Back in the dark ages of the 80s when I was taking the SAT and ACT, these tests were considered good predictors for the first semester's performance.

That's it.

I did well on both tests and did well on my first semester. It's the semesters after where my performance tanked because I didn't have some of the work habits that solid B students had. (I will also be clear and say that at least some of the problem is attributable to the university and how it handled advisors. My advisor was completely useless and let me schedule for _way_ too much hard stuff.)

There was also a really good predictor of how one would do on the SAT or ACT: NoBitH. The Number of Bathrooms in the House.

Yup. The SAT and ACT at the time were better measures of economic advantage than innate intelligence. I have no reason to believe that this isn't still the case, especially since they're more entrenched in the system than ever.

Well, sure, but first semester performance is also a good predictor of second semester performance. And second semester of third, and so on.

But more to the point: If you do poorly in your first semester and drop out, then it doesn’t really matter if the SAT would have done a good job of predicting your second semester performance.