Does the Size of Your NSF Grant Request Matter?

Posted on March 12, 2014

4


The National Science Foundation appears to favor a specific size of grant when it gives money out to archaeology but it is complicated, as I have found out.

This post came about because Carla commented on one of these NSF data examination posts I have been writing recently,

“One potentially interesting way to group the data may be by amount per grant. I’ve “heard” that grants less than around $200,000 tend to do better. Mind backing up that statement with some numbers?”

To answer Carla’s question I took a look at the NSF data I have (by the way I accept requests like this, so if you have questions of your own feel free to ask). The first thing I did is graph the grant amounts (adjusted for inflation to 2013 dollars) for the Archaeology/Archaeometry/early Anthropology NSF grant programs (see this post for look at how the data was collected):

Distribution of NSF Funding for Archaeology

Distribution of NSF Funding for Archaeology

It is a long tail model of funding but the graph is not particularity insightful. So next, I broke down the grants into counts by the range of grants given (again in 2013 dollars, all amounts are adjusted for inflation). This provides a more interesting distribution of awards:

Range Number of Grants
$0-10,000 110
$10,000-20,000 800
$20,000-30,000 234
$30,000-40,000 84
$40,000-50,000 55
$50,000-60,000 56
$60,000-70,000 65
$70,000-80,000 73
$80,000-90,000 69
$90,000-100,000 74
$100,000-125,000 145
$125,000-150,000 132
$150,000-200,000 197
$200,000-300,000 271
$300,000-400,000 89
$400,000-500,000 37
$500,000-1,000,000 34
1,000,000+ 15
Total 2540

As you can see from the table there is a strong concentration of grants given for $30,000 and below.  I looked at these awards and found that a large number of them are for PhD dissertation awards. However, there are also grants for small projects, equipment requests, conferences, etc. too. So there are other grants given for smaller amounts but I would be cautious in drawing a conclusion that very small grants are more successful given that so many of them are for PhD dissertation awards, its own special type of grant award. There is an interesting concentration of awards between $100,000 and $300,000. Though as Carla had “heard” most successful awards are for less than $200,000. I broke down the grants by percentiles to illustrate this:

Percentile
10th  $  13,614.95
20th  $  15,573.01
30th  $  17,503.22
40th  $  22,562.31
50th  $  47,675.67
60th  $  87,073.21
70th  $127,352.31
80th  $179,447.96
90th  $252,685.31

90% of awards are for less than a quarter of a million dollars.

However, as I have discussed before there are other grants outside of Archaeology/Archaeometry programs that go to archaeology projects. I added in those grants to the Archaeology/Archaeometry grants to see if it changed the distribution:

Range Number of Grants
$0-10,000 162
$10,000-20,000 932
$20,000-30,000 313
$30,000-40,000 133
$40,000-50,000 89
$50,000-60,000 94
$60,000-70,000 102
$70,000-80,000 105
$80,000-90,000 101
$90,000-100,000 99
$100,000-125,000 216
$125,000-150,000 190
$150,000-200,000 288
$200,000-300,000 396
$300,000-400,000 179
$400,000-500,000 117
$500,000-1,000,000 146
1,000,000+ 87
Total 3749

Other than adding more grants, the distribution stays about the same. There are still lots of small grants and a concentration of grants between $100,000-300,000. So what Carla has “heard” still stands up. There does not appear to be much NSF funding for projects bigger than a quarter of a million.

To make sure that these trends have been staying the same over the years I then took a look at funding over the decades for the archaeology program:

Year $0-30,000 $30,000-50,000 $50,000-100,000 $100,000-200,000 $200,000-500,000 $500,000-1,000,000 $1,000,000-5,000,000
1950 0 0 0 0 0 0 0
1951 0 0 0 0 0 0 0
1952 0 0 0 0 0 0 0
1953 0 0 0 0 0 0 0
1954 0 0 0 0 0 0 0
1955 0 0 1 0 0 0 0
1956 0 0 0 0 0 0 0
1957 0 0 0 0 0 0 0
1958 0 0 0 0 0 0 0
1959 0 0 0 0 0 0 0
1960 0 0 0 0 1 0 0
1961 0 0 0 1 2 1 0
1962 0 0 0 0 0 0 0
1963 0 0 0 0 1 1 0
1964 0 0 3 0 3 0 0
1965 0 0 0 1 0 0 0
1966 0 0 0 0 0 0 0
1967 0 0 0 1 0 0 0
1968 1 0 0 2 0 0 0
1969 0 0 1 1 0 0 1
1970 0 0 1 3 0 1 0
1971 0 0 0 0 1 1 0
1972 0 0 0 1 1 2 1
1973 0 0 0 2 1 1 0
1974 0 0 1 0 1 1 1
1975 0 0 3 2 2 0 0
1976 0 1 1 2 5 1 0
1977 0 0 1 2 3 1 0
1978 2 1 1 1 0 0 0
1979 2 0 1 2 4 1 0
1980 2 1 3 2 4 0 0
1981 1 0 3 2 1 0 0
1982 1 2 4 3 0 1 1
1983 0 1 3 1 2 0 0
1984 6 1 2 6 3 0 0
1985 3 1 3 5 8 6 0
1986 3 0 5 9 9 1 1
1987 3 4 5 16 10 2 0
1988 11 1 6 16 9 1 0
1989 23 6 6 17 13 0 1
1990 33 5 5 13 9 2 0
1991 30 7 11 12 7 1 0
1992 38 4 10 14 12 3 1
1993 48 6 9 10 10 0 1
1994 43 3 7 9 14 0 0
1995 36 6 12 10 15 1 0
1996 45 10 18 11 13 1 0
1997 42 6 20 17 9 0 0
1998 41 9 16 19 14 0 0
1999 43 2 18 18 12 1 1
2000 41 1 6 17 19 0 0
2001 39 4 14 17 10 2 1
2002 60 4 15 20 11 0 0
2003 57 8 9 17 15 0 0
2004 37 8 8 13 17 0 1
2005 50 4 9 18 11 0 0
2006 41 1 12 22 9 1 0
2007 51 2 11 22 14 0 0
2008 42 4 6 16 13 0 0
2009 53 4 16 22 30 0 1
2010 61 7 16 17 15 0 1
2011 36 5 7 19 8 0 1
2012 61 5 16 16 13 0 0
2013 44 4 12 17 11 0 0
1130 138 337 484 395 34 14

As far as I can tell these patterns tend to hold up and have stayed pretty constant. It looks like the NSF likes to fund small projects, <$30,000, and those in the low 100,000s range.

Their Success Does Not Equal Your Success

I would be highly cautious about drawing any sort of correlation between success in getting awards and size of grants. For one, I only have the final product I do not have a list of how many and size of requests for all applications. It could be than only 91 people have ever put in grants between $40,000-50,000 for archaeology projects. Comparing our hypothetical applications to the real success of 89 being funded you could say that actually apply for the range is ideal. However, I do not have that key piece of data. I don’t know if the success in the $100,000-300,000 range is because that is what most people apply to.

A second reason to have caution is that grants go through a process of review to be judged as intellectually valuable enough to receive funding. No matter the size of grant it could receive bad marks and never get funded.

That being said, it is up to the NSF Program Officer to make the final recommendation. From the NSF,

” Reviewers’ numeric ratings of proposals, while a useful indicator, are not, by themselves, a robust metric of the relative merits of proposals. Program Officers look not only at the ratings provided by reviewers but also weigh the comments that reviewers provide on the intrinsic merits of proposals. Program Officers also take into consideration other factors that might not have been considered by expert reviewers. For example, proposals for innovative new ideas often use methods or techniques that might be considered risky by reviewers and panelists. Such “risky” proposals may result in transformative research that accelerates the pace of discovery….

Program Officers will also consider broader impacts that might not be obvious to reviewers, such as an infrastructure need that will serve a large number of people. There are many dimensions of portfolio balance that may influence the final recommendation. ” -source https://www.nsf.gov/nsb/publications/2013/nsb1333.pdf 

In fact, that report states that 1.87 billion (with a b) in funds was not given to projects that scored above the average of expert scores of projects  that did get funded. 4.2 out of 5 was the average score of projects that got funding . A project can get excellent reviews and still not be funded. So the size of a grant can come into play with the grant officer as they balance their portfolio. Moreover, it will most likely vary between officers and programs, different officers will weight things differently. It also might depend on the year the grant was submitted. If lots of projects requested higher amounts of funding ($100,000+) that year it could hurt your chances of getting money in that range. So a bit of game theory comes into play which makes the amount you request a much more complex endeavor than simply staying below $200,000.

I would say that requesting smaller amounts (under $200,000) could help but it depends on many outside factors, most of which you have no control over.

Edit– Lynne brings up a great point in comments- what one asks for and receives are not always the same. I cut some text out of the quote above and it was- “Although Program Officers consider concerns about risk expressed by panels, they also see the value of funding potentially transformative research. Even if the Program Officer decides not to fully fund the proposal, proposals that do not review well at panel due to methods that are unproven or risky, can be given small awards to allow enough work for a “proof of concept”.”

Many of the very small grants, when you read the abstracts, are for these proof of concept sort of work. Lynne also mentions, “In archaeology at NSF, John Yellen will frequently request that a grant that he wants to fund lower their requested amount – he tends to discourage awards greater than $250K so that more people can be funded. It is always useful to ask the Program Officer about size of award before you prepare the proposal.”

Good advice to follow.

Posted in: Uncategorized