I wouldn’t be surprised if the two results are equivalent by the naive measure: the fraction of Putnam participants who get honorable mention or better might well be about equal to the fraction of GRE test takers who get a score of 1500 or greater.
The real issue here, of course, is self-selection. Lots of people take standardized tests because they want to apply somewhere that requires the score, which drives the GRE average down. On the other hand, most people who take the Putnam do so because they enjoy the challenge, and are more likely than average to be good at math. This drives the Putnam average up, compared to what it would be if every college student in the US took it.
Edit: I don’t have combined statistics for the GRE, but between 5% and 6% of test-takers got 800 on the math section when I took it in 2010, while I estimate 700+ on the verbal section was about as likely. Obviously the two are correlated, but it seems reasonable to suspect that the combined percentage was less than the 2% of Putnam contestants who received an honorable mention or better.
I think the real issue is that comparing the GRE quantitative section to the Putnam isn’t reasonable. I wouldn’t be the least bit surprised if there are tons of people who are not capable of scoring anything at all on the Putnam, yet have perfect GRE quant scores.
In any event, seeing researchers using the naive measure (if this is indeed what they did) to compare what are blatantly apples and oranges makes me feel a bit uneasy.
We are aware that the SAT and GRE quant sections have low ceilings, which is why you also need a very high verbal score to qualify through one of those tests alone. But some extremely intelligent people fail to clear the verbal cutoff (especially likely when English isn’t their first language), so we figured it was worth mentioning that performance on the level of Putnam Honorable Mention gets you in regardless of your standardized test scores.
Also keep in mind that these are just automatic qualifying criteria. We’ll admit a fair number of volunteers who don’t satisfy any of them.
On the GRE in 2010, 800M was ~94th percentile, while 700V was ~97th percentile. The math section is just too easy for many students in engineering, physics, mathematics, and so forth.
I wouldn’t be surprised if the two results are equivalent by the naive measure: the fraction of Putnam participants who get honorable mention or better might well be about equal to the fraction of GRE test takers who get a score of 1500 or greater.
The real issue here, of course, is self-selection. Lots of people take standardized tests because they want to apply somewhere that requires the score, which drives the GRE average down. On the other hand, most people who take the Putnam do so because they enjoy the challenge, and are more likely than average to be good at math. This drives the Putnam average up, compared to what it would be if every college student in the US took it.
Edit: I don’t have combined statistics for the GRE, but between 5% and 6% of test-takers got 800 on the math section when I took it in 2010, while I estimate 700+ on the verbal section was about as likely. Obviously the two are correlated, but it seems reasonable to suspect that the combined percentage was less than the 2% of Putnam contestants who received an honorable mention or better.
I think the real issue is that comparing the GRE quantitative section to the Putnam isn’t reasonable.
I wouldn’t be the least bit surprised if there are tons of people who are not capable of scoring anything at all on the Putnam, yet have perfect GRE quant scores.
In any event, seeing researchers using the naive measure (if this is indeed what they did) to compare what are blatantly apples and oranges makes me feel a bit uneasy.
We are aware that the SAT and GRE quant sections have low ceilings, which is why you also need a very high verbal score to qualify through one of those tests alone. But some extremely intelligent people fail to clear the verbal cutoff (especially likely when English isn’t their first language), so we figured it was worth mentioning that performance on the level of Putnam Honorable Mention gets you in regardless of your standardized test scores.
Also keep in mind that these are just automatic qualifying criteria. We’ll admit a fair number of volunteers who don’t satisfy any of them.
Christopher Chang, BGI Cognitive Genomics
Great to hear from you, that’s good to know!
On the GRE in 2010, 800M was ~94th percentile, while 700V was ~97th percentile. The math section is just too easy for many students in engineering, physics, mathematics, and so forth.