Feb 20 2015 Published by drugmonkey under Postgraduate Training
What are your program's standards for GRE percentiles and GPA?
That is, what would be the minimum score that would be essentially unremarkable, and require no other compensating attributes, to justify an invitation to interview?
48 responses so far
Speaking for a master's program here.
GPA: 3.0 for clear admission. Anything below that gets a proverbial red flag.
GRE: We require students take it, but we evaluate them very much on a case by case basis. When I show the scores to the reviewing committee, I show each subsection individually, and highlight precentiles below the 50th percentile as a red flag.
no GPA or GRE requirements....only a pulse and minimal respiration rate requirement.
Yeah, a GPA of 3.0 and above from a school I've heard of and a combined GRE above 50% is enough for a serious look-see. Our program admits directly into a lab, so the due diligence is a little more PI-dependent. I honestly think some kids, particularly Americans, think they need to be much more distinguished to get into graduate school. Not so.
For a PhD program in most of the departments in top 25 US medical school:
GRE: >50% in at least one section (overall average is often 3.0
In my university, the best source of thinking, competent labour are the undergraduates that are considering medical school. Their GREs are >90% on most of the sections and GPA>3.8. They will work with you for 1-2 years, when they take MCATs. I found it much easier to train and work with them, than PhDs.
Should we call it 'the race to the bottom of intelligence' in the PhD track?
You should ask the same question in 5 years.
No GRE (unreliable) and grades in the top 25% of graduates (GPA's are unreliable).
No strict requirements, but I'd say they have to have above 80th percentile on the quantitative GRE to be considered for our computer science program. Being American also helps.
I personally was rejected from all but 1 of the grad programs I applied to, despite perfect GRE scores. I think it was because my undergrad GPA was a bit low (3.15ish) and I was applying to top-10 programs.
For Ph.D., minimum requirements are a GPA > 3.0 and GRE scores (we just require students to take them). Ours is a direct admit, so there's a ton of variability beyond the minimum requirements.
My program doesn't have strict minimums. Research experience is the main criterion (with accompanying letters of recommendation). GPA's are almost always >3.0, but we take where the student obtained their Bachelor's into account and look at the transcripts for science and math grades. We require the GRE, but it's the least important criterion.
Program staff look at every application and triage ones that are weak in all three areas (GPA, GRE, research experience) before the apps go to committee, but almost all get reviewed by at least 3 faculty members on the admissions committee.
Based on the quality of our students, I would say that a high school diploma is the minimum requirement.
Hahah. When's the last time you interacted with a normal distribution of 12th graders, Dave?
I'm now on the admissions committee for our PhD program (well known NE medical school) and I'm both surprised at the diversity of scores, grades, research experience and bravado in our applicants. I agree that I ignore the GPA and GREs if they are >3 and >50%. In parsing the applications I typically look for a letter of recommendation that says "I would keep so-and-so in my lab if I could" and/or some clear example of independent, creative thinking. But the interviews are essential. My impression, yet to be contradicted, is that interview questions about experimental design and appropriate controls distinguish good applicants from bad.
Contemplating returning to school, DM?
I did grad at a top 20 school and helped on the admissions committee, so I saw the scores for most the kiddies, and they very much had two distinct pools. If you had no tech experience, you generally had a 3.5 or above from a good university, and really good GREs (subject and general >90% for math, 70% or better for math [stupid engineers]). If you had tech experience from a good lab that wrote you a good letter, anything above a 3.0 and 50% was probably good enough.
Also, it probably depends on the breadth of your departments. We had very tiny programs (5-6 students per committee per year), and certain traits were more valued in some departments than others. Molecular Genetics definitely tended to weight lab experience heavily, while Human Genetics was very academics focused.
Given the current state of our department's funding, current full time job that will pay your tuition as a part-time PhD student = acceptance. Other students need not apply.
50th percentile GREs? I am startled.
DM, I think the broad consensus is that the GREs aren't predictive of much other than test-taking skillz. It is apparently somewhat problematic for TG renewals to have low avg test cores.
But a <50% is a red flag, and my committee would need additional strong evidence to pull up one of those for an interview or acceptance.
Another question - what's the typical matriculation rate? We seem to be ~30-40%, although we have a large umbrella program and compete with many top schools. But our overall program size has shrunk in half in the past 5 years! Dean be cutting, big time...
Designating <50th percentile as a red flag means a decision has been made that half of the people applying to grad school should probably do something other than grad school. An assertion like that is a Rohrschach test, either eminently sensible or outrageously elitist.
You can't say GREs aren't meaningful and then use them as a cutoff or eyebrow raised anyway.
"Designating <50th percentile as a red flag means a decision has been made that half of the people applying to grad school should probably do something other than grad school."
Is that an unreasonable standard? If a program has x spots to fill, and they have 15x applications, *something* has got to differentiate one applicant from another. Given how common research experience in or in the years after undergrad is these days, using that as your sole determining criteria is probably not sufficient to get down to x offers.
I'm totally open to the debate about how good the GRE is (or isn't) as a predictor for success in a PhD program, but if you're going to use it at all I don't see why 50th percentile is an awful place to draw the line.
GPA: 3.2 minimum for the department, 3.0 for the grad school, though we can do provisional admission below that with a petition to the grad school.
GRE: No minimum, though anything below 50% of Quant is not good. We generally ignore verbal except for international students.
Of those admitted, our average GPA is ~3.8 and GRE quant scores of ~70%.
Doesn't seem awful to me at all. But, like I said, it's a Rohrschach test.
A physicist (!) once said to me that he doesn't see why the quantitative GRE should get any weight at all, since it's just a test of "middle school math" (his words, not mine). I'd be inclined to agree that success on a test of "middle school math" says precious little about one's ability to succeed in physics, but bombing a test of "middle school math" probably says quite a bit about one's prospects in physics...
Strangely, I've met a few physicists who approvingly quote some (possibly accurate, possibly not) folklore about the verbal GRE being the only good predictor of success in physics. Even though the verbal section would seem (to my perhaps naive mind) to be the one most subject to cultural biases, these guys reasoned that it tests "soft skills", and "soft skills" are "soft", so therefore filtering based on them is OK, unlike filtering on the basis of "hard" stuff. If this makes no sense to you, don't worry; it didn't make any sense to me either.
The verbal section of standardized tests is most highly correlated with IQ and other measures of cognitive horsepower.
The people I've interacted with on this subject would not speak approvingly of measuring IQ. They would speak approvingly of measuring "soft skills." Whether they are speaking from an informed perspective is a separate matter.
The problem with the middle-school math is that it often tests formulas that physicists and other quantitative students probably don't remember. For example, my GRE had a few questions to calculate arc length. These things might be derivable given enough time, but the quantitative section is notoriously long (or at least was about 5 years ago before the newer changes).
If the student spend literally no time studying, they would likely do poorly on some of these types of questions.
Keep in mind that natural scientists and engineers aren't the only people taking the GRE quantitative section. If a physics student was below 50th percentile on a test taken by people from far less mathematical disciplines, there might be a problem. And if the problem is that they spent no time preparing for something that they knew they would need to do well on, maybe that's valuable information for an admissions committee to have...
My impression is that, in the US at least, the GRE is the poor cousin of standardized tests. People freak out for months about the SAT, LSAT, and MCAT, take special cram courses & etc. For the GRE, most people seem to just show up. I know that's what I did.
I'm referring to the general test, the subject tests might be different.
we don't use GRE, which is a ridiculous but very lucrative hoax.
3.2 GPA, which is just unfair.
Alex is correct, we use it as a proxy for their ability to prepare for a graduate school by taking a pointless and expensive exam. Though, it is also used for departmental comparisons, and our higher admin is obsessed with rankings, so we do have to require it.
As a predictor, it is pretty worthless. We have found no correlation with GRE scores above 50% and success/failure. Below 50%, it is a mixed bag, but we have a small sample size.
The one thing we do use extensively is research experience. Those with meaningful research experience are what we look for, and it just so happens that these students often have great GPAs and GRE scores, though not always. Over the years, our admissions committee has gotten a bit more conservative in admitting the academically weak (GPA <3.2) student with research experience as we have a high failure rate with these students, usually at the prelim stage. This is why we bumped our GPA requirement a bit over what the university takes, though we do have a few faculty that will take any warm body.
I am in quantitative sciences, and I worked with many people with quantitative GRE score in range 10-99%. High school students - advanced postdocs.
The ones with quantitative GREs 90% could do almost everything they were asked for. Easy to explain the problem, able to find the solution without complete dataset provided to them. And all of this without years of 'training' (mostly undergrads).
GREs measure your abilities, even if there is no perfect correlation with actual performance in workplace. The fact that only the kids with scores ~ 50% apply suggests that grad school lost its charm.
Should we call it 'the race to the bottom of intelligence' in the PhD track?
Here I'm amazed at the qualifications of the graduate students we interview. Virtually all have significant research experience--with publications--and GPAs generally 3.5-ish.
We/I don't pay much if any attention to GREs. Letters from research advisors count the most.
I would not have been accepted to my department (much less interviewed) because of lack of any research experience (there are those who think this may have been a boon to science).
Why is research experience looming so large folks? Screens out those who aren't ready for the hardships? Or is it prior training? If the latter 1) why are graduate stints so long? And 2) aren't you confessing you see grad students as labor, not trainees?
Prior research experience is clearly screening out those who aren't ready for the hardships AND screening FOR those with a love of doing science. That's the biggest key for me, b/c without a love for the mostly ambiguous nature of the daily science grind, mixed in with the _occasional_ eureka moment, I think that students will have a hard time. But as with most admissions decisions, I doubt there's any hard data on this.
Yes. The way I've always heard the need for prior research experience sold it basically "doing science is really different than learning about science in a classroom; a lot of people who like classroom science don't actually like doing science."
WRT Research experience:
I don't know if it 'looms large' its just that every applicant we look at seems to have significant experience at the undergrad level. It would be interesting to hear if that is unique to us or a trend out there.
As I said, I'm pretty amazed at the qualifications (high GPAs, publications).
you asked: "... 2) aren't you confessing you see grad students as labor, not trainees?"
Research experience correlates best with grad students who will most unambiguously be seen as "great" in professorial perception (http://www.molbiolcell.org/content/25/4/429.full).
Since we already know that the #1 factor correlating with successfully getting through graduate school is access to financial support, and we know anecdotally that undergraduate research experience is a lot easier if you don't have money concerns, I'm gonna bet that we gate on prior research experience because it correlates with economic resources. And all the socio-cultural problems that entails.
My uni recently instituted what I consider to be fairly rigorous GRE standards: must be >60%ile on both verbal and quant, >75%ile on at least one. If we want to interview someone who doesn't meet the criteria, we have to write a letter explaining to the dean why they deserve a shot. It's insane.
Insanely low bar? Agree.
Also, the "research experience" bias raises the importance of school selection as an *undergrad*. No SLAC students need be considered, nor small state U students. Even if their scores are excellent.
This seems very wrong to me.
No SLAC students need be considered
That seems backwards. At least in my experience, it is especially easy to get meaningful research experience at a SLAC.
What about the analytical writing section of the general test? Do your programs look at that?
I am at large state U that has moderate research activity. We typically screen based on GPA and GRE. Many of the doctoral students we get are from abroad, do not have much in the way of research experience, have very good (typically >85 %-ile) quant scores, but verbal scores in the 40-60 %-ile range. Of greatest, concern, however is the writing scores generally being under 50 %-ile. Somewhat unsurprisingly, they are great at implementing complex techniques, but struggle with communicating structured arguments both orally and in writing. I also think this is related to their ability to critically read papers.
Am I on the right track in advising undergrad + MS trainees applying to PhD programs (biophysics, bioeng) it is their research experience (papers!), more than GPA+ GRE scores, that will enable them to complete with applicants from top-ranked programs?
"That seems backwards. At least in my experience, it is especially easy to get meaningful research experience at a SLAC."
Ditto that. The SLAC down the road from us actually has some pretty awesome gear at their disposal in part due to wealthy donors and in part due to securing NSF undergrad-oriented equipment grants and whatnot.
With regard to the low bar for grad school apparent above, it upsets me a little less than it might have a coupla years ago. I've had a few undergrad students now, often through ASPIRE/McNair type programs whose intelligence and technical competence as observed first hand doesn't necessarily correlate well with their GPAs, let alone GRE scores. The specific explanations for the discrepancy vary, but generally fall into the category of socio-economic factors noted by Becca above. A not uncommon example in my experience is the first generation student who overstretches themselves and burns out as a result of a real or perceived pressure to be the family Golden Child who makes it big. So while I've got a bit of an issue about over-recruitment of grad students, I see the why it's generally a good thing to retain sufficient flexibility to allow for the occasional strong students that have perhaps been unfairly filtered out by a screening system that is largely the invention of largely middle-and-upwards class white folks for triaging other largely middle-and-upwards class white folks.
We have a solid URM pool (~20%) in our PhD class and make a concerted effort to recruit a diverse population. We also have had a PREP (http://www.nigms.nih.gov/training/prep/Pages/default.aspx) program and recruit students who have completed these programs to gain research experience. So I'm not convinced that screening for research experience is a proxy for wealth.
So DM, how would you screen applicants? There has to be some criteria...
I think the subject GRE's are also pretty important to consider, but I don't know how many schools require them for applicants.
The reason the quant score is so useless is that the distribution is top weighted. You need a nearly perfect score to break the 90% barrier, because all the engineers taking the GRE absolutely crush that section (it's a lot of geometry and trig). Verbal might be weighted the other way, since foreigners generally do poorly on it. I remember my verbal and quant were numerically the same, but about 20% points apart.
Does no one weight the written argument score into their considerations? Or do they not do that anymore?
Re: research experience: I think I got into my program because it was clear that, despite my university having few research labs (small liberal arts), I was clearly desperately trying to gain research experience. This was evident from the absolute hodgepodge of internships in many different places over the undergrad years. A couple years later one of the faculty told me he had noticed it and it made him recommend my application.
Do most places require and undergraduate thesis for a biology degree? Mine did, and it probably would have served as a pretty accurate report of my strengths and weaknesses.
Wait, DM would use test scores for graduate admissions?
In my 'top-40' chemistry department, we don't have any automatic interview criteria. Typically we will admit 3.5 GPA or higher, as long as GREV and GREQ scores are above 60% percentile. This corresponds to about the top 35-40% of applications. But this is not guaranteed admission/interview. We had an applicant with a 3.9 GPA from a major research university, with 85percentile plus GRE on each section who we did not interview/admit. The reasoning? He was four years at a major research university and never stepped foot in a research lab. All of his letters were from teaching faculty, not research active faculty. We concluded that despite opportunity, he did not choose to pursue research options earlier. We would have had a different view if he had come with these raw scores in a setting that wasn't a major research university.
We have held our PhD programs to a very small size after cutting our intake by about 25% 3 years ago. We've actually gone away from GRE as a primary criteria, and look more closely at performance in quantitative coursework and yes, prior research experience. Having said that, post hoc average GPA is ~3.4 and the GREs (post hoc) remain competitive. Our average time to degree is about 5 years.
The other major change we've made is to go from an exclusively PhD program to a more comprehensive MS that can convert to a PhD path - in fact our growth right now is in new MS students. Our overall goal is to use the MS as a mechanism for students to try their hand at advanced work, prove themselves capable to themselves and faculty and many track into the PhD. In interviews, we're activity scanning for who might truly be a PI. Others are encouraged to pursue health professions if they are interested in research but don't show signs of being able to develop the level of independence required. These changes have had the effect of doubling our percentage of URMs in the last 2 years to about 25% of enrolled graduate students.
I think we're trying to do the right thing for the environment - control PhD over production, improve the cohort that does advance into a PhD path, increase URM access, and provide a respectable MS option that leads somewhere. Time will tell.
For the PhD program, for automatic acceptance, we use combined GRE score of 320 or above and a 3.4 or higher in science GPA.
For masters, the automatic acceptance is 310 and a GPA of 3.2 or higher in science GPA.
With that said, most of our PhD students have GPA's around 3.2-3.8 from small schools with GREs scores of 150-160. The same basic thing is true of our masters students.
The kids that generally do the best are those who generally have research experience before entering a PhD program,
I get the impression that committee selection is very unscientific from reading these comments. I only see one link to existing studies of graduate student quality. A lot of comments seem to be based on personal experience; that is, comparing subsets of their own already small nonrandomized sample. The only thing you seem to agree on is research experience is important.
DrugMonkey is an NIH-funded researcher who blogs about careerism in science. And occasionally about the science of drug use.
Site Admin | Theme by Niyaz
Drugmonkey Copyright © 2017 All Rights Reserved