40
$\begingroup$

I'm a fourth year PhD student in theoretical computer science. I'd like to stay in academia, so I'm thinking about how best to advance my career. Obviously the best way to do that is write lots of good papers, but another question is whether I should be trying to have more of those papers be single author.

So far I only have one single author paper (out of six). It's neither my best work, nor very recent. Is that a red flag when applying for postdoc or faculty positions? Should I try to have more single author papers?

This comes down to my approach to research. I like talking to people. I like telling people what I'm working on and, if they are interested, I'm keen to work with them. In other words, I'm very open to collaboration even when it isn't strictly necessary. Should I change that? That is, should I try to keep projects to myself more so that I end up with some single author papers?

Apologies if this question is off topic. I want to ask this question to people in my area, rather than on a general forum like academia.se. In particular, on all my papers, authors are ordered alphabetically. This makes single author papers a more important signal in TCS than in other areas, where author ordering conveys this information.

$\endgroup$
  • 5
    $\begingroup$ Co-authoring is a subtle thing. A general rule: if you thought for 2-3 weeks without a result, and somebody gave you an answer - he/she should be invited as a co-author. If you just told somebody in a cafe pause about a problem which you haven't thought even 1-2 days about, and somebody gave you an answer - then not. You should not "hidden" problems you are working on, you should just take a bit more time to think on them before asking others. B.t.w. serious researchers are very "resistant" against "taking them as co-authors" ... they want be sure to have made a real contribution. $\endgroup$ – Stasys Nov 13 '15 at 20:24
  • 4
    $\begingroup$ @Stasys I completely disagree with your "general rule". Authorship should be determined by the contribution made to the work, not by the time that one of the participants spent failing to solve it. In particular, if you tell me a problem in a cafe and I solve it without any input from you other than the problem statement, it sounds an awful lot like I should be the sole author of that paper, with an aknowledgment to you for telling me the problem. Or, a more likely and preferable outcome, we'd work together extending the result and both be authors. $\endgroup$ – David Richerby Nov 14 '15 at 14:24
  • 4
    $\begingroup$ @DavidRicherby Telling me a problem is a major contribution - good problems are hard to find. If someone told me a problem and I solved it, I'd generally want to include them as a coauthor. $\endgroup$ – Thomas Nov 14 '15 at 17:00
  • 1
    $\begingroup$ @Thomas Your criterion is also independent of the amount of time the problem-suggester spent failing to solve the problem. $\endgroup$ – David Richerby Nov 14 '15 at 17:14
  • 2
    $\begingroup$ @David: Note that my "general rule" points to only two extreme situations - subtle is only the "gray zone" in between: a real contribution, when a joint conversation already starts. On a "cafe problem": I meant a situation when you give me an answer before our cups are empty: a clear sign that either the problem was no problem at all or was a folklore fact. I agree with Thomas, but only if the problem indeed "smells very fresh", was not seen before. Otherwise a short acknowledgement for "turning attention" is fully enough. $\endgroup$ – Stasys Nov 14 '15 at 19:20
24
$\begingroup$

In some fields (like e.g. Economics and Mathematics) single authored papers -are- a good thing to have when you go on the job market. In theoretical computer science, collaboration is much more common, and it is not unusual for even relatively senior researchers to have very few single authored papers. It is not at all suspicious if a student on the market has only papers with co-authors. This is partly because we publish more frequently, so it is ok that each individual paper has less signal about your specific contribution. This is not to say that single authored papers are not impressive, but credit is super-additive across authors, and all things being equal, you should prefer to have a better paper with more coauthors than a worse paper with fewer. This should push you towards collaboration, since collaborators make your research both more fun and generally stronger.

Of course in hiring decisions, committees will attempt to figure out whether you were a significant contributor to the papers you worked on or not -- but this will be done largely through the letters you obtain. If you have several good papers with the same senior co-author, that coauthor should write a letter for you, since they can explain in detail your contribution. It also helps if you have a clear research agenda of your own. Many papers on -your- topic with a rotating cast of coauthors conveys that you are leading the research direction, in contrast to a series of papers on disparate topics, each in the research agenda of your various coauthors.

$\endgroup$

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.