Quote (Thor123422 @ 19 Nov 2017 22:12)
Thing is, a difference in GPA like that does not speak at all to your potential.
If it was a 2.9 versus 3.9, I'd see your point, but if you understand GPA as a predictor of success then you know this isn't the case for the situation I outlined for you.
In this case it is more likely that the 3.5 student simply didn't have as much time to dedicate to assignments, not that they came away with a lesser understanding of the material.
Quote (Thor123422 @ 19 Nov 2017 22:17)
Here's an interesting example.
I recently spoke to my department head and the concept of the GRE came up for foreign students.
What he told me was interesting. Basically, he can't trust any Chinese students GRE scores as predictors of their success in the program. The reason is that Chinese students have huge prep courses that lets them score highly on the GRE, so they all score near the top.
Nigerian students score MUCH lower, but as long as he picks the ones that are high for the region, they ALWAYS turn into much better students than the Chinese students regardless of the deficit in their GRE scores.
Raw scores aren't a predictor of success as much as relative scores from the same group.
well, ok, gpas and gre and stuff like that is maybe a bad surrogate for the much wider concept of "performance". it is of course difficult to operationalize "performance", and I have no issue with weighing in both absolute and relative performance. but I still stand by my point that performance, however we operationalize it, should be the only criterion for the hiring for merit-based positions. phd programs, positions in a law firm, hedge fund managers, and so on. soft skills ("does he fit into our team") will of course play a larger role the further one goes down the job pyramid.
also note that your example with the "3 SD above their country average nigerians" being better than the "1.5 SD above their country average chinese students", despite the chinese scoring higher absolute GREs, might not translate to other fields as easily. the more a position is about hard work and experience and the less about raw talent and top-end intelligence, the more shaky this concept becomes. when hiring for an academic position in a STEM field, you really want people with top-end intelligence, and it sounds very plausible to me that the relative performance of applicants compared to the rest of their country might be a better indicator of this trait than the absolute value of a certain test score.