physics gre = irrelevant

  • If you want to know something about the GRE subject test in physics then chances are you will find it in here.
  • If something about the physics GRE it isn't already discussed in here then please put it in here.

Post Reply
nlmlms
Posts: 2
Joined: Sun Nov 08, 2009 10:29 am

physics gre = irrelevant

Post by nlmlms » Sun Nov 08, 2009 10:44 am

for all you guys and gals out there that were able to stroke their egos comfortably after receiving some numbers from a test - wake up. for everyone who didnt do as well as theyd like to have....wake up.

this test, a 100 question multiple choice test, does not, can not, and will not determine the calibre of scientist you will become as long as you don't give up on what you want to do. i was admitted to a top 10 program for HEP theory and scored a 460 on my PGRE. this was mostly due to my research, and social nature (i dont have aspergers lol), personal statement, and rec letters. schools today, especially top programs care less and less and less about this idiotic test.

for those who did well, and believe it to be a rite of passage into the realm of graduate studies in physics, you need to change your attitudes. in 10-15 years when you may very well be professors sitting on a committee accepting students, dont make decisions based on these ridiculous tests, review all the material! and for those who did poorly, always always keep on top of what you really want, explore all your options, contact as many prospective professors as you can and get involved in dialogues with them!

theres no point to continue this tradition, and for what? to keep out the huge number of people who would jump on board to become physicists if these tests werent there to destroy their chances?!!? hahahahahah, that would be the day - what a joke. its too bad we dont go back to the way it was in the 50's, plenty of excellent scientists were educated then with no problem and no pgre.....and it seems we have a shortage of scientists now.....gee i wonder....

always remember, this is a game, theres always a way to win : )

blackcat007
Posts: 378
Joined: Wed Mar 26, 2008 9:14 am

Re: physics gre = irrelevant

Post by blackcat007 » Sun Nov 08, 2009 11:57 am

nlmlms wrote:for all you guys and gals out there that were able to stroke their egos comfortably after receiving some numbers from a test - wake up. for everyone who didnt do as well as theyd like to have....wake up.

this test, a 100 question multiple choice test, does not, can not, and will not determine the calibre of scientist you will become as long as you don't give up on what you want to do. i was admitted to a top 10 program for HEP theory and scored a 460 on my PGRE. this was mostly due to my research, and social nature (i dont have aspergers lol), personal statement, and rec letters. schools today, especially top programs care less and less and less about this idiotic test.

for those who did well, and believe it to be a rite of passage into the realm of graduate studies in physics, you need to change your attitudes. in 10-15 years when you may very well be professors sitting on a committee accepting students, dont make decisions based on these ridiculous tests, review all the material! and for those who did poorly, always always keep on top of what you really want, explore all your options, contact as many prospective professors as you can and get involved in dialogues with them!

theres no point to continue this tradition, and for what? to keep out the huge number of people who would jump on board to become physicists if these tests werent there to destroy their chances?!!? hahahahahah, that would be the day - what a joke. its too bad we dont go back to the way it was in the 50's, plenty of excellent scientists were educated then with no problem and no pgre.....and it seems we have a shortage of scientists now.....gee i wonder....

always remember, this is a game, theres always a way to win : )
whatever you said is known to most of us here. whats your point?? :?
even if we don't like this "tradition" what can we do?? go on some sort of strike or stop attending the schools??
there is no way out I believe atleast in the near future..even as I write to profs of different universities, they emphasize the importance of PGRE. sometimes one has to yield to the authority no matter how unfair it is.
and by the way, I am finding it hard to believe that you got 460 and went into a top school. because if you had that caliber in research as you tried to emphasize here , then you can never end up with that score in the first place..

nlmlms
Posts: 2
Joined: Sun Nov 08, 2009 10:29 am

Re: physics gre = irrelevant

Post by nlmlms » Sun Nov 08, 2009 12:35 pm

blackcat007 wrote: whatever you said is known to most of us here. whats your point?? :?
even if we don't like this "tradition" what can we do?? go on some sort of strike or stop attending the schools??
there is no way out I believe atleast in the near future..even as I write to profs of different universities, they emphasize the importance of PGRE. sometimes one has to yield to the authority no matter how unfair it is.
you've missed my point, obviously as an applicant you can't do much but state your situation, but i'm saying, down the road if you ever hold power over another person as a graduate committee member may, consider more than just the score. the system can change, it just takes the willingness of the people with power. and we will have the power in time.....

blackcat007 wrote:I am finding it hard to believe that you got 460 and went into a top school. because if you had that caliber in research as you tried to emphasize here , then you can never end up with that score in the first place..
yeah i find it hard to believe as well! ill give you a hint, i dont have to take quals here : )
anyways, its mostly because i can schmooze, and i have excellent research, and half of my recommenders are famous.

and obviously, you are incredibly incorrect that if i am as good a researcher as i'm claiming, that its impossible i got that low of a score.....it has absolutely NO correlation to how good i am at physics whatsoever. this is exactly where YOU (and anyone else with your sentiment) can begin changing your attitude about how this inane test is viewed by peers.

jones
Posts: 27
Joined: Tue Jan 06, 2009 6:23 am

Re: physics gre = irrelevant

Post by jones » Sun Nov 08, 2009 12:56 pm

I agree with blackcat007 -I cannot imagine how somebody with 'excellent research' and recommendation letters from famous scientists make such a lame PGRE score. The test is not perfect, and a perfect score may not mean you're great at physics, but a score as pathetic as 460 would prove how bad you are at physics. And the whole 'schmoozing' thing sounds a little fishy to me.

delton
Posts: 32
Joined: Mon Sep 21, 2009 9:22 pm

Re: physics gre = irrelevant

Post by delton » Sun Nov 08, 2009 1:22 pm

hey, I agree with nlmlms, the PGRE dosn't determine how great a physicist you are.. of course there are other factors.

This reminds me of 2 things:

First, I know one person who is really dead-set on HEP and math-phys. He placed out of physics 1 and 2 and knows barely anything about those subjects. He goofed off in E&M and thermo and had no interest in those classes. That kind of scenario might explain how someone could do poorly on the PGRE yet still be good at advanced, specialized topics. Still, I believe that if you want to succeed in any type of physics, you should have a good foundation in the basic stuff. I believe physicists should be well-rounded and have a general knowledge of the field and its practical applications. I think many grad-school committees feel the same way. Still, I suppose some might take a canidate with a low PGRE score, who is really good in a particular area, but that is a special circumstance.

Also, there is another good reason grad-schools care about the PGRE, is because it shows dedication. One of the most important factors in sucsess in grad school is a person's work-ethic, ie, their level of passion and dedication. A dedicated person will be willing to study hard for the PGRE and get a good score. So in a sense, the PGRE is a "big hoop" you have to jump through to show you are really dedicated and serious.

So, it is surprising you were able to get into a top school, and if its true it is certainly a special circumstance.

Bottom line:

Physics GRE is important: study hard for it, learn the material, show the grad schools you are serious.

But remember, its only one factor on your grad app, and of course, it doesn't determine how good of a physicist you are or will become, in any strict sense.

blackcat007
Posts: 378
Joined: Wed Mar 26, 2008 9:14 am

Re: physics gre = irrelevant

Post by blackcat007 » Sun Nov 08, 2009 3:17 pm

nlmlms wrote:
you've missed my point, obviously as an applicant you can't do much but state your situation, but i'm saying, down the road if you ever hold power over another person as a graduate committee member may, consider more than just the score. the system can change, it just takes the willingness of the people with power. and we will have the power in time.....
ok I see your point now.. true, knowing the real scenario, if we do get power someday we should act justly.

nlmlms wrote: yeah i find it hard to believe as well! ill give you a hint, i dont have to take quals here : )
anyways, its mostly because i can schmooze, and i have excellent research, and half of my recommenders are famous.

and obviously, you are incredibly incorrect that if i am as good a researcher as i'm claiming, that its impossible i got that low of a score.....it has absolutely NO correlation to how good i am at physics whatsoever. this is exactly where YOU (and anyone else with your sentiment) can begin changing your attitude about how this inane test is viewed by peers.
well true there is no such direct correlations, and i also agree that one who aced it may be a tyro in physics or vice verse,
but one familiar with the PGRE will surely agree, that there are atleast 25-30 questions in each paper which is based on just rational thinking, its just based on daily experience and stuff that we interact with.. like problems involving concepts of projectile motions or conservation of momentum and energy, simple newtonian mechanics, basic resistor in parallel and series circuit, rolling disks etc. these I think are carved in anyone's head. even if they are not, then a person can answer them intuitively. forget about SR, QM, lagrangians and hamiltonians and other adv topics.. even if you leave them.. you can easily score a scaled score >25 or so which is I believe is atleast a 550.
I neither have the authority nor will I go to that extent to ask for a scanned score report.. but I think you are engaging a reductio ad absurdum just to prove the irrelevance of PGRE for Physics PhD admission.

User avatar
quizivex
Posts: 1031
Joined: Tue Jan 09, 2007 6:13 am

Re: physics gre = irrelevant

Post by quizivex » Sun Nov 08, 2009 3:43 pm

I'm shocked someone with a 460 was admitted to a top school, regardless of research merit. It seems like connections and gender must have been a factor. No offense to nlmlms, who I fully respect, but I say this because I just don't want everybody on the forum making the wrong assumption that the PGRE doesn't matter and that any bad score can always be overcome with other things so they can just blow it off. A 460 would be an insurmountable disaster for most applicants to top programs, period! The test warrants your full attention.

Nobody said the test directly determines our potential as scientists, but what school we go to is a major factor in our development, and the test can directly affect that!

Ryalnos
Posts: 33
Joined: Sat Nov 14, 2009 1:14 pm

Re: physics gre = irrelevant

Post by Ryalnos » Sat Nov 14, 2009 1:18 pm

nlmlms wrote: and obviously, you are incredibly incorrect that if i am as good a researcher as i'm claiming, that its impossible i got that low of a score.....it has absolutely NO correlation to how good i am at physics whatsoever. this is exactly where YOU (and anyone else with your sentiment) can begin changing your attitude about how this inane test is viewed by peers.
I'm certain they are mathematically correlated to some degree.

nathan12343
Posts: 249
Joined: Sat Dec 20, 2008 9:34 am

Re: physics gre = irrelevant

Post by nathan12343 » Sat Nov 14, 2009 2:18 pm

Actually, it really does have practically zero correlation with research ability: http://blogs.discovermagazine.com/cosmi ... e-student/

It DOES have a very high correlation with prelim and qualifier scores, but that's probably unsurprising.

Ryalnos
Posts: 33
Joined: Sat Nov 14, 2009 1:14 pm

Re: physics gre = irrelevant

Post by Ryalnos » Tue Feb 23, 2010 2:21 am

Remember that those graduate students were the successful applicants. Those with lackluster GRE scores must have had strengths in other areas. I doubt that, all other things being equal, PGRE scores would have zero correlation with physics ability.

The survey does speak well of Chicago's admissions process, however. Apparently they've figured out how to weight the PGRE fairly in their evaluation of applications.

User avatar
quizivex
Posts: 1031
Joined: Tue Jan 09, 2007 6:13 am

Re: physics gre = irrelevant

Post by quizivex » Tue Feb 23, 2010 12:52 pm

Well stated, Ryalnos!

I've said the same kind of thing before. People who hate the GRE will inevitably conclude "the GRE sucks" no matter what the data or premises are.

Since admissions are based on a holistic evaluation of the many components of a student's record, it's natural that the "successful" applicants who were weaker in some area were on average stronger than the typical successful applicant in other areas, to make up for the weakness. Otherwise they would not have been admitted.

So the students with weak GRE scores probably had exceptional research accomplishments or outstanding recs, and thus were equally qualified in the first place. To test the true correlation between the PGRE and future success or ability, you must look at all applicants, not just the successful ones!

If there's no correlation between PGRE scores and some metric of grad school success/performance, then that actually reflects FAVORABLY on the test! Statistically, it means the PGRE is equally as good as the other selection factors in predicting success (think about it). That's the way it should be. If admission is based on a weighted sum of the pieces of a student's record (A*PGRE + B*GPA + C*School Reputation + D*Quality of recs + etc), then in theory, if the optimum/ideal weights were used, there shouldn't be any correlation between any of the individual admission criteria and student success.

geshi
Posts: 200
Joined: Tue Dec 01, 2009 12:01 am

Re: physics gre = irrelevant

Post by geshi » Tue Feb 23, 2010 1:47 pm

I like your argument, but I'm not sure I agree with you 100% quiz. If there was a mutli-variable linear function associated with "grad school success," then you would expect to see some kind of correlation. Let's assume we have: A*PGRE + B*GPA + C*School Reputation + D*Quality of recs + etc, just like you said. Let's set this equal to "arbitrary graduate school success." For some arbitrary candidate, let's fix the summation of B*GPA + C*School Reputation + D*Quality of recs + etc to some arbitrary constant (say z). So now we have A*PGRE + z = arbitrary graduate school success. Clearly there's now a linear relationship between the candidates PGRE and their arbitrary graduate school success. Now we would expect to see at least some upward trend of grad school success as PGRE score increases. Of course the value of A could be a variable based on the strength of the rest of the application. In other words, if the constant "z" is a very large value, then the value of "A" would be smaller.

This is, in my opinion, how it SHOULD work. The stronger the rest of your application, the less emphasis they should place upon the physics GRE. If the rest of your application is weak, they should be placing a higher emphasis on the PGRE. The physics GRE *is* somewhat irrelevant if you have a strong application outside the PGRE.

matonski
Posts: 121
Joined: Thu Dec 04, 2008 5:03 pm

Re: physics gre = irrelevant

Post by matonski » Tue Feb 23, 2010 2:12 pm

But quizivex's point is that z is not constant but should ideally vary opposite that of A*PGRE for a given "arbitrary grad school success." Keeping z constant is the same thing as saying, "all else being equal, a student with a higher PGRE will more likely be successful than a student with a lower PGRE." No one's arguing that part.

geshi
Posts: 200
Joined: Tue Dec 01, 2009 12:01 am

Re: physics gre = irrelevant

Post by geshi » Tue Feb 23, 2010 2:46 pm

No, I get that.
Of course the value of A could be a variable based on the strength of the rest of the application. In other words, if the constant "z" is a very large value, then the value of "A" would be smaller.
I was assuming this is more what quizivex meant.

I guess I didn't state my point very well.
Keeping z constant is the same thing as saying, "all else being equal, a student with a higher PGRE will more likely be successful than a student with a lower PGRE." No one's arguing that part.
I was trying to make the point that if you took multiple candidates with "z" being the same, but different PGRE scores, I doubt you would see a correlation. This would be a reflection of it not being a very good predictor. Of course this is somewhat of a moot point since we don't have access to said study from U of Chicago.

My personally opinion of the PGRE is something along the lines of:
It DOES have a very high correlation with prelim and qualifier scores, but that's probably unsurprising.
Most "IQ" exams measure a person's ability to learn in a classroom setting. IQ ≠ intelligence. I view the PGRE something along the lines of an IQ exam. It shows how well you learned classroom physics material. This is somewhat irrelevant to research. With enough work, anyone can learn classroom physics.

kroner
Posts: 217
Joined: Fri Sep 25, 2009 1:58 am

Re: physics gre = irrelevant

Post by kroner » Tue Feb 23, 2010 3:09 pm

Assuming their formula is perfect, if the total weighted score S of everyone admitted was equal, then you'd expect that performance would have no correlation with individual elements of the total score. But that's not the case, since instead they're using a cutoff. Some people will be near the cutoff and some will be much higher, so there's room for variation in S. People with a higher S will probably tend to have higher PGREs (it's possible not, but that would mean PGRE has a negative correlation with the other elements being considered which doesn't seem plausible). So we would still expect some correlation between PGRE and performance. It might not end up being very much though.

User avatar
quizivex
Posts: 1031
Joined: Tue Jan 09, 2007 6:13 am

Re: physics gre = irrelevant

Post by quizivex » Wed Feb 24, 2010 1:10 am

Yep good point kroner. The cutoff concept is important and I did consider it, but I didn’t include it in my comments cuz I was trying to keep from writing another endless post that users have to scroll down to read, (like this one lol). Z would not be constant but I estimated it had small variance when I said minimal/no correlation would be expected (so this is my response to geshi too).

Remember, the “studies” people here have referenced (Harvard, Chicago??), testing correlations of the PGRE with whatever, took place at one particular school. The students there already achieved a certain given level of success on their record, which includes a combination of many factors, to be admitted in the first place. So to go back and compare their PGRE scores to their future success seems silly to me. That’s the key IMO.

Indeed, Z is not constant. There is in theory, a cutoff below which applicants are rejected, and the current students could be anywhere above it. But note that for a typical program, students who are outliers by their standards are accepted to better schools** and usually don’t attend, so the variance of Z shouldn’t be high at one place.

I had implicitly assumed the spread of Z was small and if you approximate Z as constant, one should expect no correlation. But still, for a small spread in Z there should only be a small correlation, likely statistically insignificant, which is really what they mean when they report “no correlation”… It wasn’t exactly zero, just not statistically significant in their opinion.

If the correlation was “less than just small”… like close to zero, or even small negative, well we could come up with many lurking factors that could sway the stats that way. For instance, the glut of international students who have perfect GRE scores due to various reasons that aren’t indicative of higher potential. For instance, they have more physics-oriented undergrad studies… take less BS than we do, and they don’t get exposed nearly as much to research. Since they don’t necessarily have more long term potential than Americans, a typical international student with a 990 that is average in terms of success in the program they go to will actually lower the aforementioned correlation, perhaps making it negative if it is already near zero.

** And if you’re considering the top schools, there’s really a limit to how “good” a profile can be. Most of the students there have near perfect grades, GRE’s and come from good undergrad programs so even the most extreme outliers in ability and performance aren’t going to be outliers compared to the rest, with respect to these variables.

User avatar
quizivex
Posts: 1031
Joined: Tue Jan 09, 2007 6:13 am

Re: physics gre = irrelevant

Post by quizivex » Wed Feb 24, 2010 4:52 pm

Here's a simplified example to help explain the aforementioned ideas. I can't spend any more time discussing this but I wanted to try to make it clear once and for all.

Suppose a committee wants to predict success by giving students 2 tests. One is a physics test of a variety of problems from different topics, of varying kind, length and difficulty, including lab exercises, literature searches, and an oral talk (i.e. lots of stuff we need to do in grad school). Another is a test on how many of the first 100 digits of Pi they can memorize in an hour.

They then try to predict future success by the formula B*Test1 + A*Test2 = Z (which is what admission committees do, in principle, even if they don’t calculate anything). By common sense, we know the first test will more accurately predict success in physics grad school. But raw memorizing skill (like retention of knowledge) is also useful and important in any career, so A should still be nonzero**.

We know the better predictor is Test1, and Test2 is still relevant, but much less so. So the optimum predictor has B>>A. Suppose the committee unwisely switched B and A, giving the second test most of the weight, A*Test1 + B*Test2 = Z. Now consider a group of students of equal predicted potential, Z. Now track their success in grad school 5 years later, measured by Z’, which is optimally estimated by B*Test1 + A*Test2.

We want to measure the correlation between the test scores and Z’… Solving for the test scores:

B*Test1 + A*(Z – A*Test1)/B = (B – A^2/B)*Test1 + (A/B)*Z = Z’

B*(Z-B*Test2)/A + A*Test2 = (A-B^2/A)*Test2 + (B/A)*Z = Z’

Recalling B>>A , and Z was constant, the first and second equations correlate the following:

B*Test1 vs. Z’

-B^2/A * Test2 vs. Z’

THIS IS THE KEY!


The coefficient in the second comparison is negative, so Z’, the student’s future success, is negatively correlated with the Test2 scores, but positively correlated with Test1 scores. Thus, Test2 was given too much weight, and Test1 should be given more weight. If the optimum coefficients had been used in the first place, all students of given Z would have the same Z’, so Z’ would have been constant and there would be no correlation between Z’ and either test. (In an extended real life example, Z’ would not be constant for a group of students with the same Z, but the correlation would be a minimum if the optimum process were used.)

If the committee had neglected the second test entirely, their formula would be Test1 = Z. Inserting this into the optimum formula B*Test1 + A*Test2 = Z`, where B>>A we have B*Z + A*Test2 = Z`. Thus, you’d find that Test2 actually correlates with Z`. We’d conclude Test2 needs to have more weight.

So by interpolating, we see there should exist optimum coefficients, A and B, such that the correlation between either test and Z’ is zero (or a minimum in a real life example). The coefficient of any merit-related metric should be nonzero in an admission process, and some (like grades) should be larger than others (like VGRE). Something completely irrelevant, like hair length, minority status or softball ability would optimally have zero weight, unless they are some kind of lurking variable connected to a true merit factor.

**For instance, problems on prelims/qualifiers often require random tricks and methods that even students of exceptional intelligence and would be unlikely to dream up on test day. A big factor in succeeding on these types of tests is memorized examples of similar problems. I’ve heard a few people say they only got a passing grade on the prelim because some familiar problems showed up. So indeed, the assumption that the optimum memorizing test coefficient is nonzero is valid.

rahuldatta
Posts: 22
Joined: Sun Jan 11, 2009 8:39 pm

Re: physics gre = irrelevant

Post by rahuldatta » Thu Feb 25, 2010 1:25 pm

i completely agree with "nlmlms"... there is no point in just blaming the system.....if we (at least some of us here will certainly) reach such positions in adcoms then we must take upon ourselves the responsibility of doing whatever we can to improve the system

___
Posts: 23
Joined: Fri Feb 22, 2008 6:57 pm

Re: physics gre = irrelevant

Post by ___ » Thu Feb 25, 2010 10:26 pm

quizivex wrote:If the optimum coefficients had been used in the first place, all students of given Z would have the same Z’, so Z’ would have been constant and there would be no correlation between Z’ and either test.


Something is fuzzy here. If we have some infallible admissions committee that can exactly predict future success Z' by some formula, i.e., Z' = A*test1 + B*test2, then as long as test1 and test2 aren't negatively correlated with each other, there will definitely be a positive correlation between future success and each of the tests. If either of the test scores goes up independently of the other, then Z' also goes up. Hence, positive correlation. The only way this wouldn't work is if the admissions committee decides to only admit students with the same future success, so that if one applicant has a higher test1 then he/she necessarily has a lower test2. But that's clearly not how admissions works.

Is there something I'm missing in your logic?

kroner
Posts: 217
Joined: Fri Sep 25, 2009 1:58 am

Re: physics gre = irrelevant

Post by kroner » Thu Feb 25, 2010 11:45 pm

___ wrote:
quizivex wrote:If the optimum coefficients had been used in the first place, all students of given Z would have the same Z’, so Z’ would have been constant and there would be no correlation between Z’ and either test.


Something is fuzzy here. If we have some infallible admissions committee that can exactly predict future success Z' by some formula, i.e., Z' = A*test1 + B*test2, then as long as test1 and test2 aren't negatively correlated with each other, there will definitely be a positive correlation between future success and each of the tests. If either of the test scores goes up independently of the other, then Z' also goes up. Hence, positive correlation. The only way this wouldn't work is if the admissions committee decides to only admit students with the same future success, so that if one applicant has a higher test1 then he/she necessarily has a lower test2. But that's clearly not how admissions works.

Is there something I'm missing in your logic?

Yeah, he's assuming that Z' is essentially constant among all admitted students. The committee establishes a lower bound on Z' of accepted students, and there's effectively an upper bound as well since students who are more qualified will get into a better school and take that offer instead. That leaves a narrow band of Z' scores among the students that end up attending the program. In that case, yes the two tests will be negatively correlated among attending students because of the way selections are made.

I disagree somewhat with that premise, especially for a top school like UofC. At top schools although there should be a sharp cutoff on the lower end, I would expect a pretty big range above that. Plus there are a lot of factors beside the ranking of the program that affect people's decisions, which makes things not work out so neatly. I don't know how much one should expect those factors to throw off the results though. Maybe not much.

User avatar
quizivex
Posts: 1031
Joined: Tue Jan 09, 2007 6:13 am

Re: physics gre = irrelevant

Post by quizivex » Fri Feb 26, 2010 5:21 am

@ "___" As kroner said, a key premise was that sigma of Z is small.
kroner wrote:I disagree somewhat with that premise.
Whatever Sigma(Z) actually is, it’s clearly much smaller for the enrolled students in one program than sigma for the whole group of applicants accepted there, which itself was much smaller than sigma for that school’s entire pool of applicants. Only the latter sample, or some random sample of students who haven't already been sorted by their undergraduate record, is legitimate to test how the PGRE correlates with future success.

Therefore, when trying to correlate the PGRE with success at one grad program, the result will be much closer to my artificial example than to the legitimate analysis of an unbiased sample, regardless of how close my approximation of sigma(Z)=0 is.

How much closer? Nobody will know until someone does a valid study. However, the people who keep claiming “The PGRE sucks” with no premise at all… or claim "no correlation -> PGRE sucks" have done far less to justify their claim than I have.

As for kroner's comment on top schools, I mentioned that in an earlier footnote. Yes, students can't turn down top schools for better schools, so you’ll have students there that are talented and those who are ridiculously talented. However, there’s an upper limit on some of the numbers measuring this talent. Since the stats for students there are already so high, and many are at or near the maxima of the PGRE and GPA ranges, the most extreme talents at these schools are not going to score any higher than the rest. (They'll be distinguished by other things like research accomplishments, winning competitions etc... )

So if you study the group of extreme talents alone, you'd get no well defined correlation between the PGRE and their success because there's practically no spread in the former variable. The correlation for success at the top schools with respect to the GRE will come from the rest of the students who were very strong in some areas and just decent in others (i.e. students with excellent research and Ok GRE's like grae vs. those with excellent GRE's and Ok research like butsurigakusha.) This range of Z should be comparable to those at average schools and the result should be similar.

User avatar
HappyQuark
Posts: 762
Joined: Thu Apr 16, 2009 2:08 am

Re: physics gre = irrelevant

Post by HappyQuark » Tue Mar 16, 2010 9:49 am

While I agree with nlmlms general conclusion, that being that the PGRE shouldn't be the end all be all of grad school admissions, I couldn't disagree more with the specifics and in fact would argue that nlmlms has vastly misunderstood the role that the PGRE plays in the admissions process. It never has been and was never intended to be a singular indicator of potential for success as a student or a researcher. There are limits to what a multiple choice physics test can tell you about a persons abilities, and I imagine that is why it is only one of the many criteria that admission committees look at when judging you. With that being said, the PGRE can be a good indicator of a persons knowledge and intuitions on the subject and, if nothing else, a demonstration of a persons diligence and work ethic to study for and do well on the test.

If it is indeed true that you managed to get into a top 10 program in HEP theory with only a 460 on the PGRE (which I must admit seems very unlikely to me) then congratulations are owed to you. With that said, it is absolutely not a result you are justified in generalizing or recommending to others. Even with a 4.0 GPA, bucket loads of relevant research and even primary author publications, it would still be difficult for the large majority of students to get into a top tier school (not even specifically top ten) with such a poor PGRE score.I don't mean to put you on the spot, nlmlms, but would you be willing to provide more information about your admissions? Perhaps the other schools you applied to and whether or not you were accepted to those as well. Is the program you were accepted into a top ten physics program or did you mean to say just that the school in general is considered top ten (e.g. Yale is an Ivy League university that doesn't necessarily have a top ten physics program, let alone a top ten HEP theory program).
Last edited by HappyQuark on Tue May 18, 2010 1:09 am, edited 1 time in total.

cryingsun
Posts: 60
Joined: Wed Apr 07, 2010 2:59 pm

Re: physics gre = irrelevant

Post by cryingsun » Mon May 17, 2010 10:27 pm

nlmlms wrote:it has absolutely NO correlation to how good i am at physics whatsoever
Well I think a good researcher knows that it is hard as hell to prove the absolute absense of correlation of two quantities from observation data, and it's safe to assume that any data that is in favor of "NO correlation" is even worse than the data from astronomers :wink:

rachely476
Posts: 1
Joined: Wed Oct 13, 2010 5:13 am

Re: physics gre = irrelevant

Post by rachely476 » Wed Oct 13, 2010 5:19 am

nathan12343 wrote:Actually, it really does have practically zero correlation with research ability: http://blogs.discovermagazine.com/cosmi ... e-student/

It DOES have a very high correlation with prelim and qualifier scores, but that's probably unsurprising.
Thanks for sharing the link
Last edited by pqortic on Thu Oct 14, 2010 12:54 am, edited 1 time in total.
Reason: link in signature

larry burns
Posts: 77
Joined: Wed Jul 02, 2008 1:12 pm

Re: physics gre = irrelevant

Post by larry burns » Sun Oct 17, 2010 7:55 pm

what about the other way around? Would having a great PGRE score make up for an otherwise bad application? For example, getting an PGRE 850+, with otherwise mediocre stats such as 3.6 gpa, avg LORs, avg general GRE, no publications, top 5 state school?

I heard from some professors that the PGRE has NO relevance at all as some profs on the admissions committee don't even LOOK at it. However, I've of course heard from others that its very important


Based on what I've seen from the applicants results on this site, its hard to say as there seems to be a very strong correlation with gpa and PGRE scores. That is, I hardly saw anyone with 850+ PGRE with a 3.7 gpa or less

User avatar
WhoaNonstop
Posts: 853
Joined: Mon Sep 21, 2009 1:31 am

Re: physics gre = irrelevant

Post by WhoaNonstop » Sun Oct 17, 2010 10:08 pm

At top schools, they look at the whole profile and how well-rounded it is. Being strong in all areas of course is better than being strong in less.

-Riley



Post Reply