Some admissions committee insights

[tinuvielf]

Despite the advantages that you observed, still many non-native English speakers are at a huge disadvantage when it comes to the GRE. A Japanese/Chinese/Vietnamese/Thai/Tibetan speaker certainly doesn't have the kind of advantages that you elaborated. And after all, the GRE verbal test does not measure a person's intelligence. I wish more people would acknowledge this fact.

I do acknowledge the fact. I'm quoting myself when I say "Of course, non-native English speakers are at a disadvantage in such tests—a large disadvantage, and I don't mean to discount that."

I also acknowledge that GRE verbal measures a great deal more than just vocabulary. (And I'm not talking about intelligence.) It measures your knowledge of a given language, which involves familiarity with colloquial expressions and other nuances that simply may not translate from one language to another. Non-native speakers have it hard. I'd be the very last person to say that they are not disadvantaged by the test. I'm always in awe of the non-native speaker who manages a high score on the verbal portion of that test. My earlier post was simply an attempt to point out an area in which many (esp. Indo-European language speakers, as I also mentioned) do possess something like a wedge into doing well on the test.

If it sounded like was disregarding the hurdles internationals face, I assure you that was not my intention at all.

Edited January 11, 2010 by tinuvielf

[jacib]

Do you suppose they add them together and then do the cut off, or look at the verbal and quant separately and then do it? Because I think it would eliminate a lot of potential talent just to add the two numbers together and take the whole score together without seeing if one is really high or not.

I've posted this elsewhere, but in sociology, a discipline that can either be quite quantitative or quite qualitative, my father (a sociology professor) got the following "inside deal" from a colleague at a public university with a top-ten ranked program :

"One low sore won't hurt; two low scores need to have something pretty outstanding to counter balance;a bad GRE and a less than stellar record will have a hard time, though we admitted one such person who wona prize for the best senor thesis."

This seems to argue in favor of that whole holisticism thing we've been talking about. The guy gave some other advice that seemed quite candid so I have to assume that this advice is also completely accurate, and if we are being "lied to" about GRE cutoffs, it's not by everyone. I think it depends on the program. (Based on non-empirical evidence) I think math/non-social science programs tend to use the presort a little bit more, but then again, they have things like perquisites which a social science or a humanities program usually don't have. Many social science and humanities programs don't even require a background in the subject. I think the

Anyway, the above comment seems to argue that for this school in sociology, they consider both numbers separately, and want at least one of them to be quite good. I've got the feeling that many schools will be different (my father guessed schools that are strong in demography want higher quant scores from their students, even the ones who don't mention demography in their statement but especially those who do). From what I've gathered on the forums, in English they don't even look at the quant (so people claim). In Religion (a field where number are probably only used ever so slightly more than English), that one needs at least an acceptable quant score is the general consensus and if schools have a cutoff, it seems to be based on the sum of the quant and verbal.

Lastly, its funny that no one else on here seems to be worried about a GPA cutoff. My coworker and I were talking about this today at work and for both of us, we thought these was our weakest features. He even said, "If they see passed my 3.07 GPA..." and he's right. His major GPA is 3.5+, his writing sample is really good (especially since it was an independent research project), his GRE's are both above 700+ (I forget exactly what), he has research ability in at least three languages. Why is no one worried about a GPA cutoff? People only seem to worry about a few bad grades they got in community college twenty years ago.

[tinuvielf]

I can't imagine a cutoff being that ridiculously high...for one of the top business programs, their AVERAGE accepted GRE is 1400, so for every 1500, you have a 1300 applicant accepted. (I'm sure that's not exact, but you get my point.) It seems 1200 would probably be on the higher end of cutoffs, but who knows. Some programs seem to be more "high GRE" conscious than others.

Okay, well this is somewhat encouraging. One thing that's been worrying me though is my non-US-institution grades. We have no GPA system, and the grading scale is ridiculous. B-plusses are like A's in my department, and there's no A-minus on the scale. I'm afraid the adcoms are going to take one look at my grades and throw away my file.

Plus Yale's English department website pretty much says only the undergraduate GPA really counts in their admissions process. Here's a quote:

"Selection is based on the applicant's undergraduate record, the results of the Graduate Record Examination and the GRE subject test on "Literature in English", evidence of motivation supplied in the candidate's personal statement of purpose, evidence of ability to do advanced work as expressed in a writing sample (10-15 pages) and supported by three letters of recommendation, and sufficient preparation in languages."

All that bold type is their emphasis, not mine. (And, btw, I goofed and sent in a 20-page writing sample by accident

But what of all those people who improve in the master's years?? I guess the original link to the Chronicle Forum that got this discussion going addressed that issue: "automatic rejection." (Sigh)

[jacib]

While I understand that having good acceptance rate stats can't hurt a school, I question its importance in grad admissions. How many of you actually took acceptance rates into account when deciding where to apply? I sure didn't...in fact I don't even know the rates for any of my schools. Do rankings of PhD programs even take selectivity into account? That said, I didn't even really look at rankings when I applied, but I would be surprised if they were even part of the equation...they just seem totally irrelevant. Grad school prestige is not about being an elite club; it's about the quality of research done in the program, which I imagine can be measured much more reliably using proxies like publication rates.

They don't and it isn't.

I have not seen any non-professional graduate school ranking that takes into account acceptance rates. It's all based on peer surveys.

National Research Council's methodology: Each university granting a doctorate in an area was asked to rate each of the other doctorate granting universities in that area according to faculty quality and effectiveness of program. The resulting responses were converted to scores (having two decimal places) ranging from 0.00 to 5.00 where for faculty quality 0 denoted ``not sufficient for doctoral education'' and 5 denoted ``distinguished,'' while for effectiveness of program 0 denoted ``not effective'' and 5 denoted ``extremely effective.'' Based on these scores, rankings were obtained of all doctorate granting institutions within each area.

US News and World Report's science methodology:

Rankings of doctoral programs in the sciences are based on the results of surveys sent to academics in computer science, mathematics, and physics during the fall of 2007, in biological sciences and chemistry during fall 2006, and in other fields during fall 2005. The individuals rated the quality of the program at each institution from "marginal" (1) to "outstanding" (5). Individuals who were unfamiliar with a particular school's programs were asked to select "don't know." The schools with the highest average scores among those who rated them were sorted in descending order and appear here. Surveys in biological sciences, chemistry, computer science, Earth sciences, mathematics, and physics were conducted by Synovate.

USNWR's social science and humanities methodology:

Rankings of doctoral programs in the social sciences and humanities are based solely on the results of peer assessment surveys sent to academics in each discipline. Each school offering a doctoral program was sent two surveys (with the exception of criminology, where each school received four). The questionnaires asked respondents to rate the academic quality of the program at each institution on a 5-point scale: outstanding (5), strong (4), good (3), adequate (2), or marginal (1). Individuals who were unfamiliar with a particular school's programs were asked to select "don't know." Scores for each school were determined by computing a trimmed mean (eliminating the two highest and the two lowest responses) of the ratings of all respondents who rated that school; average scores were then sorted in descending order.

There is more information on things like response rates on each of those websites. Also interesting is that the NRC has done a more recent survey but they don't know how to use the results! Also interesting is the unexplained, very slight differences in methodology between the science and social science methodologies. In fact, it's not even clear if they used different methods, or just wrote up the same methodology differently!

Anyway, no admission rates in and of themselves aren't taken into account by anyone, but I think schools like to keep application rates high for other reasons that rankings or prestige, namely, because they are wary of discouraging potential students, especially minority students or students with very specific specialties, who they might otherwise want, even if accepting someone with those scores is a very rare event. It's probably quite difficult to predict the pool year to year, especially once specialties are taken into consideration.

[iCyborg]

All the arguments in this thread were based on only one from-no-where post.

IMO, the way they process applications varies from university to university. The bottom line is that when you check top schools' student profiles, they all are great. This means the process works. And the number of applicants who were wrongly assessed are very limited. Also, I know some guys with GRE below 1200 are in top schools now. So I think it is the assessment of all factors in our applications. If one has potential to conduct successful research in grad school, with their experience the adm comm can very easily and quickly detect him/her out from the applicant pool.

[socialpsych]

The bottom line is that when you check top schools' student profiles, they all are great. This means the process works.

Student at top school --> great profile does not imply great profile --> student at top school...it could still be that some great applicants are ending up consistently in the reject pile. I'm not saying that's the case; I'm just saying you can't just look at students at top programs to rule it out.

[moralresearcher]

I agree. I can live with not meeting the bar of a dream school - oh hai Yale -- but I resent, absolutely, being lied to.

Here I again compare this to law schools. You can judge, fairly accurately, how you stand in law school admissions based on your GPA and LSAT. It's brutal but it's transparent and honest.

This is like when guys swear up and down that they look at "inner beauty". I'd rather you say "I really appreciate models and video dancers with low self-esteem." That is so much more fair.

The problem with having hard and fast GRE/GPA rules is that sometimes the applicant pool has a weirdly high or low score compared to past pools. Also, with the GPA--grade inflation/deflation definitely plays a part. I took an advanced statistics course where the average grade was an A (bordering on an A+, not something that should happen in such a difficult course), as well as the perceived reputation of the school you come from--private schools look better than public, your US News rank throws its weight around a bit, etc. So they rank people by the other applicants. My advisor said that something several programs do is take your GRE score and divide by 200, then add your GPA. (For example, I got a 1310 and have a 3.8, my score would be 10.35. Typically 10 is considered competitive and 12 would be perfect). They then rank the applications based on that number and read LORs and SOPs. If you have exceptional LORs or SOPs, it can bump you up the list over someone with lukewarm letters and a generic SOP.

[casey825]

This is like when guys swear up and down that they look at "inner beauty". I'd rather you say "I really appreciate models and video dancers with low self-esteem." That is so much more fair.

[ladedodaday]

The problem with having hard and fast GRE/GPA rules is that sometimes the applicant pool has a weirdly high or low score compared to past pools. Also, with the GPA--grade inflation/deflation definitely plays a part. I took an advanced statistics course where the average grade was an A (bordering on an A+, not something that should happen in such a difficult course), as well as the perceived reputation of the school you come from--private schools look better than public, your US News rank throws its weight around a bit, etc. So they rank people by the other applicants. My advisor said that something several programs do is take your GRE score and divide by 200, then add your GPA. (For example, I got a 1310 and have a 3.8, my score would be 10.35. Typically 10 is considered competitive and 12 would be perfect). They then rank the applications based on that number and read LORs and SOPs. If you have exceptional LORs or SOPs, it can bump you up the list over someone with lukewarm letters and a generic SOP.

This is an entirely arbitrary but weirdly compelling system. Maybe my mind is warped from spending too much time on here with overachievers, but wouldn't there be a huge logjam of applicants between 10 and 11? Either way, the whole process is really strange, but there seems to be no logical/fair way of doing it. Hopefully we're all lucky!

[natofone]

My advisor said that something several programs do is take your GRE score and divide by 200, then add your GPA. (For example, I got a 1310 and have a 3.8, my score would be 10.35. Typically 10 is considered competitive and 12 would be perfect). They then rank the applications based on that number and read LORs and SOPs. If you have exceptional LORs or SOPs, it can bump you up the list over someone with lukewarm letters and a generic SOP.

This would end up putting much more of an emphasis on the GRE than undergraduate GPA, no?

[Lauren the Librarian]

My advisor said that something several programs do is take your GRE score and divide by 200, then add your GPA. (For example, I got a 1310 and have a 3.8, my score would be 10.35. Typically 10 is considered competitive and 12 would be perfect).

I'm totally intrigued by weighting systems. I'd be a 10.62 in that set up.

I started fiddling and came up with my own equation: GRE score divided by 10, added to 40 times your GPA. Less than 200 is very weak, greater than 300 is very strong and you have 100 points in the middle for everyone to be spread out it (so they wouldn't get clumped in the 10s). I'd be a 283! Sounds about where I think of myself in the real applicant pool.

Edited January 15, 2010 by Lauren the Librarian

[LifeIsGood]

The problem with having hard and fast GRE/GPA rules is that sometimes the applicant pool has a weirdly high or low score compared to past pools. Also, with the GPA--grade inflation/deflation definitely plays a part. I took an advanced statistics course where the average grade was an A (bordering on an A+, not something that should happen in such a difficult course), as well as the perceived reputation of the school you come from--private schools look better than public, your US News rank throws its weight around a bit, etc. So they rank people by the other applicants. My advisor said that something several programs do is take your GRE score and divide by 200, then add your GPA. (For example, I got a 1310 and have a 3.8, my score would be 10.35. Typically 10 is considered competitive and 12 would be perfect). They then rank the applications based on that number and read LORs and SOPs. If you have exceptional LORs or SOPs, it can bump you up the list over someone with lukewarm letters and a generic SOP.

Ok, let's see. GRE divided by 200.... Add that to GPA... Adjust for windspeed... And we get... 10.9.

Wait, I'm not perfect? That can't be right....

[jacib]

I'm totally intrigued by weighting systems. I'd be a 10.62 in that set up.

I started fiddling and came up with my own equation: GRE score divided by 10, added to 40 times your GPA. Less than 200 is very weak, greater than 300 is very strong and you have 100 points in the middle for everyone to be spread out it (so they wouldn't get clumped in the 10s). I'd be a 283! Sounds about where I think of myself in the real applicant pool.

I like Lauren's system better, because I think GPA does mean a lot. However, you guys are all assuming addition. It makes more sense a multiplier... your high GRE scores attest to the rigor of your undergraduate program (I know schools weight GRE scores more heavily when they aren't familiar with your undergraduate university). So here's my suggestion:

GRE/1600*GPA/4= x%

I got 81%, or a B-. (Think about grades pre-grade inflation). On Lauren's scale, I got 288.

I wish I were applying to law school, so I could play little numbers games like this. And anyway, I think the GRE is probably on some scale where it affects your GPA 1+like the square of the number of standard deviations it is from the mean applicant (or more likely mean matriculant), so that really high and low scores matter more than average scores. And the GPA would of course be fudged in some way or another, probably less scientific, "Oh I met a really smart girl from there once who had really pretty eyes... she seemed to have a rigorous education/work out routine.... so I'm adding the mental equivalent of .2 to all applicants from that school."

[Lauren the Librarian]

So here's my suggestion:

GRE/1600*GPA/4= x%

I got 81%, or a B-. (Think about grades pre-grade inflation). On Lauren's scale, I got 288.

Multiplier? Hrm... Very interesting. On your scale I got a 79%, and the difference in our stats are similar to the difference in stats when using my scale (283 for reference). I actually like your scale better than mine because it gives a number and concept we're all familiar with. I tried recalculating my score with the GPA I should have gotten (if I hadn't had so many external family issues during my undergrad) and came up with an 86% - which sounds about where I belong among my fellow collegic minds.

Whee! I could play this game all night!

[ColorlessGreen]

All of these scales still seem to end up with a bunch of applicants packed into the same range of numbers, though, don't they? Or are we just skewing the results with the particular crowd on this forum? By moralresearcher's advisor's scale, I end up with 10.9, same as LifeIsGood. By Lauren the Librarian's scale, I end up with 296, which is close enough to both her and jacib's scores as to make no real difference. And by jacib's scale, I end up with an 85%.

I think that no matter what scale the schools might use, they'll always end up with a lot of applicants who appear very similar numerically. So jacib, I hope you are right in that admissions committees must notice extreme scores more than ones lumped in the middle.

The more I learn about the application process, the weirder it seems to me. I live in fear of being reduced to a set of numbers and a few short writing samples.

[intextrovert]

Considering I'm in the humanities, I really like numbers. So I crunched mine, too:

On jacib's scale, I got a 91%.

On Lauren's, 302.8.

On moralresearcher's, 11.52.

As fun as these are, I highly doubt the programs I'm applying to would ever use such a system. For one, as coyabean noted, English programs don't care two cents about your quantitative score unless maybe it's alarmingly low, so my randomly high Q won't be helping me. I really think we tend to focus intensely on this stuff because it's something concrete and quantifiable to fret over rather than the sort of nebulous concepts like quality of SoP, writing sample, and "fit," which is near impossible for us to figure out how a department will evaluate. Not to say that there might not be some cutoffs, but I bet they do initial glances to flag anyone who's borderline but might be a good fit.

Further proof that they likely don't use such a system: I was rejected everywhere but one my first round with these same numbers, which on these scales seem to be pretty high. I'm sure people with weaker stats than mine that were less naive, better focused, and a better fit were accepted, and I'm okay with that. Because now I'm older and wiser and I focused purely on making the non-quantitative side of my app better!

[jacib]

All of these scales still seem to end up with a bunch of applicants packed into the same range of numbers, though, don't they?

Sigh, yes. Law school this is not. This is why we're judged on stupid things like our writing quality, research experience, language skills, and most of all research interests. Oh, and what professors whisper about us behind our backs. The concrete things like numbers only tell let us know that we're competitive, not that we'll get it. I think my GRE scores will get my app looked at very carefully.... but its obviously the other things that will get me in or not.

Hence stories like inextrovert's above. Which is why the process is so nerve-racking. There is really no way we can compare ourselves to other candidates, especially us folks in the humanities/social sciences. Everyone has positive SoP's.... no one knows how glowing all of theirs are.

[moralresearcher]

Yes, it is pretty arbitrary, and yes, everyone is pretty much in the middle. Alas, the normal curve. I think I like Lauren's scale the best... I think I have the best stats there. I want the whole thing to be over already.