Nature Medicine | Spoonful of Medicine

More on discrimination

One of my colleagues was telling me the other day that we at the journal have a bias in favor of the USA. She was specifically making this comment with regard to our reviewer pool, which is indeed dominated by US-based scientists. But then again, there are many more scientists here than in any other country in the world.

In terms of authors, though, there doesn’t seem to be such bias in favor of the US or against any country in particular. Have a look at this graph, which shows the ratio of published to submitted papers as a function of country.


Each color represents a year — from 2007 to 2004, top to bottom. Note that this plot includes papers submitted not only to Nat Med, but it’s pooled data from NMed, NNeuro, NGenet, NSMB, NCellBio, NImmunol and NBT. I didn’t plot the actual number of submitted papers because, of course, we get our largest number of submissions from the US. But this graph clearly shows that we don’t favor the US over, say, Italy or Spain. The graph also shows that countries that have invested heavily in science over the past few years, like Australia, show a steady increase in their ratio of successes over failures. Sure, countries like China and India still have some catching up to do, but they’ll get there, trust me.

LatAm stands for Latin America. South Pacific includes Singapore, Hong Kong and Taiwan. Scandinavia includes Sweden, Norway, Denmark and Finland.


  1. Report this comment

    Alan Dove said:

    This is fascinating. It would be even more dramatic, I think, if you could get figures on the number of full-time researchers in each of these countries. At a glance, I know that the US has many, many more researchers than Switzerland, and yet the Swiss are scoring many more Nature papers.

    That would suggest that if anything, Nature actually has a bias against US researchers, despite its American-dominated reviewer pool. Of course, such an interpretation would be absurd, which I gather is your point.

  2. Report this comment

    Alan Dove said:

    Please disregard the conclusion of my previous comment. I just realized the graph is the ratio of published:submitted papers, not total publications. So the Swiss have a higher success rate than the Americans, but that could also mean that fewer Swiss researchers submit papers to Nature.

    It would still be interesting to get the numbers of researchers in these countries, to normalize the data to the relative sizes of the scientific talent pools.

  3. Report this comment

    JCL said:

    I was going to include the actual number of submissions, but decided in the end that this was somewhat confidential data that maybe NPG wouldn’t appreciate my sharing.

    I don’t know about the number of scientists of different countries, but my impression is that getting reliable info on this is very difficult, as the definition of “scientist” depends on who you ask.

    It is nevertheless safe to say that there are many more researchers in the USA than in Switzerland, surely.

  4. Report this comment

    Alan Dove said:

    If we craft the definition correctly, we could automate the data collection. For example, we could probably get a pretty good approximation of the biomedical research population of a country by defining “biomedical researcher” as “anyone who has been an author on a PubMed-indexed paper in the last three years.”

    The basic outline of the computer script would be:

    1. Define a list of countries to be searched.

    2. Execute a search on PubMed for papers from the first country on the list, limited to the past three years (or whatever other time span the programmer defines).

    3. Dump the author names from the search into a list.

    4. Sort the list, remove all duplications, then count the number of names. That’s the census of that country’s biomedical researchers.

    5. Repeat for the next country.

    PubMed can limit searches to specific article types, so we could also eliminate letters and reviews from the set, leaving folks who are authors on actual research papers.

    Now we just need to find someone with the programming skills and time to do this. I’m not sure I have either, but I may toy with the idea and see what I can accomplish.

  5. Report this comment

    JCL said:

    good idea, although i think you’ll find that doing this through scopus, instead of pubmed, is a lot easier.

    if you don’t have access to scopus, i may do it myself, as it’s not too many countries, after all.

    incidentally, i just presented this graph at a talk in madrid, and people were quite interested in the data.