Thread by @azfaust

Since starting to work with NSF grants a few years ago, after coming in with years of NIH experience, I am constantly comparing and contrasting the two agencies.

One of the first things that stood out to me with NSF is that the POs are often rotating - meaning most are not professional POs like at NIH and instead are academics on "loan" from their home institution.

NSF does have professional POs but they usually manage the entire division/directorate and deal little with general review minutiae.

This makes establishing a relationship with POs trickier. You should still email/ call and talk to them but know that they may not be there in one or two years (this can be a blessing if you don't hit it off).

There is no SRO at NSF. The PO is both the PO and SRO - meaning they review the application, identify and assign reviewers, manage the study section panel (if organizing one), and make decisions which grants to recommend for funding.

There is a portion of POs at NSF that don't believe in study section meetings because they believe that those with strongest personalities can overtake the meeting. So they simply collect the reviews and rank the grants according to those reviews.

Some Directorates function more as teams and will meet and make funding decisions for divisions together. Some give complete autonomy to the POs on who to fund. Those decisions can be overruled but it does not sound like this happens often.

There is a permanent file on each person that submitted a grant as a PI or co-PI. POs have access to this file whether the previous submission was to their division/directorate or not. Some POs don't consult this file before making a funding decision. Some do.

(POs cannot access files of their collaborators or institutional peers.)

Why does this matter? While NSF does not have a formal resubmission track like NIH, this does allow a PO to look back at previous grants and evaluate if the PI has been responsive to reviews.

Learning the culture of your directorate or division is important. Some are not bothered by multiple simultaneous submissions to NSF but some are (even if they are to different divisions/directorates) and knowing where yours falls on this scale can help you plan.

There is no numerical scale for reviews, instead it's qualitative (excellent, very good, good, fair, poor). The POs generally want the reviewers to spread the evaluations since it's easier to select "good" and "very good" but harder to reject "excellent" proposals.

There are a lot of "consultants" who recommend setting up in-person meetings with NSF POs in order to build that relationship. I have yet to talk to a PO who liked this trend. Instead, go to meetings and meet with them there.

Broader impacts will not win you the grant but can be a difference maker for getting a grant. POs have told me when they have grants that seem to have similar intellectual merit reviews, they will go by BI reviews.

It appears that the most successful BI sections have the PI plugging into an already existing outreach and education program. Setting up new programs takes a lot of effort and there are always questions if full-time faculty have time for that.

CAREER grants have a higher success rates than regular grants in almost all divisions that I looked at except for SBE.