> I'm sure the writers of the SAT are well aware of this, but full IQ tests just don't scale.
First of all its generally faulty to presume what others are thinking.
Secondly, standardized admissions testing in its current form should be eliminated and there is a fair amount of data to support this[1][2]. If schools really want to discriminate on performance they should test for originality, composition, and decision capacity all of which can be done objectively with automated administration.
Sorry, but I write tests and I have absolutely no idea how you propose to measure these things objectively and automatically, without producing exactly the kind of "convergent" test you said you didn't like. For example, take "decision capacity". I don't even know what you mean by that, and I don't know how this will be assessed automatically without fixing a pre-designed correct answer.
Incidentally, the GRE has a system for grading essay composition automatically. It also absolutely sucks. It has no idea what your arguments are actually saying, so points are allocated based on irrelevant features like average sentence length or the total number of paragraphs. Researchers have succeeded in getting perfect scores by just copy-pasting the same sentence 25 times. In my book, this is worse than multiple choice.
For that, there's an "inbox test", a set of incoming messages to be dealt with in order. Such tests are widely used to screen candidates for manager jobs.
Wow, I just learned something new. These look legit and well-matched to the actual work. It does seem hard to make a "standard" decision test though, it seems each would have to be tuned to the job it was made for.
There is a lot of evidence that this is absolutely not the case. On of my links in a prior comment indicated there are now many universities that are standard test optional and keep stats comparing the performance between standard test students versus other students to find no difference in performance.
Consider your example. You are a famous developer well known for a performance oriented programming language. Of the many computer science students that graduated from MIT I cannot name any who are as well known for such. It is impossible to say that if your SAT scores were closer to perfect you might have gained admission. Regardless you have clearly excelled where others have not regardless of institution or institutional entry. At the least this suggests the incorrect combination of parameters were assessed given real world performance.
First of all its generally faulty to presume what others are thinking.
Secondly, standardized admissions testing in its current form should be eliminated and there is a fair amount of data to support this[1][2]. If schools really want to discriminate on performance they should test for originality, composition, and decision capacity all of which can be done objectively with automated administration.
[1] https://www.forbes.com/sites/kimelsesser/2019/12/11/lawsuit-...
[2] http://www.nea.org/home/73288.htm