Last week, the National Science Foundation announced this year's awardees of the Graduate Research Fellowship Program (GFRP). It’s a big deal: only 2,000 out of about 13,000 applicants win the fellowship annually. Moreover, the GRFP is one of the few substantive grants that targets aspiring scientists - undergraduates in the grad-school application process, and grad students in their first or second year of study – and covers their expenses for three years of a five-year window. I am a 2014-19 fellow, and the GRFP allowed me to switch into a lab with a better advising fit, reduce my teaching requirements, and participate in non-traditional training – like the writing program with Massive.

I'm eternally grateful for the opportunities the fellowship gave me. But I also empathize with the digital hand-wringing that accompanied this year's announcement in academic circles: as in years past, the awardees were disproportionately from elite academic institutions, which points to deeper systemic issues with the GRFP program that penalize applicants that don’t have access to the advantages of such places. And it's these students who need funding most – elite schools usually award grad school packages that will cover them for five or more years. Less elite universities tend to offer three years – less than the average four to seven years it takes to earn a PhD.

Still, Harvard – which serves around 7,000 undergrads, and where I'm a PhD candidate – received 43 GRFPs this year. In contrast, the entirety of the California public university system, which serves more than 400,000 undergrads, received 50 awards. There's no question about how awesome the awardees are – they’ve earned it! – but rather about how awarding fellowships primarily to students at elite institutions serves to perpetuate inequality in science.

I applied for the fellowship twice. My first attempt was as an undergraduate on a huge scholarship at Roger Williams University, a small liberal arts college in Rhode Island (experience tells me you probably haven’t heard of it). I didn't get the fellowship. The following year, when I applied as a graduate student at Harvard, I did.

While the broad strokes of my proposed research project were the same across the two applications, the second application was far stronger thanks to the resources I had access to at Harvard.

These came in the form of a writing group led by a professor who had a side-interest in creative communication. She shared her insights into good writing and into the NSF from her own recent experience of writing full grants. But more importantly, she had piles of old applications and reviews from the writing groups of years past, had overseen the writing process behind those applications, and had reams of advice from (anonymous to us grad students) former GRFP application reviewers based at Harvard.

With access to help like this, it's hard to see how students from less connected schools stand much of a chance in the application process.

The NSF states that the purpose of the GRFP is to “…help ensure the vitality and diversity of the scientific and engineering workforce of the United States." And, in fact, for the 2016-17 academic year, the NSF changed the rules of the GRFP application so that grad students can only apply once, instead of twice, in hopes of increasing applicant diversity, stating that the change "should result in more individuals applying as undergraduate students... a more diverse population than admitted graduate students." But the number of undergraduate awardees has remained unchanged – while an increasing percentage of these awardees come from the top 30 schools in the nation.

This could be because the the expectations of applicants that are set by reviewers are becoming impossibly high. Many applicants who were not awarded were told that they couldn’t receive the award due to a lack of publications. My personal experience corroborates this: my highly unusual set of two undergraduate publications were among the stated reasons I won the award. More typically, undergraduates outside of elite institutions have limited access to research experiences, whether it’s from a lack of active faculty researchers at the institution or of money to pay the enormous publication fees demanded by scientific journals. It’s also hard to do research when you have to work part-time. The experience GRPF judges appear to be looking for is really only available to the most privileged undergrads and early graduate students among us.

It takes tremendous effort to write a GRFP application, including its personal essays and, hence, the personal commentary from the reviewers. That plus the low success rate and the increasing standards for awardees makes it seems likely that the current system is deterring excellent future scientists who come from less-advantaged backgrounds from applying. Why apply at all if you don’t have any papers? Or if you have limited research experience because you couldn’t afford to take an unpaid position?

Within the GRFP system, there is still a lot of room for improvement. For starters, I’d like to see more transparency about the people who judge the applications. Review panels often lack voices from underfunded universities, and faculty from elite universities are less likely to engage in mentorship with female or minority students than faculty at public schools. While it is likely that much of this disparity in awardee demographics comes from unconscious biases, it is alarming that there is a real possibility of the same biases playing out in the GRFP review system. Implementing institution-blind reviews could help prevent biases about the institutions the applicants attend from arising in the first place.

I’d also like to see more panelists share some of their insights into the reviewer experience publicly, beyond the generic (and somewhat exclusionary) advice we typically see. For example, in the writing group for my second GRFP application, I learned that the reality of the enormous expectations on a scientist means many can often only spare a minute or two to look at any one proposal. This dramatically changed how I formatted my application – moving to a model with bullets and a small but striking figure, easily digested in a skim. There's no reason this information should only be accessible to people attending Harvard. (There is now a database of people who have volunteered to field questions, but as an undergraduate applicant outside the elite institutions, I wouldn’t have felt comfortable reaching out to a complete stranger from a far better university about my application.) Virtual writing groups or forums for the GRFP, which go deeper than posting old essays, could perhaps help serve more applicants.

More broadly, we can work to change systems which keep underserved students from achieving the successes more readily available to students from elite universities. We know that GRFP – and grad schools – prefer applicants with research experiences. So we can increase opportunities for underrepresented students in general.

First, we can avoid offering unpaid research positions that disadvantaged students can’t afford. We can focus more of our recruiting of summer researchers in the NSF’s Research Experience for Undergraduate program from disadvantaged universities. And we can provide more support for undergraduates at scientific conferences, where students are able hone their presentation skills, learn more about current work in their field, meet their peers, and connect with potential supervisors.

In recent years, overcoming inequity and fostering diversity have become major parts of our collective consciousness. In the science world, we have begun to take our own steps toward these ends. But we are far from done. There are reasons more significant than the CV line item for a Harvard grad student to want a GRFP award, but these benefits, and the accumulating opportunities that accrue from them, need to be shared more equally.