They're objective measures (don't depend on the person applying them), created subjectively (people choosing criteria based on their preferences/ideas), and chosen subjectively (people deciding which ones they want to rely on more). I meant objective originally, as in many people claim the whole process is objective.

They're obviously objective. That doesn't make them good. Think of Harvard's goals, maybe some more complicated version of a combination of:

1. Meritocracy: Give a chance to the students with the best innate chance at real-world success

2. Self-preservation: Give a chance to the students with the best chance at real-world success

3. (implicitly) Don't let too many people in who don't further (1) or (2).

Those measures (SAT, high-school GPA, gender, color, income, ...) are weak predictors of (2). How couldn't they be? We live in a world that encourages feedback loops, making it difficult for the most intelligent and ambitious people to break through class barriers with any reasonable degree of success.

They're not good measures because (a) they're not even particularly strong predictors of goal (2), and (b) they're piss-poor predictors of goal (1).

By way of contrast, a compelling essay is much harder to assign an "objective" score to, but it's a stronger predictor of both (1) and (2) than the rest put together, especially at the top end.

The important thing to keep in mind with all of those though is that they're proxy measurements. We can't directly measure the future, so we come up with tests to try to guess less incorrectly. It doesn't matter which measure you pick; they'll all be "problematic." If you recognize that though, it's easier to move past a shallow thought like whether the measure is objective or not and toward a system that better align's with the university's goals.