How do they work? What are the differences between the two? How competitive are they? How long does it last?
It looks like you are looking at Mathematical Sciences Postdoctoral Research Fellowships here: https://www.nsf.gov/pubs/2016/nsf16558/nsf16558.htm The website has this to say:
"The Research Fellowship option provides full-time support for any eighteen academic-year months in a three-year period, in intervals not shorter than three consecutive months;
The Research Instructorship option provides a combination of full-time and half-time support over a period of three academic years, usually one academic year full-time followed by two academic years half-time. This option allows the Fellow the opportunity to gain teaching experience during the two half-time academic years. The full-time fellowship support will be provided during the first year except in extremely unusual circumstances, with any exception subject to approval by the managing program director."
The answers to all of your questions are in the program solicitation. Both of them give the equivalent of eighteen months of full-time financial support for you to carry out the duties you propose.
In a research fellowship, you’ll be focusing 100% on research - probably conducting it under the supervision of a full-time faculty member or other kind of researcher at a research institution of some kind. In the research instructorship, you’ll be taking on at least some teaching duties; it appears from the language that the typical grantee spends the first year 100% on teaching, followed by two years teaching 50% time.
My guess is that the option people take is based a lot on the kind of position they want. If you want to be a professor at an R1 or R2 institution - the kind where professors may only teach one or two classes a semester (or in a whole year) and where they focus most of their time and energy on building their lab, performing research, and writing grants, then you’d want to do the research fellowship.
If you are interested in balancing teaching with research - perhaps being a professor at a liberal arts college or smaller regional university - then the research instructorship is probably the better choice.
If you want to know how competitive they are, or any of that other inside baseball information (like what makes a good proposal), you can always reach out to one of the program officers listed at the top of the program solicitation. It is literally their job to help people get these grants, and savvy researchers spend a lot of time talking to their program officers before they submit their grants. Their office telephones may not reach them, but shoot off an email if you’ve got some curious questions (but please do read through the entire program solicitation before you reach out; try not to ask them any questions that are answered in there.)
What I don’t get is for research fellowship they are providing 18 months over 3 years. Does that mean they expect you to come up with funds for the rest? Probably with a mix of teaching position at host institution and possibly fund from again host institution? If you have to fund half of the three years by teaching, you are teaching about half of the time only slightly less than research instructorship. And with the instructorship, you have to come up with one year support since they are funding only 50% for the two years. Whatever support one has to come up with is likely from the host institution. Does that mean it is really up to the host institution which one to apply? It looks like the host institution will get two years of free teach with research instructorship while supporting the grantee for a year that could be teaching or research at their choosing. With the fellowship, the grantee is counting on the host for 18 months support either by teaching or getting internal research funding.
“What I don’t get is for research fellowship they are providing 18 months over 3 years. Does that mean they expect you to come up with funds for the rest?”
Yes.
“Probably with a mix of teaching position at host institution and possibly fund from again host institution?”
Yes, or a mix of any other research and teaching duties. For example, you could have a 2-year postdoctoral fellowship from your institution; getting the NSF would allow you to extend your postdoc to a 3.5 year postdoc. This is especially useful if your institution’s fund has a limit on how long you can receive a postdoctoral fellowship.
“If you have to fund half of the three years by teaching, you are teaching about half of the time only slightly less than research instructorship.”
That’s true, but most people who take this research fellowship are probably not teaching to cover the rest of the time. They’re probably on some other research fellowship or source of funding that allows them to do research 100% of the time.
“And with the instructorship, you have to come up with one year support since they are funding only 50% for the two years.”
Yes, I believe the idea is that the other 50% of funding would come from a research grant. It is possible to have a research grant that only funds 50% of your time; in fact, when you write a grant, you have to specify how much of your time will be dedicated to the activities in the grant (and thus what percentage of your salary it should pay).
“Whatever support one has to come up with is likely from the host institution.”
Not necessarily. It can be another grant or another postdoctoral fellowship.
“Does that mean it is really up to the host institution which one to apply?”
Not necessarily. First of all, again, your other 50% of funding could be from another grant. But second of all, many institutions will allow you to specify in what order and for what duration of time you would like to take your funding, within reason.
“It looks like the host institution will get two years of free teach[ing] with research instructorship while supporting the grantee for a year that could be teaching or research at their choosing. With the fellowship, the grantee is counting on the host for 18 months support either by teaching or getting internal research funding.”
This is not really the right way to think about this.
Generally speaking, academics have to justify their existence and specify what activities they’ll undertake to get money. At research universities, it is not uncommon for professors and other faculty to have to fund 80 to 100 percent of their income through research grants. What they get in exchange, though, is a great amount of autonomy and flexibility in their careers. Consider some examples:
Let’s say Heather in her first year of a postdoctoral fellowship PhD at UCLA. Her current postdoctoral fellowship is supported by NSF - in the sense that her PI wrote a grant that included funding for a postdoc. Heather has realized that she wants to be a liberal arts college professor. She has no teaching experience. So she applies for this NSF grant; she writes the grant with the aim of having a teaching role at Pomona College (which she of course works out with Pomona beforehand, gets letters of support, etc.) She takes a leave of absence from her postdoctoral fellowship for the first year to teach at Pomona full-time; then she returns at 50% time to her postdoc in the next two years and teaches at Pomona half-time.
Or let’s say Stephen had a two-year postdoc at Stanford. He’s coming into his second year of his postdoc, and he knows that he needs another year or two to really prepare his portfolio and resubmit his NSF CAREER Award. So maybe he negotiates another six months of support on his PI’s grant, then applies for this grant to get an additional eighteen months of research support so now he has two additional years.
Technically, does the institution get two “free” years of teaching? Yes - actually, it’s more than that; they actually get PAID to keep the postdoc on board, because federal grants also come with ‘fringe’ (an amount of money paid to the researcher’s host institution to cover infrastructural things like electricity, benefits, facilities, etc.)
But having a grant (especially an NSF) grant is a really good thing for a postdoctoral fellow. Being able to win federal grants looks great on your CV and makes it more likely you’ll get future grants. And it provides you, the scholar, with the freedom to do whatever you want (and have written into the grant) rather than doing what your PI or institution wants. There’s also the case that often there is no job until you can find yourself. Consider Heather’s example: Pomona wouldn’t necessarily even have a half-time position for a postdoctoral fellow to take, but because she’s bringing the money and the fringe benefits, Pomona can offer maybe some cool and different classes to their students for three years and Heather gets the teaching experience she wants.