AAS (American Astronomical Society) calls for removal of GRE and Physics GRE in admissions

[TakeruK]

In Astronomy, we have had discussions of eliminating the use of both the GRE and Physics GRE test scores in admissions for quite a while. People have written papers about their unfair use (see references in link below).

Today, the President of the AAS wrote an draft open letter to all department chairs in Astronomy with the following recommendation:

Recommendation: Given the research indicating that the GRE and PGRE are poor predictors of graduate student success, that their use in graduate admissions has a particularly negative impact on underrepresented groups, and that they represent a financial burden for many students in pursuing advanced degrees in the astronomical sciences, the AAS recommends that graduate programs eliminate or make optional the GRE and PGRE as metrics of evaluation for graduate applicants. If GRE or PGRE scores are used, the AAS recommends that admissions criteria account explicitly for the known systematics in scores as a function of gender, race, and socioeconomic status, and that cutoff scores not be used to eliminate candidates from admission, scholarships/fellowships, or financial support, in accordance with ETS recommendations.

For the full text of the letter (which comes with a bibliography) as well as the AAS President's comments accompanying the letter, please see: http://aas.org/posts/news/2015/12/presidents-column-rethinking-role-gre

To me, this is an exciting step in the right direction that will improve our field and help us recruit better junior members (and better colleagues for me!).

I'd be interested to hear what others think and whether you know if your field's national society have been thinking about similar actions.

[Queen of Kale]

While I understand there are systematic issues with standardized testing I also was told by my current PhD advisor that he only even considered my application because of my GRE scores. I was a nontraditional student, from a no-name school which I picked because it was affordable and where I lived, and he had no way to calibrate my application against the others in his stack besides the GRE. I had high grades but without knowing anything about my undergrad school he didn't know if those were meaningless or not. I can imagine for other nontraditional students or students trying to make a giant prestige-leap that standardized testing can be crucial (for better or worse). On the other hand, I can see how terribly things might have gone if so much weight had been put on this one test and it hadn't gone so well. Yet what other way can we calibrate students with such vastly different backgrounds? I am not convinced there is an alternative, and until there is, I think GREs should be at least an optional component of one's application (although the fee is outrageous, but that is its own issue).

[rising_star]

So to me, this is a really exciting development, and one which ties in with several other things happening in higher ed these days. The explicit recognition and acknowledgment that there's no correlation between GRE scores and PhD completion is important. Why? Because it demonstrates that students who struggle on the GRE may still be very qualified for programs AND because it points out that programs which ignore this are potentially doing themselves a disservice by missing out on qualified applicants. In this way, it reminds me of why many prestigious colleges and universities have gone test-optional for high school students. Again, the same (lack of) correlation between SAT/ACT scores and graduation is at the heart of this and has been explicitly cited by some of those institutions.

In addition, the timing of this proposal is interesting in light of the various student protests at many campuses. One of the things many of those students are calling for is greater faculty diversity. Anything which improves the pipeline by easing barriers to entry for diverse (first generation, underrepresented minority, female, etc.) potential graduate students will ultimately assist presidents and institutions who are seeking to diversify their faculty. Right now, such calls are difficult for colleges to act on because there are a limited number of diverse PhD students graduating each year. So, if going test-optional for the GRE can enhance the ability of diverse PhD students to gain admission, then it's entirely possible that there will be more diverse PhD-holders in the future. 

So yes, all in all, this is a good thing. I'm curious to see what happens in January when the AAS council discusses this. I'm even more curious to see if graduate departments will adopt this recommendation. And, if they do, whether it will be limited to astronomical sciences or will spread to related fields like physics.

 

(P.S. YAY! Finally something I wanted to say enough to make my 5,000th post here!!)

[TakeruK]

@Queen of Kale: I definitely agree with you that the GRE can be a way for non-traditional students to prove themselves. For example, in my field, there are many people who live in countries with either no opportunity to pursue a Physics BSc or high pressure from family and society against a Physics route. So, many students take Engineering degrees. The Physics GRE allows the admissions committees to compare these non-Physics students with Physics students on the PGRE. It's not clear how much correlation there is between PGRE score and Physics knowledge. So, personally, I'd take the approach of making the PGRE an optional item---i.e. just a chance for a candidate to demonstrate excellence instead of a tool to filter out those disadvantaged by the test.

@rising_star: Although the timing is coincidental with events happening nationwide, this is something the field has been considering for awhile. For example, here is a 2014 article: http://www.nature.com/naturejobs/science/articles/10.1038/nj7504-303a I do think though, that given what has been happening, this is the right time to push for a change like this! 

Now I'm a little sad that I won't be attending the AAS meeting this year! But I'll look out for an official announcement during that week (first work week in January!).

One of the authors of the article (MIller) linked above came to my school to give a talk about why the GRE and PGRE are not good indicators of grad school success. We had a small meeting with students and Miller. Miller said that he thinks the test does more harm than good and so he would not even recommend allowing optional score submission. He still felt this way after presenting a very similar argument to what Queen of Kale wrote above. I still think there is more good than harm to be done by allowing optional scores, but I guess it's something still up for debate!

[shadowclaw]

I think this is a great discussion at both the undergraduate and graduate level. I think to some degree, high SAT or GRE scores can indicate that someone has had a good education as far as math and reading comprehension go, but it doesn't necessarily mean that they are going to have the motivation or skills necessary to make it through a degree. There's also the problem that some people suck at standardized testing and some people are amazing at it. I have a friend who did amazing on her SATs and got a good scholarship based almost exclusively on those scores, and she ended up dropping out after two years. Clearly, the SAT scores were not a good predictor there. When contacting POIs last application cycle, I was in contact with one who told me everything about my profile was great except for my GRE scores and that I'd never be accepted unless I scored above the 90th percentile on all sections because they just don't want students below that cutoff. That's crap. My ability to sit in a room for several hours and ace a standardized test on math and reading has very little to do with my motivation or ability to do ecological research. In fact, I would argue that the GRE has nothing to do with being able to do research, and that's what graduate school is all about, at least at the PhD level. I also believe that the test disadvantages students in lower economic brackets - they can't afford all of the expensive test prep material or afford to take it again and again until they get the score they want. If high-quality test prep materials were freely available and testing fees were low, I think that would go a long way to making the test more fair. The same can be said about the SAT's... wealthy parents can afford to get SAT tutors and software for their kids, or at the very least, those kids can go to after school SAT prep classes. Many students in lower economic brackets get jobs and don't have time to go to those study sessions.

I find it very interesting that the research cited has found such a weak link between GRE scores and success. During the application process, I have seen it stated on several university websites that there is a strong correlation between GRE scores and graduate student success, which is why they place so much weight on them. I wonder where those schools got their information from. Interestingly, the program that I got my masters degree from actually didn't require GRE scores. Perhaps the graduate admissions committee is very enlightened.

[TakeruK]

It sounds like the AAS Council has voted to approve this letter and the AAS itself (instead of just the President) will be taking an official stance discouraging departments in the US from using GRE and Physics GRE cutoffs in graduate admissions. They will do more than just issue a statement---it sounds like a powerpoint "fact sheet" presentation is being put together for faculty to use to convince their departments of this position. This is just what people at the meeting have reported, I think a more official announcement will come shortly.

[RunnerGrad]

I've always found it odd that the U.S. places a significant emphasis on standardized tests both for undergraduate and graduate admissions.  These tests are a significant financial burden for students from lower income families, especially if they are ill on the day of their first test and need to retake it.  As a Canadian, I'm just not used to such widespread use of standardized tests, with a few exceptions (MCAT for medical school, LSAT for law school, GRE for some graduate psych programs).  Even so, some medical schools in Canada don't use the MCAT because it disadvantages students whose mother tongue is not English.  So universities with large populations of Francophone students (see UOttawa or NOSM), or exclusively Francophone students (see most of the Quebec medical schools) don't use it.

Edited January 6, 2016 by RunnerGrad

[TakeruK]

Here's another update. So far, three schools that I know of have indicated that they will no longer require the Physics GRE as part of the admissions process for Fall 2017. I am keeping an updated list over at PhysicsGRE.com but I thought I would share some info here as well. This is timely because the next subject GRE test is this Saturday, so this information is helpful so that students can use their free score reports elsewhere! 

Astronomy PhD programs that do not require the Physics GRE score, but allow for optional reporting, for the Fall 2017 admission season:
Last updated Sep 14, 2016
1) University of Arizona Astronomy: See https://www.as.arizona.edu/application-requirements-and-procedures

2) University of Washington Astronomy: See http://depts.washington.edu/astron/academics/graduate-admissions/

 

Astronomy PhD programs that do not accept the Physics GRE score (not even as optional reporting), for the Fall 2017 admission season:
Last updated Sep 14, 2016
1) University of Texas Austin Astronomy: See http://www.as.utexas.edu/astronomy/education/admit.html

[ctg7w6]

I honestly don't see the problem with the GRE besides the cost. Of course, waivers are available for low income people, so there is that (and I received the equivalent of a waiver, though not from ETS).

 

I agree that the GRE does not provide a good indicator of PhD success in terms of motivation, etc. Nevertheless, that is why there are other aspects of the package. The GRE is a one-off test. Only a sustained commitment (visible on CV, transcripts, etc.) can show motivation, and even then it isn't entirely accurate. The GRE is the graduate level equivalent of an IQ test in some ways. You can study for it, but there will simply be some people who will never get a perfect score or even a very good score.

 

A good GRE score shows that you either studied a lot, are very smart, or a combination of the two. It is one part of many in any application. I certainly don't agree with GRE cut-offs that immediately get your application thrown into the trash bin, but it can be a useful measure if used properly.

 

The real killer is the $27 fee for sending the scores. And schools already use the GREs generally well. I.e. the verbal section counts for very little for math and science, and the quant counts for little in verbal disciplines. 

 

We can call just about anything a significant barrier for certain groups of people. College itself can be expensive, a four year commitment is tough for people with extenuating circumstances, low-funded colleges don't provide opportunities for adding to a CV, GPAs aren't always representative, the GRE is tough for people, application fees for PhD programs are expensive, etc. I'm not saying we can't change things to make them better, but I guess I kind of look at it as if any requirements on an application give me another avenue to prove that I should be there. If I can't succeed well in a particular requirement, there are many other requirements that I can shine in.

[spunky]

On 9/14/2016 at 9:52 AM, TakeruK said:

Here's another update. So far, three schools that I know of have indicated that they will no longer require the Physics GRE as part of the admissions process for Fall 2017.

So... does this pertain solely to the Physics GRE or to both the subject-specific (Physics, in this case) GRE and the general GRE (you know, verbal + quant + analytic)?

[TakeruK]

3 hours ago, spunky said:

So... does this pertain solely to the Physics GRE or to both the subject-specific (Physics, in this case) GRE and the general GRE (you know, verbal + quant + analytic)?

The original statement by the AAS recommends decreasing the use of both Physics GRE and General GRE. The studies cited by this letter includes studies that investigated the effects of one or both tests. 

The recent changes in admission criteria, for the three schools mentioned, so far, only refers to the removal of the Physics GRE requirement. It's my understanding, based on my knowledge of this effort, that this is the first step towards removing the General GRE entirely. The Subject GRE is an easier first step, I believe, because the departments have the ability to decide if any such test is required, but the General GRE requirement may be a university wide decision. 

[spunky]

Thank you. It's just that I was chatting with a few friends from when I used to intern at ETS (and who now work there full-time) and they were telling me about all the expansions ETS is undergoing within the U.S. next year. Like they're opening more test centres and hiring more staff at the existing ones to accommodate greater demand for their  products, particularly the TOEFL and the GRE. 

I found it curious that as standardized testing draws more and more criticism, the demand for these same standardized tests also grows more and more. When I read your post I thought "well, maybe this is specific to Astronomy and that's why they're reaching a happy compromise of getting rid of the subject test but still asking scores for their other  "base tests". But it seems like they're aiming to just scratch the whole thing, at least from what you said.