Average GRE scores of admitted students, by discipline (Duke)

[chateauheart]

One of the first things that newcomers to this forum notice is how staggeringly difficult it seems to be to get into an economics PhD. There is now some evidence of that.

 

The data are from the last three years. Note that economics at Duke is not particularly high ranked compared to their other departments.

[newton108]

May be that's the why Duke is not particularly high ranked in econ.

[AkiraHatake]

@newton108: Do you mean they are focusing too much in GRE?

 

The battle is crazy now

[thagzone]

Well, Duke is particularly highly ranked in economics, just not relative to the rest of its departments. It is a T20 program.

 

This is pretty common among all T30 departments. The median combined GRE for the incoming class at my program (T20ish) was around 330 as well. Go browse scores for people reporting acceptances to T30 departments on Gradcafe. They're pretty much around the 330 range, and higher on average for native English speakers who tend to do much better on the verbal portion of the test.

[chateauheart]

It's also important to note that we basically don't select for GRE verbal and we don't really care about GRE quant either. But the median GRE of econ students is still significantly higher than several disciplines that care about both (sociology, psychology), and also slightly higher than political science.

 

The very high scores of admitted students despite of its low weight in the application process seem to be a good indicator of the unusual level of intelligence of incoming economics PhDs across all levels of programs.

 

This is also true of mathematics, where GRE verbal is similarly irrelevant and yet incoming math PhDs tend to have higher verbal scores than almost all social science and humanities programs.

[tm_member]

It's also important to note that we basically don't select for GRE verbal and we don't really care about GRE quant either. But the median GRE of econ students is still significantly higher than several disciplines that care about both (sociology, psychology), and also slightly higher than political science.

 

The very high scores of admitted students despite of its low weight in the application process seem to be a good indicator of the unusual level of intelligence of incoming economics PhDs across all levels of programs.

 

This is also true of mathematics, where GRE verbal is similarly irrelevant and yet incoming math PhDs tend to have higher verbal scores than almost all social science and humanities programs.

 

Alternatively, you could view GRE scores as being extremely important but now a useless signal due to right tail censoring. Maybe it only looks like the GRE doesn't predict admission cause everyone has high scores already?

[chateauheart]

In theory, there's probably some of both substitution and right tail censoring. I think the former is a much bigger effect.

 

One interesting fact is that, at least a few years ago, the GRE scores of incoming students to second-tier econ programs were consistently higher than that of first-tier econ programs. This points to substitution: admits to first-tier econ PhDs usually have such a compelling profile that their test scores are a noisier signal than the rest of their profile, so it is discarded, and thus ex-post GRE scores among incoming students have high variance. Second-tier econ programs tend to take a lot of students with more variance in their potential (e.g. international students), so the GRE is comparatively the less noisy signal, and there is more selection on GRE scores.

 

If it's only right-tail censoring, then the GRE scores of top-tier students should still be slightly higher than that of second-tier students, since the median is nowhere the max (post-reform).