Comment by rsj_hn | Hacker News Reader

Comment by rsj_hn | Hacker News Reader

Parent Full thread

Waste of time, huh? You have differences in IQ between black, hispanic, white, and asian americans leading to very big differences on, say, standardized math and verbal tests. Now, the teachers are accused of racism, because their black students are scoring much lower than their asian students. Then they are judged unfairly and put into the impossible situation where they are expected to graduate students with no test score differences between races, because no one is willing to admit that there are IQ differences between races, which means a series of weird evasions where schools have to stop administering objective tests, or hide the results of the tests from the public, or go through ritual handwringing about "systemic racism", "unconscious racism", "undetectable racism", because these racial differences remain across time, cultures, and even nations, and seem stubbornly resistant to any intervention other than one involving no longer collecting the data, or no longer reporting the data.

Or, we can just accept that there are differences in IQ which should lead to differences in test scores and normalize correctly. That would be useful. It would not put teachers into impossible situations. We would not be tying teacher pay to impossible to achieve goals.

But no, we can't honestly approach this subject and must constantly contort everything because honest baselines and metrics would be a "waste of time" over the unrealistic and impossible to obtain outcomes.

Same with incarceration rates, and some other group differences. Oh, wouldn't you like to know which group differences are biological and which are due to the shared environment, so that we know which interventions have a chance of working and which don't? No, we can't know that, as obtaining this information would be a waste of time. Honesty would also stop enormous wastes of human capital being currently spent on simultaneously trying to cover up differences in outcomes as well as punishing people arbitrarily when they notice said differences.

https://www.nature.com/articles/s41588-018-0147-3

tl;dr for those who didn't read: they looked at genes (SNP's really) instead of "race" and found... a 7-10% correlation with cognitive performance (that is, 7-10% of the difference in individuals scores is determined by the differences in the genome subset studied).

Yet you look at "race" (a social construct that itself correlates at best with about 50% actual genetics) and see a much larger effect. That by itself should immediately put this "black people are dumb" notion to bed (i.e. even if it's measuring something, it's clearly not measuring heritable diferences!). But instead everyone just doubles down on the nonsense, because the science is "polluted by cancel culture" or whatever.

And that's what this opinion piece is. It's an attempt by the political right (the Journal's editorial page is increasingly skewed these days -- most papers at least pretend to have a diversity of viewpoints) to explain away the science they don't want to be true by attacking "science" as a field.

Wow, that's misleading. They found specific 10 SNPs (single nucleotide polymorphisms) which they identified as relevant to intelligence and for these 10, a 7-10% effect. This type of research is very useful for trying to identify specific genes responsible for X but not for the effect of all genes on X.

The generally accepted range for genetic influences is much higher -- say 50% to 2/3 although we know little as to which genes do what when it comes to something like intelligence.

They identified 10 specific related SNPs on the X chromosome as part of a side study. The study itself was of 1271 sites. That number is right there in the first sentence of the abstract! And the summary sentence is awfully clear: "A joint (multi-phenotype) analysis of educational attainment and three related cognitive phenotypes generates polygenic scores that explain 11–13% of the variance in educational attainment and 7–10% of the variance in cognitive performance".

Who's being misleading here? That "generally accepted range" you're trying to invoke is precisely was is being studied here. And it was wrong, largely because the people generally accepting it were confusing social factors with genetic ones.

all genes on X is a meaningless statement unless all genes are involved in X... if you read the paper the authors seemingly put quite a bit of effort into the rigor of their model.

they used multiple validation cohorts, they controlled for environmental factors, they controlled for non-autosomal factors, applied multiple methods to boost the predictive performance of their model. adding more snps typically leads to an overfit and poor performing model on validation cohorts, which is why they use so much rigor and end up with usually 100-400 snps for each of the phenotypes.

in anycase they found a lot more than 10 SNPs, and they did more than one analysis in the paper.

Parent educational status is a significantly better predictor than the polygenic risk score for educational attainment. heritability (also known as r^2) for cognitive performance they estimated at an even lower ~7-10%.

Reply on news.ycombinator.com