Diversifying the Sciences

[rising_star]

I read this article recently courtesy of NPR and found it very interesting. In particular, I'm intrigued by the GRE score analysis they reference. "("You can basically entirely explain the makeup of our graduate programs based on GRE scores," Stassun says.)" There's more but that piece is particularly interesting. So, do you think these types of analyses and articles will lead to a change in how graduate applications are evaluated?

 

Also, I'm curious to know more about these bridge programs. Has anyone here participated them or is hoping to? Do you think they're enough to "level the playing field" between women, minorities, and others interested in STEM PhDs? If not, what else needs to be done? Does it even matter whether STEM programs are diverse?

[NoSleepTilBreuckelen]

Great topic! Someone sent me this article earlier and it's been on my mind. I certainly hope the the GRE (and SATs for our college-bound counterparts) gets phased out in US grad applications in light of evidence like this. In addition to the author's point that GREs are written in such a way that they are skewed to certain groups and against others, we all know they are SOOOOO expensive, which can skew the results by socioeconomics (e.g. whether someone can afford to take/retake, and how many an applicant can afford to send out).

 

Great STEM innovators/scientists come from all backgrounds (race, gender, socioeconomics, etc.), so it's frustrating to see something like the GRE that's acting against diversity in the STEM fields. I was about to say that I'm comforted by the fact that I don't get the impression that universities put that much weight on GREs, but even by requiring the GRE they are demanding that potential students shell out a minimum of $185 just to be able to apply. And usually it's more than that, once you factor in 1) study material, 2) transportation to the test, 3) sending the test to more than 4 schools. With a retake that could hit upwards of $500 - yikes!

[juilletmercredi]

I did not do a bridge program like this; I did other kinds of minority/diversity programs in undergrad - the major one being a two-year NIMH-funded fellowship in my last two years of college.  It actually sounds quite similar to the Columbia one, with the exception that it's during junior and senior year rather than post-baccalaureate.

 

Anyway, no, I don't think these analyses/articles will change graduate admissions significantly - at least not in the near future.  I feel like admissions committees at top schools simply are not reading them.  My department requires a minimum of 310 on the GRE; before that the minimum was 1200.  My advisor was the DGS and I've been somewhat involved in admissions here, and I know that that minimum is pretty rigid.  Individually, scores of 155 are about the 65th percentile on both sections; I'm not sure what the distribution is combined (information on that is difficult to find).

 

What's more, I've yet to see any evidence that GRE scores are correlated with graduate school success.  I think they're weakly to moderately correlated with first-year GPA in a graduate program, but in research-based PhD programs, who cares about that?  Getting through the coursework in a science PhD program is the easy part; plus we have undergraduate science GPA as that proxy.  But I think GRE scores are one of those things that universities do to try to get better rankings or something, and it's a game of chicken - Columbia's not going to dismantle theirs until Harvard does, but Harvard won't do it until Michigan does, and Michigan won't do it unless Princeton or Penn State or Wisconsin does...so on and so forth.

 

Also, no I don't think these bridge programs are the answer to leveling the playing field.  They're a start - but funding for them is limited (especially given sequestration); conservatives will always argue over them and try to defund them; and they have a small number of slots by necessity so that the program can give the attention to each student that they need.  Plus, you have to know about them.  I go to Columbia, am in one of the departments the bridge program serves, and I have never heard of this bridge program until now.

 

What I think needs to happen is

 

-Science faculty acknowledging that diversity in the sciences is important and actually does affect the good practice of science, and making actual changes in their admissions policies - either lowering or eliminating GRE score cut-offs; considering applications holistically; but - MOST importantly - doing recruiting and outreach where talented minority undergraduates gather.

 

I worked in a summer program for underrepresented minorities interested in public health for the last two summers, and did advising with the students.  I told them about graduate programs and careers that they never even heard of before, and spent an enormous amount of time sitting down with them to discuss this stuff.  Now I didn't mind because that's exactly what I signed up for, but I thought it was a shame that these talented and mostly competitive undergrads were AT my university, clearly interested in public health but hadn't really heard of my university's unique joint program.  (*I* hadn't heard of it, either; I Googled it when I was a senior in college.)

 

I've suggested to my department (which is, by their admission, very interested in recruiting diverse undergraduates) to set up a recruiting table at http://www.abrcms.org/'>ABRCMS.  One of our other doctoral programs already goes, we could just hitch up with them and do a joint table.

 

-Science faculty and career services offices at these undergraduate schools becoming aware of these programs and recommending them to their students.  Especially at minority-serving institutions and HBCUs.  And faculty at these places - the "feeders" for minority undergraduates - attempting to make connections with their colleagues at the top departments so they can help funnel their students there.

 

-Free or inexpensive GRE prep, while programs are still sorting out how to change their process to be a more holistic one.

 

A really great model in my own field is the University of Michigan's psychology PhD program.  Every year Michigan has a recruitment weekend, and their psych department has student groups dedicated to Latino and black psychologists.  They have special events for undergrads of color and even invite up ones that haven't be admitted yet, pay for you.  I know a couple friends who've done it and they say it's a great time.  Michigan's in the Midwest, a place where people tend to complain it's difficult to recruit minority students.  Now take a look at the department's pictures, especially in the developmental, personality, and social* areas (http://www.lsa.umich.edu/psych/programareas/personalityandsocialcontexts).  And Michigan is a top 5 program in the field.  I think it may be #1 or #2.

 

Clearly, I have tons of thoughts on this.

[sys88]

What's more, I've yet to see any evidence that GRE scores are correlated with graduate school success.  I think they're weakly to moderately correlated with first-year GPA in a graduate program, but in research-based PhD programs, who cares about that?

 

You haven't done your research. There's actually quite a bit of evidence that GRE scores can be predictive. GRE scores can predict success in classes (GPA) but also drop-out rates.

Here's a link to ETS's list of resources.

Here's a case-study report.

This source claims that GRE scores do not predict landing a job, but that first-year scores do. However, if you follow other sources you'll find that GRE scores can predict first-year scores, meaning that there is possibly a correlation between GRE scores and finding a job post-graduation.

Of course, I'm sure some of these correlations are more field- and program-dependent than others, but the fact stands that the GRE scores are not a completely useless metric. You could also argue that GPA is not a good indicator due to grade inflation and differences in rigor (undergraduate institution prestige). Should GPA cutoffs be lowered or removed?

I am not a fan of standardized testing myself, and feel that GRE should be a factor but not a cutoff. On the other hand, I've never had to parse over a thousand applications so perhaps it is reasonable to use it as a cutoff in situations like that. It is fine to debate the current system but let's not get silly and completely discount potentially useful metrics. GRE scores can be useful in context of a complete application.