A senior professor once said to me,

you expect one (top-ranked) journal paper for every $120,000 in funding.

I’ve always assumed it was an unverifiable figure but I’m interested to find out if work has been done.

Obviously this figure will differ for fields, funders, and your definition of top-ranked. My question is: Are there any meta studies that have looked at this funding versus quality issue?

Caveat: I’m aware that ‘quality’ of science is something that we all agree can’t be measured. But people do measure it. And I’d like to know if those people have produced any work looking at quality per unit funding and the circumstances that improve the ratio.

  • "The circumstances that improve the ratio". Are you sure that you want that? As your question states for 120k$ you are expected to produce one high quality paper to make the funding agency happy. Why would YOU want to produce more papers for the same money? 120k (minus the institution's overhead 20% or more) should be a good (but not extravagant) paycheck for about a year for one researcher. One very high quality paper per year, seems like a nice ratio to me. – Alexandros Oct 8 '15 at 9:34
  • 2
    @Alexandros I understand the question as "how can the funding agency distribute the money so that they can get more papers for the same cash". – Davidmh Oct 8 '15 at 12:56
  • If you wanted to try producing something yourself, you could try looking at the UK's Gateway to Research data - it would require a bit of manual aggregation and deduplication, as grants are presented idiosyncratically (each institution's share of a joint grant is presented individually, & outputs may be duplicated), but the core question of "how many papers came out of a grant in relation to funding" is probably answerable with a bit of work. – Andrew Oct 8 '15 at 15:39
  • This isn't directly an answer, but thought it might be relevant to your question. Measuring the quality of science is very difficult, but one ongoing project which attempts to do something close to this for the US is the STARMETRICS. I saw the founder of this program (Julia Lane) give a short talk about what they are trying to do. They have massive amounts of data on who gets what funds and are trying to use it to identify the benefits. Looks similar to the program @Andrew mentioned (which I hadn't seen before -- looks cool). – cc7768 Oct 9 '15 at 19:28

I think any figures will be moot. Let's take a Nature paper that is signed by people from US, Chile, Sweden, France, Israel, India, and Japan. Whose's funding would it correspond to? All of them? That is a lot of people, so a lot of money, we'll get a terrible ratio. Only the main author? In big collaborations is difficult to pinpoint who that would be. It is easy for a particular manuscript, but not a good representation of the whole.

Let's assume you can trace it down to a specific group. Typically, they have multiple sources of funding, which one should get the cake? Say I publish the paper "On self frying potatoes", whose experiments were funded by the grant "Development of self frying potatoes", it seems obvious. But I could only do it because I attended a workshop on "Tuber theory" paid by my other grant "Theoretical agriculture". We could make the counting at a research group level. How much money they get, how many publications get out, regardless of the source and purpose. That would still have the problem of how to count collaboration papers.

| improve this answer | |
  • 1
    Why not just count the grants acknowledged in the paper itself? Many (most? just NSF?) agencies require acknowledgement, so it might be a decent measure. – Bill Barth Oct 8 '15 at 15:34
  • @BillBarth That gives you a list of the grants, but not how to partition it. – Fomite Oct 9 '15 at 4:23
  • @Davidmh, this seems like an internal measure to me. I.e., a senior professor says to a new postdoc or junior prof, "Hey, kid, you know you need to win about $120k per paper you can publish in a top journal, right? You should keep that in mind on your way to tenure." I don't think it needs to be made precise. Certainly each set of authors could pull their own numbers together. – Bill Barth Oct 9 '15 at 13:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.