9
$\begingroup$

What is the average number of publications (including conference proceedings), per year, for CS postdoctoral researchers in the US? What are the ways to find out or estimate this number?

$\endgroup$
  • 1
    $\begingroup$ As it is usually discussed, conference is the way to publish for CS. From your (short) question, I guess that you are writing a report on the matter. I would recommend specifying specific fields or ask people to do so. Plus, stating the reason you asked this question might motivate people and bring attention to your question. $\endgroup$ – chazisop Jul 30 '11 at 9:50
  • 2
    $\begingroup$ I second the comment on specifying the field: in addition to the variation on publication frequency between fields in computer science (say, HCI vs Database), there are also variations between fields about the average length of time one spends as a post-doctoral fellow before finding a position (academic or not). I imagine that to estimate this, you would need to survey the institutions awarding postdoctoral fellowships, and postdoctoral fellows themselves. $\endgroup$ – Jeremy Jul 30 '11 at 10:24
  • 11
    $\begingroup$ @Leo: For what it's worth, the number of publications alone isn't worth very much as a gauge of the status of the researcher within the community, if that is why you are asking (I may be reading too much into the question, which is why somebackground would help). It is relatively easily to publish a large number of minor results. Not all papers are created equal. $\endgroup$ – Joe Fitzsimons Jul 30 '11 at 13:10
  • 5
    $\begingroup$ @ Leo: Web of Knowledge has absolutely appalling coverage of CS. If you only use it you'll miss a large portion of papers. $\endgroup$ – Joe Fitzsimons Jul 30 '11 at 20:38
  • 4
    $\begingroup$ Rather than Web of Knowledge or even Google scholar, I would try using DBLP to look up a few people who you know are at a similar point in their careers. The way DBLP breaks things down by year and type of publication (journal, conference, preprint, etc) makes it a pretty good way to get a feel for someone's publication rate. However, as Joe already said it won't tell you as much about the quality of the publications, which is I think more important. $\endgroup$ – David Eppstein Aug 1 '11 at 6:11
23
$\begingroup$

The main motive behind this question is to understand where am I standing with my own performance and productivity.

Then you're asking the wrong question. The right metric to examine isn't the number of publication, but rather your visibility and reputation (and ultimately impact) within the research community. If the intellectual leaders of your target subcommunity are eager to write you good recommendation letters, it doesn't matter whether you publish one paper a year or ten. Similarly, if the intellectual leaders in your target subcommunity are not eager to write you good recommendation letters, it doesn't matter whether you publish one paper a year or ten.

$\endgroup$
  • $\begingroup$ @JeffE, thank you for your comment. I don't thinks that it qualifies as the answer, though. $\endgroup$ – Leo Aug 1 '11 at 19:52
  • 14
    $\begingroup$ @Leo: Whether you call it an answer or not is up to you, but I think that you are extremely lucky when someone points out that you are heading in a wrong direction. Most people do not care or bother telling you when you walk in a wrong direction. $\endgroup$ – Tsuyoshi Ito Aug 1 '11 at 20:36
  • 3
    $\begingroup$ @Tsuyoshi, I'd like to know the average number of annual publications for CS postdocs. Please, leave aside your opinion about where I'm heading. $\endgroup$ – Leo Aug 1 '11 at 20:51
  • 28
    $\begingroup$ Five. The answer is five. $\endgroup$ – Jeffε Aug 2 '11 at 4:42
9
$\begingroup$

I'm guessing nobody really knows this statistic, especially in computer science where it's not clear what counts as a publication. Are all conference papers actually "publications," even in non-selective venues? Do all workshops count or only the selective ones? Do journal papers count in addition to their conference counterparts? What about letters papers? And where does arXiv stand?

Even worse, there's no good tool for you to get an estimate. For example, in machine learning, two of the top conferences (NIPS and AISTATS) aren't always represented in DBLP. The best is probably to visit the websites of other postdocs in fields similar to your own and see what they've done.

I'm also a postdoc, so I might not have have the correct perspective, but I also think that the number of papers doesn't really matter; what matters is the quality, consistency, and impact of your work.

Finally, "where you stand" relative to your competition depends on what job you'd like. My feeling is that, in general, people who get research positions at good labs or faculty positions at research universities probably have at least a couple publications in the top venues of their field per postdoc (and/or end-of-Ph.D.) year.

$\endgroup$
  • 12
    $\begingroup$ To add to this from the perspective of faculty recruiting, the number of papers may sometimes be used as a negative filter (candidate seems unusually unproductive), but is rarely used as a positive filter (lots of papers? hire them!). In both cases, people do look beyond the numbers - maybe the seemingly unproductive candidate is working on a big problem, or the prolific one had a lot of collaborators, and so on $\endgroup$ – Suresh Venkat Aug 1 '11 at 18:40

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.