Why do funding agencies like the NSF not publish accepted grants?

Question

When the NSF decides to fund a project, they post an abstract of the project, but they do not post any of the material that the project's PIs actually wrote. Why not?

Some good reasons for publishing these proposals are:

1. It increases transparency, so that the public directly knows what was funded with tax payer money.

2. It would help future PIs find example successful proposals in order to create better proposals.

Edit to make the question a bit more concrete: Has the NSF ever published a justification for not publishing accepted proposals online?

Answer 1

Why do funding agencies like the NSF not publish accepted grants?

Has the NSF ever published a justification for not publishing accepted proposals online?

While it is correct in a technical sense that the NSF "does not publish accepted proposals online", I think this framing of your question is a bit misleading. The NSF has a process for allowing the public to access awarded proposals. This process, based on the Freedom of Information Act, is designed precisely to balance the public's need for transparency in government spending with other important things, such as the government's need to function properly (which is why even FOIA won't gain you access to top secret military information, for example), and the completely legitimate desire of researchers to pursue their research plans on their own timeline without undue stress or fear of having their ideas scooped and exploited by others.

Some good reasons for publishing these proposals are:

1. It increases transparency, so that the public directly knows what was funded with tax payer money.

As I was saying above, the key word here is "balance". You can't have maximal transparency, because all sorts of things would go wrong if you try to have it - researchers would protest loudly, and/or eventually start including less and less useful details in their proposals, making the evaluation process more difficult and the outcomes less meritocratic; and/or eventually seek alternative funding sources or try to get by without NSF support, with detrimental results to their research. On the other hand, no one is advocating for complete secrecy either - note that the current system is already a compromise between privacy and transparency: researchers do understand that their proposals will be reviewed by a panel of experts in their fields, and accept whatever small risk of scooping/misuse of ideas may exist given this situation. Writers of proposals also understand that in the end their proposals can in fact be looked up by curious members of the public, journalists, and people in Congress, who may end up accusing them of wasting government money on useless research or other unpleasant things. Everyone accepts that when you take public funding it comes with some strings attached, including a certain amount of public exposure and some level of disclosure of your research ideas. So again, I think your framing of the question is slightly loaded, as it risks painting an exaggerated picture of the grants process as being shrouded in secrecy, and of researchers as somewhat paranoid people who are always fearful of having their ideas stolen. The truth is much more balanced and reasonable than that.

2. It would help future PIs find example successful proposals in order to create better proposals. The current system of asking friends for copies of their proposals is super frustrating and seems like it is likely fostering an atmosphere of cronyism.

I am not aware of any "system" of asking friends for copies of proposals. That's something that some people do and others don't (I didn't). And calling it "cronyism" seems like a huge stretch. In fact, I think it's rather the opposite; from my experience with the grants process (including serving on several NSF panels), while the grants process is far from being without flaws, it is as meritocratic a system as anyone has been able to come up with for a complicated process of handing out several billion dollars a year to many thousands of people based on a very messy and noisy dataset. And yes, it's true that the process is very competitive and can be frustrating for many people. That would remain true even if all awarded proposals were made publicly available.

Has the NSF ever published a justification for not publishing accepted proposals online?

The NSF page I linked to above offers an explanation of how to request access to proposals, and what kind of access will be granted:

National Science Foundation policy is to make the fullest possible disclosure of information, subject to restrictions imposed by the Freedom of Information Act (FOIA) and Privacy Act, to any person who requests information, without unnecessary expense or delay. [...]

Copies of awarded proposals are available upon request, there may be applicable fees. Personal and proprietary information will be removed from the proposal documents before they are released.

The reference to personal and proprietary information provides some justification for why the process works the way it does.

Answer 2

I think the simplest way to address your question is to note that you have asked it upside down. Transparency is not the default option now, has not been in the past, and it is at least uncertain as to whether it might become the default in the future. So then the question is, why has an agency like the NSF not changed their practices to be more transparent and open, given that the default is opaque and closed?

Put this way, the answer is quite simple: it is because no person or group has exercised the will to make the change; there has been insufficient interest or ability to change the status quo of closed by default. It is not that people have not called for increased transparency in government, and there have always been arguments that the NSF should be run by civil servants and professional administrators/managers, with decisions explicitly made in favor of social welfare or public interest.

Most actions, both governmental and non-governmental, are generally conducted on a need-to-know basis. And in the US at least, most institutions are not set up to be direct democracies quite intentionally, as masses of people were feared to be ignorant, short-sighted, temperamental, and easily manipulated. This is even more true of organizations like the NSF, part of whose design and organization was intended to allow scientists to engage in autonomous decision-making to advance basic understanding across a wide-variety of fields without insisting upon any particular idea of return on investment, popular political sentiments, etc.

At least, that seemed to be the political reasoning that is reported to have won out originally. The changes since then seem to be towards a direction of more non-scientist involvement in the process of review and grant-making, which has included major cuts to social science, banning of funding for political science, political criticism of what is funded and what isn't, legislation that sets funding requirements for different directorates, etc. One may see these as good or bad, which presumably will strongly color your opinion on whether the changes that brought them about were good ones or not!

So it is not that openness has not and was not ever considered - it was considered and rejected, historically. This doesn't mean that things cannot or should not change in the future - but one must also remember that interventions in complex social processes always bring complex effects. More tax payer awareness can mean more attempts to manipulate and misinform, including adversarial attacks meant to direct research away from sensitive or valuable areas. Less obstacles to involvement can result in centralization of power to the most socially powerful, as power in one area is used to gain power in another, rather than decentralization and devolution of power to the many.

Currently, the main thrust of transparency movements seems to be in wider-ranging policies, such as the Freedom of Information Act and its many amendments. As noted in comments, the NSF has policies publicly stated that allow people to make requests to receive documents such as successful grant requests. The requester and provided documents are themselves logged and made publicly available. It may be seen as socially weird between scientists to request such grants without contacting them directly, but the depths of social weirdness in many common human interactions is beyond my ability to quantify :)

Answer 3

In addition to other good points made... if everyone writing proposals had in mind that whatever flights of fancy... or novel insights... they were inclined to put into the proposal would potentially be quickly public, I'd think this would too much inhibit many people. On another hand, it could inappropriately encourage people with crackpottish impulses, since they could get potentially get publicity without having to have actually done the work, if only they could make a convincing case.

Also, related, this could give a way for people to claim priority on "ideas", even if they hadn't made them work, because their proposal was funded (=peer review/approval) and made public.

And, again, to avoid scooping (which does happen...) it seems best to wait a year or two after the award of the grants.

(Also, really, the notion that someone wants to improve their chances of getting a grant by imitating the grant proposals of someone else seems problemmatical. Will imitators cite? Will imitators give a share?)

Answer 4

Grant proposals must go into some detail, which the reviewers are forbidden to use or divulge, but which would be highly interesting for other scientists, and would allow those competitors with enough free money and the right staff on hand to gain an unfair advantage.

You might even put unpublished stuff in there, for example results from diploma theses, which would otherwise not be easy to find. The grant proposal gives everything, the result, and the name of the freshly graduated student which could just be hired away to some more illustrious institution.

Two years later, when students have shown posters on conferences and the first papers are published, that info is mostly stale and outdated. I don't know why the proposals are not published then. Probably because nobody would be interested any more, or because it would often embarrass the proposers.

Answer 5

I can think of a number of reasons why it is not done, most already covered. Adding to them (apologies if I missed them above): i) a scientist may be hesitant to have preliminary findings published. Preliminary means that they may be wrong, or misinterpreted. ii) She or he may also not feel comfortable with retrospective examination of what was proposed and what was done (novel findings may steer the work in an unexpected direction). iii) Because a proposal is confidential, the ability to patent an invention described in the proposal is preserved.

If we are talking about publishing the meat of the proposal (project description), there are several reasons why it may be a good idea. In addition, it would be worthwhile to publish the reviewers comments and perhaps the authors' post-review commentary. The reasons why this would be useful are:

1. Proposals represent a considerable part of a scientist's scientific output. They encompass preliminary observations, ideas, hypotheses, and proposed experiments. Without publications they could and will most likely be lost.

2. A proposal might document important scientific ideas and help establish proper timeline of scientific innovations.

3. As mentioned, other scientists might benefit from access to successful and unsuccessful examples.

NSF could publish the proposals in a journal-like format. NIH has already a few real proposal examples. These are extremely useful. Last, publishing should be voluntary on the part of the PIs.