NIH seeks input on capping grant awards
June 9, 2026
The NIH has posted a request for input on a proposal to limit the number of grant awards a given PI may hold at any one time.
NOT-OD-26-086 Request for Information: Proposal to Cap the Number of Simultaneous Research Project Grants per Principal Investigator to Support More Researchers and Maximize Scientific Productivity and Innovation
The NOT is careful to say this is about a cap on the number of awards, concluding with “The NIH recognizes that RPGs have different budgets and durations. It will be up to applicants to decide which RPGs to apply for and what budgets to request.” The NOT also appears to focus on the size of research teams, as per “All of these effects could be due, at least in part, to PIs being unable to effectively oversee and manage research teams once they become too large and complex“.
However, this seems like utter nonsense given the rationale sprinkled through this NOT mentions taxpayer investments, marginal returns as grant funding increases, best use of taxpayer’s dollars, broadening distribution of funding, supporting more PIs, freeing up funds for other uses, etc, etc. None of the apparent goals can be accomplished or meaningfully addressed if each current PI is converted to a single grant with the same number of aggregate NIH dollars. Right?
It’s about capping the total dollars to any given PI. It seems most likely that the RFI is phrased this way to try to avoid responses that specify what that dollar limit should be. I guess.
So what is “a grant”?
We will focus on the scope of NIH R01 awards, since this mechanism remains the standard point of reference.
The modular budget system was put in place 6/1/1999, at a time when the majority of R01 proposals came in with requests at $250,000 or less per year in direct costs. Placing the cutoff at $250,000 was clearly and explicitly intended to cover “most” proposals and in the initial years it did. This old post of mine links to now-disappeared NIH pages and a PDF file. If you want evidence for my assertions, you’ll have to go to the wayback machine.
“NIH data indicate that almost 90 percent of competing individual research project grant (R01) applications request $250,000 or less in direct costs. On the basis of this experience, the size of the modules and the maximum of $250,000 were selected…..initial few rounds after introducing modular budgeting found that 92.4% of grants were submitted under the limit and that 41% requested either $175,000 or $200,000 in annual direct costs.”
I’ve been keeping track of the spending value of nominal dollars since the modular limit was established, since it has never been adjusted for inflation. This uses the BRDPI (Biomedical Research and Development Price Index) for adjustment (2025 is identified as projected). This analysis is benchmarked to $250,000 in 2001 dollars. The red bars indicate the erosion of spending power of $250,000. The black bars indicate the nominal dollar value required for spending power equivalent to that inital $250,000 in 2001.
The spending power of a full-modular grant (without any cuts upon funding) is now $120,481. A “grant” is half of what it was, in this analysis.
The current dollar figure needed to equal the spending power of $250,000 in 2001 is $506,651 direct costs per year. This bumps up against another limiter, i.e., the longstanding obligation to get formal Program Officer permission to submit a grant with direct costs of $500,00 or more. This has now been removed, for reasons that are not entirely clear.
One of the quickest ways to spot-check R01 size across fiscal years is to search RePORTER for 1 R01 (new competing) in a fiscal year and to gate on some reasonable proxy for modular versus above-modular budgets using the total cost field. We have to make some assumptions about the indirect cost rates to bracket and because of this there is going to be some error. I used $400,000 in total annual costs to try to capture this.
In 2010 62.7% of new R01 were funded at $400,000 total costs or lower. In 2015 this had fallen to 53.6%, in 2020 to 29% and in 2024 it was 11.9%.
This was 8.2% in FY2025.
A far cry from the 92.4% of proposals submitted under the modular limit after it was first introduced.
The dam has broken. Frankly I did not notice this in real time and only cottoned on far too late. No longer are even the newest of PIs reluctant to submit above-modular proposals. No longer are study sections looking askance at anyone for going above that limit. No longer are POs heavily discouraging submissions requesting above-modular amounts.
I will acknowledge that it is possible that the big change here is in smaller scope proposals not faring as well at study section. We cannot access success rates for R01 proposed below or above the $250,000 limit. But I suspect that this is not what is going on, based on my limited recent study section experience and what junior faculty are telling me.
This all tends to suggest that “a grant” is, or should be, one that is awarded $500,000 in direct costs.
This brings us to the concern in the RFI about effectively managing research programs of a given size. This reminds us of some productivity per grant dollar analysis pioneered by Jeremy Berg when he directed NIGMS.
We can no longer rely on NIH web pages staying up, but for now NIGMS then-Director Jeremy Berg’s blog post from Sept 27, 2010 is still available. I’ve grabbed the key figure, in case this eventually disappears.
His conclusion was simple:
Nonetheless, clear trends are evident in the averages for the binned groups, with both parameters increasing with total annual direct costs until they peak at around $700,000. These observations provide support for our previously developed policy on the support of research in well-funded laboratories.
That meant that more dollars meant more publications, of increasing journal impact factor, when PIs were managing (the equivalent of) three grants. In the modular limit days.
Using the CPI inflation calculator (instead of BRDPI), I make $700,000 in 2010 dollars to equal $1,050,716. Two grants worth, in the current era of proposals jail broken from the modular limit and pushing up against the $500,000 advance-permission limit.
One thing that makes is hard to know how to respond to the RFI is that the NIH is not great about providing information on the scope of the (perceived) problem. How many PIs have so many grants that they cannot manage and oversee the necessary research team? Anecdotes about highly salient BSD investigators abound. Angry talk from the frustrated grant seekers assumes a lot of such overfunded PIs exist, but don’t really have evidence for this.
Some time ago, the then head of NIH’s office of extramural research (i.e. the head honcho for grant awarding) Sally Rockey posted a “mythbusting” blog in which she presented the percent of PIs that held from 1 to 8 Research Project Grants (RPG). This category includes all R-mechs and not only R01. It turned out that from FY1986 to FY2009 roughly 90% of PIs had 1 or 2 grants. In FY2004-2009, only about 6% of all PIs had three grants. Only a tiny sliver is left in the 4 or more grants categories (and again, this is all RPGs, not just R01s).
Yes, it would indeed be nice to get some updated information on these kinds of distributions. I assume, in advance of evidence, that the breaking of the $250,000 limit dam and the continuing hellish competition for each new award pushed these distributions left. But we don’t know for sure until the NIH shows us their data.
Going back to the NOT language about PIs requesting what they need, POs awarding based on need, and this not being a dollar cap, well, it is pretty clear what PIs think is their requirement for “a grant”. It’s $500k direct. Just like it was in 1999-2000 when the modular limit was established. And not very many PIs are able to sustain more than this, year in, year out.
Two grants is pushing $1M in annual direct costs but that
So is the cap going to be one grant? That would seem necessary to really move the needle on grant distribution.
Or, is the NIH going to be told to limit award sizes of “a grant” in some way?
When funders interfere in the conduct of the science
June 1, 2026
it is unusual.
A recent article in Science Insider from Jocelyn Kaiser had a paragraph that caught my attention
An HHS spokesperson said on background that NIH staff are still making final grant decisions, and that requests like those coming from HHS are routine. “It is not unusual for NIH staff to suggest changes to research plans in the interest of strengthening new or ongoing research programs,” the spokesperson said.
This is highly unusual, in my limited experience. I have never had a PO suggest changes to our work under a R-mech funded project. I have certainly never had a PO demand changes to our work. Not before the funding of a new competitive award, not during the months of support and not following submission of a project report and before the non-competing award. I have never experienced the non-competing award being held up because NIH wants me to alter the research program.
I have only ever heard a PO make a suggestion about the course or conduct of a research proposal when I have sought advice about how to get a better score at complained about study section.
Admittedly, I have never been PI of a U-mech. The U01 cooperative agreement contains language that indicates NIH staff will be involved in the direction of the research. E.g., from RFA-NA-26-001:
there will be substantial Federal scientific or programmatic involvement. Substantial involvement means that, after award, NIH scientific or program staff will assist, guide, coordinate, or participate in project activities.
The NIH funds a lot of U01s so perhaps that is cover for the “not unusual” claim? Which now makes me wonder about the disclosures / financial support sections of any papers that result from U01 funding.
But this is not standard for R-mechs, as far as I know.
In fact I have published many papers which include a statement along the lines of “The NIH had no influence on the conduct of the research, the data analysis or presentation, the decision on when to publish or the content of the publications that resulted from their grant support.” I am not alone in this, by far.
I often read articles that have a disclosure with phrasing similar to “The content is solely the responsibility of the authors and does not represent the views of the funding agencies that supported this work.” This is less specific but probably implies the same thing to most readers.
Often, the inclusion of such a statement is mandatory for the journal to accept the paper for consideration, and for publication. The journal’s guide to authors sometimes even includes the specific phrasing of such a statement. In some cases I have seen journals adopt a specific heading of Role of the funding source in the disclosures.
The context of this is that journals wish to avoid publishing scientific papers that are more propaganda for the funding body (typically commercial ones, but it generalizes) than independent investigation. In my field, the tobacco industry is often referenced, given their history of pumping out “science” papers which happen to exonerate them of any health damage caused by their products. You might also think of the pharma industry supporting academic research and wanting to make sure only the stuff that supports their commercial interest is published, and anything against those interests is for internal consumption only.
Interestingly, the AI is giving me this:
But clicking the links fails to find any text which supports these statements on independence. One fears these sentiments have been recently scrubbed from the NIH website.
I don’t know how we can continue to publish statements of independence if NIH is asserting that NIH staff (or, really, HHS staff) habitually suggest changes to research plans and hold the funding until PIs agree to these changes.
NIH funding skips
May 26, 2026
I’m a proponent of the multi-tiered system whereby the grants that get picked for funding have been selected by multiple inputs, including peer review, program priorities expressed in funding opportunities, institute medium term strategies, emerging health issues, priorities of individual program officers and the will of the IC Directors. I have frequently made the observation for new and not so new comers to the NIH funding system that regardless of whether this should be the case, it emphatically IS the case.
Once NIH started publishing NIH-wide and IC-specific data on funding by percentile, long after then-NIGMS Director Jeremy Berg pioneered such transparency, it was easier to make the argument stick. These data from the FY2025 awarding of R01-equivalent grants are the most recent example.
Peer review has a major influence on what is selected. Always has. Grants that are in the very top percentile rank (lower is better in NIH parlance) are almost guaranteed to win funding. There is a percentile rank past which almost nothing is going to be funded. In between there is what we call the grey zone in which the probability of funding is a descending function of distance from the percentile below (better) which everything funds. The reasons for this pattern vary and individual Institutes and Centers of the NIH also differ from each other in their practices. Some ICs (famously NINDS, pre-2025 NCI) tend to stick very closely to the peer review scores, including by publishing a payline, i.e., the percentile score below (better) which almost everything funds and above which almost nothing funds. Other ICs (e.g., NIMH, NICHD) have a process which lets in other priorities and many of those therefore claim not to have a payline. These semantics, in my view, do not change the principle of a payline, i.e., a score below which almost everything will be funded.
Almost everything. I do not usually focus on the skips because they appear to be relatively rare. And we assume, perhaps wrongly, that these are cases many of us would agree with. A project that has been funded some other way in the time during which this decision process is made, either with NIH funding or otherwise. Violations of IC policies about total funding to one PI. Extreme cases of scientific overlap. PIs who, for whatever reason, have decided to quit being a NIH PI. Things of this nature.
The FY2025 data are, as we know, quite a change from past fiscal years in terms of where the apparent (implicit, aggregate) payline sits. The point including and below which almost everything funded was 4%ile. The same report for the FY2024 funding puts the critical percentile rank at 10%ile.
We’ve seen quite a bit of social media complaining about 2-4%ile grants that have not (yet?) been funded for FY2026. Some wag on bluski made the point that because of the random and subjective nature of peer review scoring it is not inherently more tragic that a 2%ile grant is not funded compared with those scored a 9%ile or even a 15%ile. It’s a multi-tiered system. Still, there is a certain resonance here for people who don’t think too hard about this funding system.
It struck me that not only had the payline dropped severely for FY2025 but that there may be more frequent skips. So I downloaded a few Fiscal Years worth of data and calculated the skip rate. Because the payline had dropped so far, I just considered those proposals that scored in the top 4%ile bins. In FY2022 3.2% of these proposals were not funded. It was 3.5% in FY2023 and 3.2% in FY2024. Looking very similar.
In FY2025, consistent with my eyeball assessment, this went up. Approximately double the percentage of high scoring proposals were skipped.
In FY2025, 6.6% of the R01equivalent proposals that scored in the 1-4 percentile were not funded.
This is the sort of thing that had me fearing that the NIH was not going to even publish these data for FY2025. After all, the less we can see in public about changes in funding, the less we can complain about it. To Congress. To the press.
Talk about how the NIH has previously been funding all the wrong stuff and moves to alter Advisory Councils and arguments for “geographic diversity” and arguments for and against various types of science from a political perspective have heightened concerns.
As has the decision to alter the review triage line to ~2/3 of submitted proposals and the nebulous designation of the middle third of proposals (based on initial scoring from 3 assigned peer reviewers) as Not Discussed, but Competitive. With the latter seemingly construed as easier for NIH ICs to fund. Based on various Program priorities.
Which is a process I just said I support, yeah?
Well sure. When the decision making process is based on balancing a portfolio, breaking through the inherent circular conservatism of peer review by the successful awardees, getting out ahead of developing / novel health concerns or scientific opportunities, etc, etc. I have little confidence I will be in support of a lot of decision making going forward, particularly that decision making that appears to accord with the political needs of the regime.
It strikes me that transparency may still be under threat. One of the reasons some of the ICs have always insisted they do not have a payline is because they did not want to have to get into arguments with disappointed PIs about why their proposal was not funded. When it is the rare 3% of proposals that are skipped within what appears to be a hard payline, well presumably the reasons most would agree with are sufficient. But in the grey zone across 10 percentile ranks where anything from 20 to 80 percent funds / doesn’t fund? That’s a lot of hard to make arguments with individual PIs about why their 14%ile was skipped and that other PI’s 17%ile was funded.
The regime is shaping up to be resting itself on a series of petards, just waiting to blow itself up. On the one hand they are insisting that “pure merit” should drive grant selection. This is when they are arguing that funding opportunities that mention diversity or research topics that they don’t like are inherently less meritorious. Merit, in this case, as defined by peer review outcomes.
On the other hand they are trying to argue that the way NIH has always done business is flawed. Peer review has produced outcomes they do not like, so therefore there is something wrong with it. Too many grants awarded to coastal elite institutions and not enough to institutions that reside in Red States.
Everyone suspects that we are about to see selection from a post-review Program approach that more aggressively meddles. All ICs are now disallowed from the strict payline approach. The idea of ND, but Competitive, is potentially going to “allow” the selection of grants at percentile ranks that would have been rare to win funding in the past.
I cut my teeth as an applicant with ICs that were opaque about their funding practices relative to grant scores. I cut my teeth in an era when NIH itself was not transparent with their funding, as per this report. We had to try to figure out what was what by comparing notes with other applicants about their scores and what POs happened to say to them. Usually off the record.
This brings me to a concern that was just expressed to me.
What if the NIH stopped telling applicants what their score was for their proposal?
We would have no grounds to complain or to go tell Congress what terrible things were happening. We couldn’t verify the degree to which new political priorities were altering previous scientific priorities.
Keep your eyes peeled, folks. And start thinking about what we should do if this comes to pass.
New reporting in Science from Jeffrey Brainard and Jocelyn Kaiser informs us of new policy that is apparently being communicated by POs in relationship to progress reports.
The takeaway message is contact your PO if you plan to involve foreign co-authors on papers which acknowledge NIH funding for support. Yes, even if that co-author’s participation was not compensated in any way by NIH funding.
The scope is very broad:
NIH officials are telling grantees who submitted annual progress reports for this fiscal year—which NIH reviews when deciding whether to continue funding for multiyear grants—to remove papers that name co-authors affiliated with foreign institutions if NIH had not previously approved a foreign component for the grant. Those co-authors could include visiting colleagues, students, or postdoctoral researchers temporarily working in the U.S.; overseas researchers who donated research material but didn’t take part in the research; and scientists who moved abroad after conducting the work in the U.S.
The piece links to a general NIH grants FAQ page search for “foreign component” which appears to hinge on “the performance of a significant scientific element of the NIH-supported project outside of the United States.” It then goes on to indicate a postdoc who works in a domestic lab with their entire salary paid by a foreign government would not qualify.
At this stage, since there is no new policy that is published and this is coming from POs telling PIs to remove the offending publications from progress reports it is very unclear what this means. This direction does not, seemingly, tell us we cannot publish with foreign-co-authors, it merely says we should not claim these.
Why?
Well, this sounds very much like POs telling PIs to remove certain words from their proposals and progress reports. To avoid triggering any delaying review from the overlain political monitors. But additional information from the article in Science warns us what may lie ahead for NIH funded institutions:
NASA has told some grantee institutions they may be in violation of the Wolf Amendment because their researchers co-authored papers with scientists affiliated with institutions in China, even if no NASA funding went overseas, West says. And, she says, NASA has informed some grantees that violators may be subject to a lawsuit under the federal False Claims Act, which prevents willful, fraudulent use of government grants and contracts.
Back to the NIH:
Units of the National Institutes of Health (NIH) are privately directing grantees to request permission in advance for any co-authorship with a scholar affiliated with a foreign institution, even if all the work was done in the United States.
And there we are in very uncertain territory. Will it be a simple email outlining the circumstances to which the PO replies “noted, sounds great)? Or will this inevitably trigger a definition of the project as involving a foreign component and therefore require a bunch of additional paper work be completed? New approvals for every different country that relevant postdocs happen to come from? In the ginormous labs creating Glam level (e.g. in Science) publications this could be participants from 3-5 foreign countries, no problem.
This is bad. Bad, starting with this reporting on the nebulous risks of involving non-US authors. It will have a quelling effect. No PI is looking for extra trouble. And this raises the trouble (meaning paperwork and hassle and delay) for involving foreign collaborators to a higher level. Which may not be worth the benefits. And, since we are assuming previously there was a good reason to involve the other authors, because of course there was, well this obviously diminishes the science output under that grant award.
I don’t necessarily object to notification of NIH when such collaborations are being first planned. If the only thing a PI has to do is tell the PO “hey, we’re thinking about publishing with Hilda Haaraldsdottir’s lab” and the PO says “ok, I’ll note that” and all is fine, well this is acceptable. Sort of in the same tone as that NIH FAQ page which seems to be emphasizing Other Support and overlap issues more than anything else.
I do object, however, to the way this is telling us that any collaboration with foreign colleagues has to be approved in advance, even when there is no NIH funding being spent outside of the US.
So it would be nice if the NIH would issue some clear guidance on how this is supposed to be handled.
Thoughts on each author taking full responsibility
May 20, 2026
A tweet from a person identifying themselves as a “Chair CS Section of ArXiv” has started a shit-storm on Twitter.
In case of the entirely predictable future, one Thomas G. Dietterich posted:
Attention @arxiv authors: Our Code of Conduct states that by signing your name as an author of a paper, each author takes full responsibility for all its contents, irrespective of how the contents were generated.
This was stated in the context of generative AI, the use of which led to mistakes, biased content, misleading content, in appropriate language, incorrect references, etc. The tweet thread also included a warning that the penalty would be a 1 year ban from posting to ArXiv for all authors on the pre-print.
The shit-storm seemed to focus on the charge of “hallucinated references” which then brought up the more general issue in science whereby citations in a paper may not be entirely correct and appropriate.
There was the usual face-off of normal scientists with the purists who cheered the policy, said that surely all authors should be equally responsible for everything in a scientific manuscript and that only egregiously bad scientists would claim otherwise.
At a very basic level, there is no point to scientific papers that have more than one author if this is the real standard. Because it requires that each of them do the experiments, analyze the data, generate figures, create the drafts, edit the drafts, etc, etc.
Because the only way to “take responsibility” is to not trust the work of anyone else. And to do it yourself.
This, of course, violates the most basic principle of the vast scientific enterprise and should not be supported* by anyone.
WE TRUST THAT OTHER SCIENTISTS ARE PRESENTING US WITH VALID DATA, ANALYSIS AND INTERPRETATION.
Sure, it is a “trust, but verify” situation**. We don’t have to check everyone’s work in real time, nor do we obsessively verify everything. But we repeat experiments, we trouble shoot methods, we think about results and outcomes and we try to muddle towards a better understanding of what is likely to be replicable (“true”?), what is limited by certain experimental circumstances ,what might be true in the sense that it occurred for this published work but does not generalize…or was a chance occurrence.
I cannot possibly watch my lab staff performing each of our experiments. I cannot be personally responsible that they indeed conducted the experiments as planned/designed, reported the data validly and warned of any problems that rise to the level of compromising the data.
Ok, I could. I could only allow experiments to be conducted in my physical presence.
But this is not how most science works these days. Science reports include the work of multiple people. Many of whom are credited with an authorship line. And they trust that the other person is doing their own individual job to the best of their ability.
If my part of the job is formal data analysis, creating figures and reporting the stats, for example, I can take responsibility for what I have done. If part of my job is to write introductory material and also to discuss the findings, I can take responsibility for what I have done. Including the citation of other papers.
Other authors on the paper could, if they so desired, replicate any and all of that. They have access to the data. They have access to the literature. They may not, however, have the skill set and acquired experience to do this.
Just like I have no idea how to implant an intravenous catheter in a rat.
They trust me to get it right. I trust them to get their part right.
The focus on citations, and the charge that only bad scientists would push back in any way over each author taking responsibility for every citation in the reference list, got totally ridiculous.
In part because the outrage and calls for punishment outpace the crime. By some margin. Because the fields of science that I am familiar with, and that get complained about in online discussion spaces, occasionally find fault with citations.
People get citations wrong because they make errors. Maybe they just clicked the wrong entry in their reference software. Maybe they cited a paper that they’ve repeatedly cited and never noticed it was finally retracted last year after a decade or whatever.
Sometimes those errors might be chalked up to laziness. Citing papers because some other paper cited it, without reviewing it yourself. Citing papers by Abstract, i.e., for those authors’ claims which may not actually be supported by the data presented.
Or maybe slightly more pernicious reasons for citing. Citing papers to curry favor with a Big Wig. Citing papers to seem like you are in the game in some way. Citing papers for reasons of personal (h-index, baybee!) or institutional (LOL check the practices of NIH intramural investigators) interest.
These are the kinds of citation practices slash citation errors that go on all the time. It’s not great. Nobody defends it as good practice when errors are made. But the harm is, quite obviously, minimal. If you, as the reader of a scientific paper, take each citation as Gospel Truth..this is on you. And you are evading the exact same “responsibility” that you are demanding of each author when you try to punish them for the little oopsies of some other author.
Perhaps my framework for what citations in academic science papers mean is old fashioned, due to me starting my deep reading of science in graduate school before there were effective and universal topic-search computerized databases. And certainly because it was before we enjoyed immediate access to copies of the papers without leaving one’s office.
Citations are there, in my not so humble opinion, to assist the reader in pulling the threads of scientific evidence. To assist them in finding other relevant papers which might inform said reader on the field in general, or on a specific topic. I understand the citation I view in a paper not as some authoritative Ground Truth, but as assistance to me to finding the relevant papers. Papers which it is my responsibility to read and interpret. And to judge for myself, as the reader, whether a given concept or understanding is supported by those additional works.
This underlines my recommendation to any newer authors about how to cite their own writing. One occasionally is asked something along the lines of “how many papers should I cite for this statement?” by trainees.
My default framework*** is that you should cite [First, Best and Recent] papers.
Scientific priority of the First to publish is not something that I have a whole lot of respect for but it IS something I have to acknowledge is important in many subfields and subcultures. It doesn’t have to be obsessive but it also tends to help the reader to show them where the subfield threads start.
Best, well, duh. If there is some paper in your field that just knocks it out of the park for the point you are making, cite it. Pretty simple.
Recent paper(s) is/are helpful because they have a tendency to also cite recent work and give your reader a foothold on where that subfield of investigation has arrived.
Can someone please train the AI in my [First, Best, Recent] rubric?
*It maybe obvious, but we should probably not be providing illegitimate fodder for the antagonists that wish to collapse the scientific enterprise.
**Including for our own work.
***framework. It’s a starting place. Not an imperative for every citation. Just in case any of those purist wackaloons venture over here.
NIH fills GMS labor gap with….postdoc volunteers?
May 18, 2026
I’m just noting this because it just seems so absurd we have reached this point.
Grants Management Specialists are the administrative professionals that handle the nitty gritty paperwork details of getting grants awarded to your institution once an I or C has decided to fund it. Under ideal circumstances you, the PI, would have no reason to know which GMS is assigned to your grant. Because in ideal circumstances everything goes smoothly, there is nothing for the GMS to request of you or your institutional grants office, etc.
This is true for new awards and even truer for non-competing continuations, assuming that your ducks are probably lined up most assuredly for the latter.
Well, NIH has been indicating that the normal operations of the GMS pool have been under stress. Yes, due to various slowdowns in NIH ICs deciding when to fund grants, or the political review interfering, which delays the normal pace of grants being ready for the assigned GMS to issue/award. But also because a lot of people got pushed out of service during the DOGE process and ongoing forced reductions.
You can check out the up to date status of the pace of grants issued in 2026 with the graph over at Grant Witness. Spoiler: Non-competing awards are still off the pace of 2025 (down 2804 awards, or 14% lower) which itself was off the pace of prior (normal) fiscal years (down 3,457 awards, or 16% lower) at this time last year. Reminder, NIH has had the same $ budget to spend in each of FY 2024, 2025 and 2026.
New reporting in Nature from Max Kozlov contains this little window on how bad this has become.
At least one of [It is NIMH] the NIH’s 27 institutes and centres has lost so many GMSs that it has asked early-career researchers, including postdocs and graduate students, who work in the NIH’s own labs to consider working temporarily as a GMS on a volunteer basis, according to internal documents, meeting notes and e-mails that Nature has obtained.
Wow. These are trainees in intramural labs. Not on the Program side, at all. Being asked to “volunteer”. So clearly they cannot be paid as outside contractors or similar.
Look, yes, the job of a GMS is about navigating the red tape involved with making an award that NIH wants to make and the institution wants to receive. It sounds simple. But these are contractual government matters that involve large sums of money. I am certain to my bones that the normal GMS staffer requires a lot of training. There was a stray tweet that flew by asserting the GMS are made personally culpable for screwups, potentially to the tune of millions of dollars. I hope this is all more threat than ever a reality but still.
This is unambiguously serious business.
Putting volunteers on the job is not ideal.
The piece goes on to detail knock-on consequences of this staff shortage including a prioritization of non-competing awards over new competing awards, the corresponding pressure to use multi-year funding (more money expended for the same GMS time, I assume), a deputy director of an IC being re-tasked with helping make awards, the suggestion at least one IC was down to 3 out of 11 GMS seats filled at one point, etc, etc.
Will you survive?
May 15, 2026
There was a distressed post that flew by on Blüski this week.
How are we to survive this?
How will academic scientists get through this chaos and assault on our profession?
You either will, or you will not. Some will exit the profession. Some will hunker down and survive. It’s okay to do either.
I used to say my one detectable talent for this business was my ability to take a punch.
I’ve had a lot of reason to doubt this over the past several years.
We take the punches until we cannot.
That’s it.
That’s the gig.
Grant review homophily
May 13, 2026
A small literature on citation homophily in science describes a bias in which white men cite papers authored by white men more than is warranted, and cite papers authored by women or authors of color less than is warranted. Obviously it is tricky to figure out how often any given paper should be cited, but for now let’s stick with some results from a partially overlapping set of authors regarding citations in a subset of “top*” neuroscience journals.
Bertolero and colleagues presented Racial and ethnic imbalance in neuroscience reference lists and intersections with gender in a preprint in 2020. So far as I can find, this was never published in a journal, for those who find this important. The critical categorization in the paper looked at the four first/last author dyads involving Authors of Color (AoC) or white authors. The darker tone distributions in the figure represent actual citations and the lighter tones depict predicted citations based on random draw models using a bootstrapping procedure, with the former normalized to the latter (hence the 0% central tendency).
These data show that White/White author dyad papers are cited more than expected and AoC/AoC dyads cited less. This is true regardless of whether the citing authors are white or of color, but the effect size is much much larger for white author publications. The outcome of this analysis was not improved by attempts to use the characteristics of each paper to match to a population of similar paper. The conclusion remains. Papers authored by white authors are cited more often than they should be, and papers authored by authors of color are cited less. This effect is primarily driven by the citation practices of white authors. In neuroscience.
Entirely unsurprisingly, this also holds for papers with man/man and woman/woman author dyads. Dworkin and colleagues published The extent and drivers of gender imbalance in neuroscience reference lists in Nature Neuroscience in 2020 (doi: 10.1038/s41593-020-0658-y). This panel from Figure 4 of the paper depicts temporal trends which give you a feel for how increasing proportions of publications that include women authors would be expected to increase the citations to women. This shows that if the citing papers have Man/Man author dyads, they tend to cite other Man/Man papers more than warranted. Correspondingly they cite papers with women authors less. As with the above analysis, the pattern of overciting M/M dyad papers is present, but much less in magnitude when citing papers have women authors. Interestingly the M/W and W/W citations rates are right at expected value when the citing papers include women in the first/last authorship dyad.
This brings me to a study of the review of research grants at the Veterans Health Administration. Boyer and colleagues published Analysis by Gender and Race and Ethnicity of Reviewers and Awardees for Intramural Research Funding in the Veterans Health Administration in 2023 in JAMA Network. These data drew from five review cycles for VHA grants from 2018-2020 and involved thirty six study sections, 664 reviewers (381 were women; 81 were racial or ethnic minority individuals) and 146 (77 to women; 81 to ethnic/racial minority PIs) funded proposals. Yes, it is frustrating that any outside analysis only has access to funded projects. So the study could only report on frequency of investigators by sex or representation status within the funded grant pool. There is no indication of success rates.
The authors were able to group the study sections into quartiles based on percentage of women reviewers, out of a range from 20% to 80%, or a median split based on underrepresented minority percentages (range 0 to 39%, median 7.3%). The frequency of women investigators in the subset of funded grants was five times higher if reviewed study sections with the top quartile of woman participation compared with the lowest quartile. Which, you can see from the Figure 1 from the paper, included quite a few study sections with 40-50% women. The analysis for underrepresented individuals found a three-fold increase in the proportion of funded grants in the subset of funded awards reviewed in study sections with minority participation in the upper half (8-39% URM) compared the lower half (0-6% URM).
Interpretation of the meaning of these studies, obviously, depends on a fair number of assumptions. About whether expected citations to papers can be modeled if your sample size is large enough. About whether the statistical models used consider the necessary factors when determining funding proportions. And whether we can tell anything without the success rate data.
But the whole Ginther / Hoppe / Lauer saga from NIH grant review sets up a pretty strong null hypothesis. I.e., that this reflects a genuine funding disparity for Veterans Health Administration awards that is driven by homophily bias whereby white reviewers favor proposals from white applicants and male reviewers favor proposals from male applicants.
We have not seen this sort of study section by study section analysis of the disparity of funding at the NIH, likely this will never be done from the inside so we can get Ginther style success rate numbers. One might have thought it would be a nice dovetail with the “topic based” analysis that arose with Hoppe and the excuse based on funding ICs championed by Lauer. It would have been interesting to know how the percent of Black PIs on study sections affected funding success. The aggregate percent of Black reviewers was 2.4% in the Hoppe report. My recollection is that I’ve only ever been on NIH study sections that were 3% Black as a low bound and mostly higher, ranging up to about 9% or so. Were these sections that exhibited less bias? Is there something systematic here in which there just tend to be more Black PIs doing reviews in study sections on selected topics, in this casue substance misuse disorder research? Perhaps so, given NIDA is the one IC out of the three or four that we have any data for that appeared not to have a Ginther gap problem (in one brief snapshot).
The VA study is also useful for making reasonable inferences about what it takes to change a bias. The lower quartile participation for women was at 50%. This is, in my experience, the high water mark for NIH panels. The median split for URM reviewers (all URM, not just Black reviewers) broke between 6% and 8%. I’d love to see the CSR wide histograms for reviewer participation and to see some analysis of how that does, or does not, predict application success rates based on the characteristics of the PI.
*Nature Neuroscience, J. Neurocience, Neuron, Brain, Neuroimage
Demographics of the PIs of suspended NIH grants
May 6, 2026
A new paper in the Lancet by Fielding-Miller and colleagues reports on the principal investigators of grants that were terminated from January to May in 2025. This includes terminations based on the specific content of the grants, the answered funding opportunity or identity of the PI as well as University wide terminations at Harvard University, Columbia University and Northwestern University. Their approach was to identify affected grants via the GrantWitness database, find institutional email addresses for the PIs and invite them to an online survey.
There were 1918 investigators with terminated grants who were approached of which 941 (49.1%) responded. Of the termination types, 56.2% of those with targeted terminations responded and 38.3% of the institutional termination PIs responded.
There is a lot in here so I will just hit a few highlights. The investigation compared the percentage of PIs meeting various demographic criteria against the percentage of all PIs funded in 2024 for the relevant grant categories. The first observation is:
Compared to the all NIH-funded investigators in 2024, investigators who received targeted terminations were more likely to self-identify as women (n = 406, 56.0% vs 42.4%), Hispanic/Latino (n = 122, 17.7% vs 7.0%), Black/African American (n = 86, 11.9% vs 4.0%), American Indian/Alaska Native (n = 19, 2.6% vs 0.3%), and Native Hawaiian/Pacific Islander (n = 5, 0.7% vs 0.1%) and less likely to identify as white (n = 484, 66.8% vs 70.2%)
We can express that as a grant being 1.32% more likely than the base rate of funded investigators to be terminated if the PI was a woman. The risk for grants with Black PIs was 2.98% more likely to be terminated. I remind you this is for funded grants. And we know from Ginther et al 2011 and Hoppe et al 2019 that the R01* proposals of Black PIs are only about 60% as likely to gain funding in the first place.
The paper also addressed the disadvantaged socio-economic background criteria that used to be under section c. in the NIH’s long standing statement on their Interest in Diversity (now rescinded, see this blog post for reference to the criteria and the wayback machine link). Respondents were asked if they qualify. It turns out that grants with investigators who had disadvantaged backgrounds were 2.54 times more likely to be terminated compared with the grants with non-disadvantaged PIs.
The overall picture did not change when “restricting the analysis to exclude investigators whose only terminated grant was a diversity supplement,“, i.e. this was not only about the funding opportunities that were explicitly about increasing workforce diversity.
Also, because of the demographics of NIH funded investigators at Harvard, Northwestern and Columbia, about 20% of investigators who had grant terminations justified in an Executive Order on the grounds of combating antisemitism at the institutional level were themselves Jewish.
There’s more in the paper, and you should read it, but the takeaway verifies that which we already suspected:
Investigators in our sample who received targeted terminations were significantly more likely to be women, Black, Hispanic, American Indian/Alaska Native, and Native Hawaiian/Pacific Islander than the publicly available demographics of 2024 NIH-funded investigators.
And that’s great for us that the authors have provided a data based account of what happened in 2025. We have something citable and specific. Even if we suspected that the topics and funding opportunities targeted for termination would just so happen to be correlated with relevant PI demographic groups. Suspect because this is related to what was shown for bias against the funding of R01 proposals on topics of disproportionate interest to Black PIs in Lauer et al 2021 and indeed the fact that Black PI proposals were less likely to be selected for funding across the quintiles of topic-associated funding rate as described by Hoppe et al, 2019.
I do want to end with a caution. It is tempting for people to promulgate a sort of top-level, executive summary from this. Such as the following from the end of the Results section:
Among our sample of investigators with terminated grants, BIPOC women and trans/non-binary people had nearly three times higher odds of having an equity-related termination compared to White cisgender men (aOR 2.69, 95% CI: 1.71–4.25)
I saw a popular science blogger/tweeter type post something along these lines today, with a tweak that misrepresents the findings.
But it has been suspending NIH grants to women and minorities at three times the rate of neutrality.
In fact the rate was 0.32 times for women and 3 times for Black PIs, presumably the overall BIOPOC stat was likewise about 3 times. This is a 10-fold higher hit for PIs of color than it is for women in general.
*The statistics in this paper appear to include every mechanism of grant award including research grants, fellowships and supplements.
Jeremy Berg posted a word cloud analysis for terms that appeared in the title of NIH grants last year but have been removed in the latest non-competing interval. N.b., this excludes any grants which have not (yet?) been renewed.
I have no idea how unusual this is for the NIH as this is a new consideration for me. I don’t believe anyone had any reason to even suspect this occurred with much frequency in the past. In my experience grant titles are never changed within the non-competing interval. Nor are Abstracts, even if the Progress Report indicates that there has been a change in the Aims, something which is perhaps not common but is far from rare.
Changes in the title from one competing interval to another, should one be fortunate enough to obtain a competing continuation, do occur but mostly just reflect a change in the scientific focus as a natural progression of the work previously accomplished.
As noted in Berg’s Alt text,
“The most commonly removed words are “equity”, “diverse”, “disparate”, “racial”, “gender, and similar terms.”
This analysis does not prove cause, although every bit of chatter suggests that Program, broadly construed, has been requesting these changes be made prior to issuing notices of grant award.
The Administration claims, via Director Bhattacharya and others, that there is no list of banned words that, if included in a proposal, would lead to it being prevented from funding.
Maybe technically true in some way, or maybe not. But this analysis sure suggests there is a list somewhere of words which are likely to prevent, or significantly delay, the issuance of a notice of award. Program Officers are communicating that to PIs and requesting/demanding that the titles be altered.
The chatter on bluski included a position that this is all no big deal. That altering the title words for form’s sake has no real impact on the science that will be conducted and that this is all just routine as long as the funds are awarded. One individual called it “transactional”.
I think this is naive.
I don’t know exactly what was meant by this but my longer term readers will be thinking about a constant mantra around these parts.
A grant is not a contract.
This is intended to remind people, for various circumstances, that one is not bound by chains to conduct every experiment mentioned in a R-mechanism NIH proposal. One is expected to make progress under the proposed Aims, yes, and that may include doing what you said in some circumstances. In other circumstances, changes in the advance of science in the PI’s lab or in the field at large may dictate doing something not mentioned in the proposal itself.
There is some grey area here, for sure. Program is touchy about PIs veering too far outside the scope of what has been peer-review approved but “too far” is a very, very VERY slippery concept. Statements in the Progress Report may be viewed with different levels of seriousness by PO and PI if it ever came to a debate.
PI: “Hey, I mentioned we might do this cool new experiment in one sentence in the progress report and you didn’t say anything and grants management issued the Notice of Award so that means we got Program approval right?“
PO: “Wait, you didn’t say you were going to totally throw every resource this year into what I see as a totally novel direction!”
There is some poorly understood black and white as well, such as that you cannot spend grant funds on research using a vertebrate animal species not mentioned in the original proposal unless you do a bunch of paper work with OLAW and your PO requesting a change.
This is under normal times.
We are not under normal times. Grants are pulled individually and University-wide for various offenses against the policies of the regime. Long court battles ensue and even if the University wins, much time is lost.
I think it naive for anyone to think that a PI who was asked to remove the above listed target words from their proposal is just being asked to change words. They are being told they are not supposed to be working on their research from a specific set of angles and perspectives.
I think it is very likely that most PIs in this boat are in considerable discomfort and indeed fear about how they should proceed. Fear that the usual flexibility afforded to PIs about what they work on under a NIH grant may be suddenly violated in specific cases. Fear that anyone who only “transactionally” alters their title or Aims and continues to work on, oh I don’t know, gender affirming hormonal therapy or racial equity of health care of refugees, may be sued by the government to replace the allegedly misspent dollars.
It is spectacularly tone deaf to ignore these as legitimate concerns.
UPDATE: Sigh. I was kindly reminded of NOT-OD-26-007 issued last November which serves as a:
Reminder to NIH awardees that changes in scope represent new terms and conditions, with which recipients must comply.
It goes on to define a change of scope as:
a change in the direction, aims, objectives, purposes, or type of research training, identified in the approved project. NIH GPS, 1.2. This can include changes to Aims, Objectives, Titles, and Abstracts.
So yeah, a change of Title is a change of scope which become(s) new terms and conditions of award, and the recipient must comply with those changes. And in case you are really thick, “This includes changes in scope that were renegotiated to align with the agency’s priorities.“
On reforming the NIH
April 29, 2026
Yes, I realize I’ve spent the past 19 years yelling into the void about various ways that the NIH should, in my views, alter the way it does business. On Feb 8, 2007, I observed:
It is dismaying to realize that by the time he received his first R01 (the major NIH research grant) Mozart would have been dead for 7 years (tipohat to Tom Lehrer). The official noises coming from the National Institutes of Health, and even some individual institutes such as the National Institute on Drug Abuse (scroll for comments on the young investigator) are positive, sure. We’ve heard such sentiments before, however, and most objective measures show long, uninterrupted dismal trends for the young and developing scientist.
…and things went downhill from there.
Today, however, I’m focused on comments on the bluski from one of the “Abundance” authors and podcast participants I mentioned yesterday. I was having trouble with their assertion that there were only two poles of opinion on the NIH- DOGE destruction or those that did not permit any criticism of the NIH and pretended it was perfect in every way. For the obvious reasons this didn’t sit well with YHN, given my credentials as a complainer about the NIH. It also didn’t sit well with many NIH-interested folks on the bluski.
On prodding, one Derek Thompson first insisted that he didn’t mean us. Of course not. “i never said “SCIENTISTS don’t want to reform the NIH.” that would be fucking stupid. scientists love complaining about grants and bureaucracy, and Bsky complains about … everything! the problem is that the NIH itself is resistant to large-scale reform.“
Glad he clarified. It certainly did not come across in the transcript of the podcast.
The discussion finally arrived at this:
which, in the event bluski craters, reads: Well, I wrote a book chapter after talking to dozens of scientists and the NIH itself about how the NIH should reform—to experiment with new grant approval methods, to award more funding to younger scientists and newer ideas, and to reduce grant-application burdens—but thanks for your feedback.
Finally something actionable. Criticism topics beyond mere “paperwork”. And it is just as absurd as you imagined.
The community of extramural scientists has discussed these topics ad nauseum of course. From the fate of younger scientists, to frustrations about getting “newer ideas” funded (ask me about my hot off the presses summary statement) to potential strategies for reducing grant application churn, we’ve gone over it all.
However, despite my grouchiness, one has to admit that the NIH itself has frequently discussed these problems and done things to try to change. The fate of newer/less tried/unfunded investigators has a long and storied history, NIH even has a summary page. New ideas have been serviced with Director’s New Innovator grant this and Stephen I. Katz award thats. To stabilize funding we’ve seen MERIT and MIRA and PECASE, oh my. Policies to limit the number of award dollars to any given PI have been announced. and practiced, at least in some ICs. I’m not sure what “new grant approval methods” means. If it means strict payline versus Program pickup behavior, well, this is already represented. Grant lotteries? Well, I don’t know for sure but this probably has been considered within the halls of NIH- of course this one runs into Congressional language and is not simple to change. (Lottery schemes are stupid and nobody really means a true lottery anyway.) Could happen, probably been discussed once or twice. Grant-application burdens…well, one might see the A1/A2 policy machinations as attempting (badly) to address this.
I AM going to give this guy a partial pass about interviewing NIH folks, particularly on the record. I have frequently mentioned my frustrations about NIH folks expressing what seems like the company line, believing in their own…smoke….a little too much…zombie mantras about revising grants…pretense about non-paylines from those Program Priority institutes….saying “we can’t” in response to many “we don’t wanna” issues. and my outrage about various grant review issues where NIH folks who should know better express beliefs that are in flagrant opposition to the experience of nearly everyone after their first one or two study section experiences.
But still. It is just plain ridiculous to stick to this straw man that every supporter of the NIH within the NIH refuses to entertain even the slightest critique of how they do business. Or to refuse any measures of reform to try to do their business better. We can add SABV policy into the mix here, in addition to the above-mentioned issues. And Collins’ creation of the National Center for Advancing Translational Science. And Congress backed initiatives (hint, they do not invent this up themselves without NIH input) on the BRAINI or, some time ago, HIV/AIDS research.
It happens all the damn time. With much of this in public, on the public record, easily viewable and understandable in the history of how NIH has operated.
Podcast doods operate on the perspective of the moment. They have no need to understand historical context. They have no requirement to justify their critique of the hour or, more importantly, to address which interests will pay the price for prioritizing their own schemes.
Take the fate of young investigators…it SOUNDS great to complain about how terrible it is that some can’t get funded or that the age of first major award hasn’t changed in decades from over 40. And to bleat about how ‘new ideas’ are somehow unique to the younger folks. But do we ever get serious input on how many new PIs we actually need? or what happens five years down the road when we fund all the ESIs who apply? or analyze how many truly new ideas gain traction via funding an ESI versus funding an existing vibrant lab on their new ideas that can’t get funded?
Or take childhood cancer- OMG it’s just terrible and we haven’t solved that yet, give more moneeeee! Of course, NCI already gets the lion’s share. If we devote more of the zero sum pool to this particular health condition, we take it away from elsewhere. Like Fetal Alcohol Syndrome. Or substance misuse in parents or during pregnancy…which affects other children adversely. Or, um, childhood vaccination against infectious disease. ahem.
Everything is a tradeoff. More money for heart disease? Less money for multiple sclerosis. More money for AI eleventy to do…whatever? Less money advancing promising drug leads to the point where a company might choose to pick up the ball and advance it to an approved medication.
Etc.
Everything has a tradeoff.
And recognition of competing interests and ideas about what is the most important thing has a tendency to look to the podcaster critique of the moment like intransigence. As if the NIH should not be criticized.
But that’s not it. The frustration is with Johnny-come-lately types that cannot understand that their certainty about priorities is discordant with good functioning of an agency that is supposed to be addressing the concerns of all Americans.
It isn’t that NIH can’t be criticized.
It’s that your criticism is poorly informed and displays a lack of consideration of inevitable tradeoffs.
Interviewing “dozens” of the PIs who submitted the 62,592 research project grant proposals for FY2025 does not even remotely bring one up to speed on the diversity of viewpoints and priorities that are in competition for limited funding. Only 8,161 (13%) of these were funded, awarded to 10,090 PIs (a third of which were first-time awardees). This represented about 20% of the 49,609 unique PIs who applied for funding in FY2025. How do we evaluate the opinions of the 39,519 unsuccessful applicants against the opinions of the successful? How do we decide that one-third of awardees being first-timers is too low?
How many of the disappointed experienced applicants are at the end of their first interval of support, and only a mere 5 years ago were the new and untried investigators podcasters now purport to be so concerned about? Do they really think that the way to advance science is to continually fund untried, new investigators for five years and then boot them off the rolls so we can get in the next new penny?
I dunno, guys. Maybe I’m the problem because I think the NIH does some really good things, has a demonstrated history of activities that lead to almost miraculous advances in human health and necessarily operates in a tenuous tradeoff of priorities and investments. Maybe my reading of science history that says it is an inferior approach to try to engineer efficiency in advance by picking winners from too narrow of a perspective makes me a knee jerk defender of the status quo.
On criticizing the NIH
April 28, 2026
Someone on the socials tagged me in on a few comments about a NYT transcript of an Ezra Klein podcast. Apparently he wrote a book about “Abundance” and somehow this relates to thoughts about re-configuring the NIH. I had a quick scan through, didn’t see anything before I got tired on my phone, and couldn’t immediately work out what Abundance means or how this related.
Later I had time to get to it on a computer and search out NIH (nothing. N.I.H. did work.).
The blog host and, I guess, his co-author of course set up a criticism of convenience in support of their blather about how they have the one true solution.
It starts with the premise that NIH needs some shaking up and gee isn’t it a shame that NIH is uncriticizable in polite company.
Consider three approaches to the N.I.H.: a pro-establishment liberal approach, an anti-establishment MAGA approach — which we’ll call just current policy in 2026 — and an anti-establishment, abundance liberal approach.
No worries, they know all about the current regime.
The current anti-establishment, MAGA approach essentially says — for a variety of reasons that are too complicated for me to go into right now: We hate universities, we don’t trust scientists, and we really don’t like mRNA vaccines. So we’re going to attack the universities. We’re going to destroy a lot of their scientific programs. We’re going to cut the N.I.H. grants by billions of dollars and also basically defund mRNA research because Robert F. Kennedy Jr. and Donald Trump don’t like it very much. That’s catastrophic.
Fair enough. They view themselves as good guys, despite expressing admiration for Elon Musk earlier. They purport to believe in the aspiration of what NIH is doing, or supposed to be doing. But they want to be significant critics and reformers so….
The establishment approach would be to say: The N.I.H. spends $40 billion a year and is the jewel of global biomedical research. It is one of the most important, successful institutions in America. You cannot criticize it; you cannot touch it. It exists on a kind of spectral plane that we simply cannot broker any criticism of.
mmhmm. I’m getting a distinct whiff of straw here. In what sphere does anyone operate where the NIH can’t be criticized? Are they unfamiliar at the very least with the Golden Fleece award from Senator Proxmire? And the succession of generally right wing Congress Critters who took aim at the NIH now and again as their target for reducing government expenditure? They definitely do not interact with many scientists if they believe this is true.
Here’s their genius criticism, which nobody would allow in polite company so thank goodness for their podcast. and publisher, apparently.
But then you come to category No. 3, and the abundance liberal approach….it says: You know what? Current policy is horrific. What’s also quite embarrassing is the fact that, according to their own testimony, American scientists who are funded by the N.I.H. spend up to 40 percent of their time filling out paperwork. These are the smartest people in the world, whom we entrusted with coming up with the most important breakthroughs about the cosmos and the human body and curing diseases. And what do we do for almost half of their time? Force them to check boxes. That’s a failure, and it’s a failure that we inscribed with decades of cover-your-ass rules that force scientists to essentially become bureaucrats.
Current NIH policy is horrific. In many ways. Past NIH policy, before 2025 was horrific, in many ways. And if these folks would talk to actual scientists who work under the NIH extramural granting system, well, they would get an earful.
However. “Paperwork”? “box checking”??? We waste 40 percent of our time on this? I don’t know who they have been talking to but I suspect they have mistakenly conflated a real complaint with something they can understand. I suspect that the scientists who complain about 40 percent of their time taken up with non-science tasks are referring to the grant chase. The amount of time devoted to writing and submitting grant after grant after grant for very low success rates, just to be able to do their work.
Perhaps I am projecting but trivializing NIH funded scientist’s major complaints about the system as complaining about “paperwork” seems wrong to me.
They then go on to insist their goal here would be to ask:
What do we want to accomplish with the N.I.H.? Don’t we want an abundance of scientific breakthroughs? Isn’t a good way to do that — to unleash the productivity of scientists and unburden them from some of the paperwork requirements that we’ve added in the last few decades? Let’s find a way to allow scientists to be scientists by reducing that burden. That’s an approach that I would like to see a “good DOGE” lean into in 2029.
I dunno, unique genius guys, even if the complaint is about “paperwork”, believe you me that nobody would hesitate to sound off on the NIH system about this. Is “paperwork” the annual progress reports? Are we to be freed from that from some future “good DOGE”? GREAT! Are they suggesting that IDC rates be increased even further so that PIs can get proper administrative support to fill out travel reimbursement reports? GREAT!
But it cannot be about this. It just CANNOT be “paperwork”.
The strangling of scientific productivity is because of the grant chase.
The need to submit many, many proposals just to land one research grant and the need for preliminary data for those endless submissions. The relative uncertainty of continuing research programs, the hoarding of precious funds, the reluctance to take on employees/trainees, hesitancy to publish until a grant has been awarded… etc.
Fix that with your “Abundance”, guys.
Acting Director Bruce Reed posted a blog post on the consolidation of all NIH grant review within the Center for Scientific Review (CSR). There are some tidbits of interest including:
In 2024, CSR reviewed 66,697 applications….With consolidation, CSR became responsible for managing the review of about 30,000 additional applications per year.
How did they accomplish this? Well, by leaning on the people previously doing the job within the IC study sections.
CSR expects to create 52 new chartered study sections to manage the review of R01 applications previously handled by funding ICs. These study sections are essentially being created out of existing study sections, which are now oversubscribed. Overflowing study sections will be divided to cover different, scientifically coherent sets of topics.
CSR is recruiting reviewers who had previously served on panels convened by funding ICs to retain valuable expertise…Most of the review staff who were at IC branches are now staffing CSR meetings.
Pretty reasonable stuff. This apparently says they are drawing in experienced SROs and the same super set of potential grant reviewers to cover the same old topics.
So far, so good.
There’s some blah-de-blah about making review more consistent via minimizing “significant local variation in review practices” without specifying what that may mean. Standard guidelines….training…yeah yeah, sure.
Here’s the howler slash head scratcher. One of those things that makes you wonder if possibly the longest running, most experienced and most senior person left at the CSR has any idea how grant review and award actually works.
Having review meetings with applications grouped according to science, rather than the IC that administers them, promotes appropriate competition. Better competition should ultimately result in better review outcomes.
Huh?
Let’s see how this works with some closely-related ICs and study sections pulled totally at random.
We already know that the supposedly already-in-progress ENQUIRE process reviewed NMB, NAL and BRLE this past cycle, and decided to keep BRLE as-is, but to revamp NMB and NAL. From my point of view of having a lot of proposals reviewed in BRLE, some in NMB and a handful in NAL, as well as being empaneled for a term of service on BRLE way back when, I cannot help drawing the direct connection.
NAL stands for Neurotoxicology and Alcohol but it has always, to my limited view, had a big domination of alcohol-related proposals that would be funded by NIAAA. In fact I once learned the hard way that this panel did not in fact actually have interest in non-alcohol neurotoxicology. NMB (Neurobiology of Motivated Behavior) has typically been dominated by proposals that would be funded by NIDA. BRLE, at least in times past, has a bit more of a diverse mission. When I was on it we seemingly got NIDA and NIAAA directed proposals, as well as a fair number from other ICs such as NIMH, NICHD, NINDS, what have you.
NIAAA also has maintained AA4 (NEUROSCIENCE AND BEHAVIOR) as an in-house study section. I never was very clear on why they did so and why some proposals would be reviewed there instead of, say, BRLE, NMB or NAL.
I don’t think NIDA had any lasting generalist panels like AA4, but they sure did have SEPs convened for targeted FOA/NOFO. It probably worked out about the same.
There is a certain species of my colleagues, and I love and sympathize with y’all, that expresses very strong opinions about how those guys over there get sweetheart insider club review and funding privilege. While they themselves, of course, are subject to true and competitive merit-forward review.
In general, the in-house reviews (standing IC study section or IC SEP) are more likely to be accused of inferiority, lack of true competition and a smooth path for the insider club.
Within the CSR panels, well, there is some of this kind of grumbling based on the degree of IC capture that is represented with an added dose of complaining about “gatekeeping”.
I have colleagues on both sides of the divide. They are not shy when complaining about their latest grant disappointments.
The funny part is that it is all true. And it is all not true. In approximately equal measure.
Review is highly subjective. Bruce Reed of all people should understand this by now. The major driver is the attitude of the assigned reviewers and the panel at large regarding the importance / significance of the work. And people have a pronounced tendency to think the work they do is the most important. Talk to a NIAAA insider about how important it is to cure alcoholism and alcohol abuse. Or to a NIDA person who is, er, cocaine forward, about basic mechanisms of substance misuse and the way to make “real” progress. Or to a NIMH funded person about research funding for their favorite mental condition that doesn’t happen to involve substance abuse.
They may even quote you some epidemiological and funding stats that “prove” their position.
So how does Reed’s “appropriate competition” work in these more generalist* panels? As per always, our default hypothesis has to be one of percentages. A panel that is 80% “NIAAA folks” will give higher scores to alcohol proposals. A panel that is 80% “NIDA researchers” will score alcohol proposals lower.
How does this help respective Program officers? NIDA isn’t looking to fund any alcohol-focused work and NIAAA isn’t looking to fund a cocaine-only proposal. It is not necessarily “better competition” OR a more useful review outcome to pit these against each other all of the time. Who cares if the 1-5%ile is all NIDA proposals and the NIAAA ones score from 6-10%ile?
This is likely to occur, broad strokes. Statistically speaking. Particularly when one IC is smaller than the other, thereby generating fewer applications, thereby requiring, on the face of it, fewer peer reviewers on one of these newly generalist panels.
Don’t get too smug, NIDA-folk, NIMH is larger than us and the same dynamic could quickly put us on the back foot. And let’s not forget about Big Daddy NCI’s tobacco addiction portfolio.
The issues that are actually important to, e.g., NIAAA Program Staff for making funding decisions may not even be rigorously addressed, if the panel is basically responding “meh, it’s alcohol, pass“. Right? Sure, they can pull up anything they want from the ND-Competitive pool but these will not have the benefit of discussion or panel vote.
Look. I favor generalist study section panels that tend to get a diversity of IC assigned panels. I thought my formative years on BRLE were frikkin awesome in terms of the kind of cross-pollination of reviewers and proposals that occurred. The focus then was on behavior, regardless of the broad spectrum of disorders to which is was applied. But I always had the suspicion this only worked because there were more specialized / IC-captive CSR study sections as well as the in-house study sections.
Regardless, it is not trivially obvious that this more general competition will result in better review outcomes, in so far as the primary outcome is SUPPOSED TO BE assistance to the Program staff in deciding what to fund.
*Neuropathophysiology of Addictive Substances seemingly replaces NAL; Neural Basis of Motivated Behavior seemingly replaces NMB. Keep your eyes on them for rosters and, ultimately, funded grants to see how this re-organization plays out in practice.
The response of many Boomers, particularly those who went through higher education in California, to talk of ballooning student loan debt and the unaffordability of University education is often a lack of sympathy. Mostly due to this.
“I paid my way through college working minimum wage jobs so these damn kids these days are just spoiled. Get to work!”
They, as is very typical for Boomers, conveniently forget the investment the prior generations made in them and how their selfish Reaganism (taxes bad, government bad) sympathies withheld similar support for subsequent generations. It worked a bit like this, yes, even in allegedly left of center, education favoring, high tax California.
For today, we are going to ignore the major investments that built the fabulous University of California, California State University and the Community College systems.
We can start with this constant dollar chart of the Tuition and Fees charged to undergraduates at the UCs and CSUs over time. It is very obvious that costs were a lot lower from 1965-1980 than they were in 2010. You will recognize this interval of time as prime Boomers going to University years. In fact it is almost comical to see the costs start to rise literally after Reagan was elected President and then to bump up again when the Clinton Presidency (and interestingly the economic boom of the tech revolution) turned the Democratic party fully to their “triangulation worked so this must be ground truth” adoption of the central premises of Reaganism, i.e., that taxes and government spending were bad things.
We’ll get to that part in a moment.
Meanwhile, the same selfishness of the Reagan era and the Reagan/HWBush recession put the hurt on compensation for minimum wage jobs. Here we can see both the Federal and California Minimum Wage standards across time. This is represented in constant 2011 dollars to match up with the static data on Tuition and Fees, above. First, California and Federal were pretty much aligned from 1954 to 1998, and it is only after that when California started treating its lower compensated workers better. (N.b. the grey lines represent local municipalities adopting minimum wage standards even higher than the California statewide standard.) It is pretty clear that from 1965-1980, when University costs were low and relatively stable, minimum wage workers enjoyed the highest level of compensation. For example the adjusted CA minimum wage rate of $9.47 per hour in 1979 was not regained until it hit $9.40 in 2018.
For comparison’s sake, adjusted CA minimum wage was $8.00 in 2011, when the graph of Tuition and Fees, above, ends. Minimum wage was 85% of what it had been in 1979, whereas the costs at a CSU had ballooned to over 6 times what they had been. 600%.
This is the first fact-smack you need to apply to your local Boomer or other person who should know better. Minimum wage was down 15% in 2011. It was $6.70 in 1987 and in $6.43 in 1995. That’s 71% and 68% of 1979 compensation for minimum wage jobs. You will recognize this interval of time as prime GenX going to University years.
The conversation likely then drifts into blaming the victim territory by bemoaning why University Tuition and Fees are seemingly ever increasing. Take a pause for the part where you have to explain inflation to your Boomer. And try not to get sidetracked discussing how a given student is not responsible for the general competitive trend where Universities think they have to provide more and more luxury and services.
This brings us back to the fact that generations prior to the Boomers chose to support higher education with their tax dollars riiiiight up until, you guessed it, Reaganism.
The Public Policy Institute of California reports on higher education support in California and leads with this graph on the amount of money California has spent on UCs and the CSUs out of the General Fund. I.e., out of taxes. I am not positive but this appears to be in constant dollars and not inflation-adjusted constant dollars. Which, to belabor the obvious, means the decline is actually worse than depicted here. According to the national CPI inflation calculator that ~$10,000 per CSU student in 1986 equals $20,930 in 2012.
CSU spending per student was relatively constant from 1980 to 2002. Constant here represents a decline in spending power and sure enough if you look at the first graph, CSU tuition and fees were increasing in constant dollars across this interval.
Spending on the UC system was increased in the late 1980s (cushioning the impact of the Reagan recession) and continued until about 1991 or so. You can again look to the Tuition and Fees chart above and see how the constant dollar costs to students started to rise when the State investment declined. Another rise in California State investment in the UC budget that peaked around 2000-2001and fell thereafter was matched by another stabilization and then steep increase in student costs. State support continued to decline from 2001ish to 2012 in nominal dollars and costs to UC students in constant dollars continued to rise. According to the national CPI inflation calculator that ~$25,000 per UC student in 1986 is equal to the spending power of $52,325 in 2012.
We don’t know what UC and CSU got from the CA taxpayers in the 1970s. But in the early to mid 1980s, as the last of the Boomers were attending University, UC was getting 2.5 times as much per pupil from the taxpayers as in 2010. This is in nominal dollars. So you have to increase this to account for inflation.
The difference was made up by charging the students more. A LOT more. Five times as much in tuition and fees, in constant dollars.
While they were able to only earn 85% as much in a minimum wage job.
Any conversation about how we return to a decent society can look to the past in part, a la making America great again. And it was indeed great when higher education was affordable, and students could work their way through college with entry level jobs. But that was only possible because taxpayers agreed to support institutions of higher education. And felt that educating the youngsters was a high priority. And were grateful for the taxpayer supported benefits that they themselves enjoyed and are enjoying.
No, there is not a “double bind” for women in NIH funding
April 10, 2026
I can’t locate it right now but a recent blueski conversation about geographical distribution of NIH funds that touched on the Ginther gap, resulted in (to my poor memory) a throwaway about how Black women surely have it even worse.
Of course I don’t know the breadth of the context intended. But I often get this assumption/question when I introduce the Ginther Gap to academic audience. In short, Ginther and colleagues showed in 2011 that the NIH grant proposals with white PIs had a ~1.7 fold advantage in funding and Hoppe and colleagues followed this in 2019 to show that a subsequent sample of grant applications showed the exact same disparity.
I can’t seem to find any mention of the answer in a quick search of my blog archives. However, Ginther and colleagues examined the fate of grant applications in their sample by race, gender and PhD vs MD of the PI in a followup paper published in 2016. So this is a good opportunity to mention their findings.
The answer to “Is There Evidence of a Double Bind for Women of Color?” raised in the article’s title was pretty emphatic.
No.
Figure 1 tells the tale. The upper panels show that per-application rate that is the subject of the major outcome in Ginther and Hoppe. The lower panels show a per-applicant rate that differs slightly from the dodge promulgated by Lauer and co. In this case the per-applicant rate is calculated as the probabability of a PI who applied in the 2000-2006 interval (the sample for the Ginther et al 2011 analysis) having received at least one R01 award from 1980-2006. No matter how the data are sliced, Black women are not at a substantial disadvantage relative to Black men. The biggest gap is apparently MD-holder application award probability.
The paper gives more breakdowns in Table 1. The male/female difference in award probability for Black PIs is negligible for both new and experienced investigators. Of new investigators under the age of 50, Black PhD women had a single-submission R01 award probability of 51% versus 37% for men, a significant difference.
Note that the male/female differences for other racial / ethnic categories were also minimal.
The point of this is not to dismiss the feeling women get about being disadvantaged in the NIH grant getting game. Especially women of a certain age who got their start, or hit mid career, before much of the data that we have readily available to us.
The point is use other successful examples to design remedies for persisting problems with grant award. In this case, there was previously a funding disparity for women and it was fixed. OK, the gap was closed. There are still some disparities for renewals and big mech funding that have come up in more recent times. But as I pointed out in at least one prior blog post,
something changed in 2003. All of a sudden a sustained advantage for men disappeared.
I don’t know what changed. The abruptness of the effect and the persisting success rate of women juuuuuuust below the rate for men year after year for fourteen years* does make me suspicious about some sort of corrective process rather than a fix to the systematic bias. I.e., NIH ICs making sure to balance success rates across PI gender when they make their overall funding decisions. Otherwise wouldn’t we expect the small difference to be more randomly distributed across the 14 years?
Even these data on the left side do not reflect what I assume was a much larger prior disparity for women. There was a time in the past where a group of women sued the NIH over this. It resulted some sort of agreement by the NIH to change. To enroll more women on study sections, for example. A LOT more women. Women comprised** 27% of standing reviewers and 24.5% of non-standing reviewers in FY2000. This rose to 32.9% of standing, and remained at 24.5% of non-standing in 2004.
I don’t know that I’ve seen comprehensive stats since then but I think somewhere I saw more recent data showing this pushing up close to 40%ish. For some reason the RePORT page for peer reviewer stats is rudimentary.
A quarter to a third to forty percentish is enough to Do Something. To Move the Needle. To oppose a systematic bias associated with male preference for the scientific work of other men.
In contrast, seeking a target of Black study section members that matches the percentage of Black PI applications for R01 funding (1.4% in the Ginther sample, 2.3% of only white+Black in the Hoppe sample) is not sufficient to budge any bias needles on study section. Not even if we assume*** Black peer reviewers on study section will express a bias in favor of Black PI applications that is equivalent to the bias of white peer reviewers that is connected to the 1.7 fold advantage for white PI applications.
Women were PIs on 37% of R01-equiv awards in the past two Fiscal Year. Women were also about 37% of the PIs on RPGs for FY2025. If we assume women are about 54% of the population, their underrepresentation as funded PIs is about 31.5%. Since Black individuals comprise about 14% of the population, representation as about 2% of funded PIs is 85.7% underrepresentation. As discussed above, a similar under-representation holds for study section membership.
Low numbers are pernicious. With aggregate study section participation of at least 25% women, and maybe up to 37%ish in recent years, on study sections this almost guarantees quite a few women on each and every panel. In contrast, 2% overall means that many panels do not have any Black reviewers. And this may interact in strange ways with grant topics, and study section domains. I cannot recall being on NIH review panels of normative size (20+ reviewers) where I was the only Black person. If some scientific domain review panels are regularly including 6-10% or more Black reviewers, well, some other panels must be horribly deficient.
*This difference was mostly driven by Type 2. It may be the case that Type 2 renewals were driven into obscurity by the crashing success rates of the post-doubling and this eliminated the source of the gender disparity. Winning?
**Per CSR Data Book, FY2004 dated 5/23/2005.
***this is unlikely to happen, just given the selection of these reviewers from the ranks of reasonably successful Black PIs, implicit biases, the tendency for those on the short end of bias sticks to favor scrupulous “fair” decision making, etc.
DrugMonkey 