaži ovu nit -
Next I used the method from https://www.nature.com/articles/s41467-018-04191-y … which compares genetic differentiation of populations (measured as Fst) between SNPs associated with a trait and 10,000 sets of SNPs not associated with a trait but similar in other aspects. Here's what SNPs height looks likepic.twitter.com/BO9hzW1oq2
Prikaži ovu nit -
When I used the GWAS results from an educational attainment and cognitive ability study I found this pattern which is not suggestive of natural selection. It also shows that SNPs related to these traits are not very different in frequency between Africans and Europeanpic.twitter.com/7U3jTcovhE
Prikaži ovu nit -
One thing I can do with that knowledge is compare how much genes related cognitive ability vary between groups & how much IQ varies between groups. Using national IQ data sets from hereditarians, I find that <15% of variation in IQ scores can be explained by genetic variation.
Prikaži ovu nit -
Using the most reliable methods right now I show no signs of divergent natural selection acting on intelligence or cognitive ability and that genetic variation explains a very small portion of IQ score variation. When you hear about Murray's book keep this all in mind!
Prikaži ovu nit -
I owe a lot of people thanks for this paper: but in particular thanks to
@JeremyJBerg,@DocEdge85, &@EvoRoseman for very helpful feedback on analyses &@AysuOkbay who graciously provided within-family data from the EA3 study. Look for it in a peer-reviewed journal soon!Prikaži ovu nit
Kraj razgovora
Novi razgovor -
Čini se da učitavanje traje već neko vrijeme.
Twitter je možda preopterećen ili ima kratkotrajnih poteškoća u radu. Pokušajte ponovno ili potražite dodatne informacije u odjeljku Status Twittera.