Friday, April 27, 2012

Should you be suspicious of papers written by a rugby team?

I wonder if over the past decade the number of co-authors of condensed matter papers in high profile journals has increased significantly. It seems quite common now to have more than ten authors.
Is this really justified?
Have all of these people really made a substantial contribution to the paper?
Are they all willing to stake their scientific reputations on all of the results in the paper?

Many of these journals require a statement of the contributions of the different authors. But, most of the statements I read are quite generic.

I realise that some of these papers involve theory and experimental collaborations. Some report measurements using complementary probes (e.g., x-rays + neutrons + optical spectroscopy). However, for many of these papers I would have thought that the numbers would be:
1-2 people make the sample
2-3 make the measurements
1 is a friendly theorist who helps in the interpretation.
This adds up to 4-6 not 15!

Why should we care? Here are some possible concerns.
  • Some junior people who do the bulk of the work are not getting due credit.
  • There is a dilution of responsibility for the content of the paper.
  • The long author list may be hoodwinking us into thinking that the paper reports a more substantial body of work than it does.
  • There is a problem with joint theory-experiment papers.
Is this something we should care about?


  1. When applying for jobs, is it useful to have a large collection of papers to one's name, even if they have large authorship groups? That would be an obvious incentive to increase the number of authors on each paper.

  2. A large number of papers always impresses bean counters. It also increases the chance of increasing ones h-index since that metric is not sensitive to number of authors or author order.

    My view would be that unless you are first, second, or last author having your name in the middle may count for little. Furthermore, it can be a negative if the ratio of the number of middle author to first author papers is less than one.
    I think having a few middle author papers can be a plus because it shows you are a team player and/or have something broader to contribute.

    Ultimately, people want to hire leaders not minions.
    On the other hand, at the postdoc level some people want to hire a minion with a particular skill, and so if they have contributed to lots of papers that would be seen as a plus. However, I am not sure such a person has much of a future.

    I should say that I am not claiming that any of the above is right, just, or fair. It is just the way things are...

  3. the preponderance of user facilities (esp. synchrotrons) can certainly boost the number of authors on the average experimental paper. Staff scientists, if if not directly involved in the experiment, are often included if they help maintain a given instrument. I think this is the policy in many larger experimental groups as well...