What school districts/states need to do is give all students a better shot by holding prep sessions all throughout high school, and offer free or low-cost individual prep tutoring.
We didn’t even need a study for this. It has been obvious for years with all of the rampant grade inflation in the schools. These days every kid is above average!
But... that's not what it purports to measure, and there are other explanations for why a high-scoring student might have more success in life (e.g. getting accepted to a more prestigious school, regardless of performance, which doesn't seem to have been controlled for).
That is not what that says. Those graphs show that GPA is not correlated with that long-term outcome and that SAT scores are only weakly correlated to that outcome, given those controls.
r = 0.07 is definitely not effective, well-designed and informative.
The two ways to do well on the SAT are having tremendous ability or working your ass off in preparation. In high school, you can cheat, cajole, beg, etc.
But perhaps the more useful metric will be, not the relation of SAT and income, but how the SAT is related to the ADDED value of a college degree? I.e., a correlation between SAT and a variable [(average income of people with and SAT X)-(the same without college)]
The idea that eliminating standardized tests to improve equity in admissions (while other adm. criteria like HS GPA, essays, and extracurriculars have even bigger equity issues) has always seemed a little half-baked
Really weird analysis frames. Why 1300 SAT at the bottom? Why 3.3 GPA at the bottom?
Does all correlation break down below those points? Is the researcher hiding that? Clearly they couldn't get away with the pretty 3.6 GPA as the lower bound, but 3.3 GPA is just as absurd.
