The College Board just released this year's SAT scores!
I thought I'd go ahead and put everything in familiar terms and make some plots.
This thread will have a lot of pictures. First up: How did everyone do?
All of the typical race differences are there. Blacks did roughly 15 IQ points worse, Hispanics did about 10 points worse, Asians did similarly better, etc.
If we scale all that by the sizes of the populations who took the tests, we get this:
Another way to look at this data is to stack everyone into a single population, like so:
Alternatively, we can get a picture of tomorrow's elites and underclasses by looking at deciles.
The top decile is almost as Asian as it is White! Combined with limited local university prefs, this helps to explain why Asians have been displacing Whites at elite universities.
Now, you might wonder: Isn't the national data plagued by representativeness issues because of selection into testing?
Yes! But not in Michigan. In that state, practically everyone takes the test, so the results are more representative of its population.
In Michigan, the results are all quite similar, except Asians have noticeably higher variance.
The decile composition of Michigan is also far less Asian at the top, because Michigan is less Asian than the national sample and it has more Whites.
There are other states with representative sampling of their local populations, and D.C. is in that group too. There are also some states where practically everyone takes the ACT, and we definitely need to look at those results as well.
What about sex differences? Well those are always important, and they replicate the differences seen in other years!
Beyond the variance differences, we see that girls outperform boys in Michigan (representatively sampled) while underperforming nationally (selectively sampled).
This research directly militates against modern blood libel.
If people knew, for example, that Black and White men earned the same amounts on average at the same IQs, they would likely be a lot less convinced by basically-false discrimination narratives blaming Whites.
Add in that the intelligence differences cannot be explained by discrimination—because there *is* measurement invariance—and these sorts of findings are incredibly damning for discrimination-based narratives of racial inequality.
So, said findings must be condemned, proscribed.
The above chart is from the NLSY '79, but it replicates in plenty of other datasets, because it is broadly true.
For example, here are three independent replications:
A lot of the major pieces of civil rights legislation were passed by White elites who were upset at the violence generated by the Great Migration and the riots.
Because of his association with this violence, most people at the time came to dislike MLK.
It's only *after* his death, and with his public beatification that he's come to enjoy a good reputation.
This comic from 1967 is a much better summation of how the public viewed him than what people are generally taught today.
And yes, he was viewed better by Blacks than by Whites.
But remember, at the time, Whites were almost nine-tenths of the population.
Near his death, Whites were maybe one-quarter favorable to MLK, and most of that favorability was weak.
The researcher who put together these numbers was investigated and almost charged with a crime for bringing these numbers to light when she hadn't received permission.
Greater Male Variability rarely makes for an adequate explanation of sex differences in performance.
One exception may be the number of papers published by academics.
If you remove the top 7.5% of men, there's no longer a gap!
The disciplines covered here were ones with relatively equal sex ratios: Education, Nursing & Caring Science, Psychology, Public Health, Sociology, and Social Work.
Because these are stats on professors, this means that if there's greater male variability, it's mostly right-tail
Despite this, the very highest-performing women actually outperformed the very highest-performing men on average, albeit slightly.
The percentiles in this image are for the combined group, so these findings coexist for composition reasons.