11

I am a PhD student in applied probability looking for postdoc jobs after submitting my thesis. What is the value of arXiv papers (not yet published in a journal) compared to papers published in a journal while applying for postdoc? For example, if I have two journal papers and one arXiv paper, whereas someone else has three journal papers, will he gets any benefit over me in applying?

My arxiv paper got rejected from a very top-tier journal after almost two years of review. I am working on the reviewer's comments; then I will submit to a new journal.

  • 13
    Can you specify a field. In most chemistry fields for example a preprint is worth nothing at the moment. – DSVA Jul 16 '18 at 7:59
  • 2
    In general, a paper that's already been published in a journal has much more academic value than one that still needs at least some work to be published in that journal. Much of that value comes from removing the uncertainty that it's actually publishable in that journal; some of it comes from having actually done the work to make it publishable in the journal already, freeing up your time for other things. So all other things being equal, the other guy should win (in principle). In practice, of course, all other things are never equal, so it depends :) – Stuart Golodetz Jul 16 '18 at 9:43
  • 9
    This is a great question for your advisor, who knows the academic job market in your field better than Random Internet Strangers do. – JeffE Jul 16 '18 at 9:58
  • 1
    The quality of a paper depends on the journal it is published in. He can have 10 low-quality publications and you may have one great one, and a good committee may prefer you over him. It is true that often hiring committees tend not to be able to make a determination of a paper's quality, but that also depends on where you apply – user Jul 16 '18 at 11:43
  • It may also depend slightly on the country where you are seeking employment. E.g. in the UK, hiring committees in maths are quite journal-conscious these days, so it could be better to have two in e.g. Annals Henri Poincare and one on arXiv rather than three in e.g. Proc. AMS. But I think 2 in Proc AMS beats 3 on the arXiv, if that makes sense - at least in the current UK scene – Yemon Choi Jul 17 '18 at 3:26
33

ArXiv papers are preprints, not publications, so their value in the context of a job application is definitely lower than a paper that has been accepted for publication in a peer-reviewed journal. In your example, assuming all other things being equal, the answer is that yes, the job candidate with 3 journal publications would be at an advantage relative to you with your 2 journal publications and 1 arXiv preprint.

Of course, all other things are never equal, and papers differ enormously in their quality, scope, and significance, so that it’s quite possible that your one arXiv preprint might still end up overshadowing any of the three journal publications of the other candidate. So I would advise you not to get too caught up in these sorts of numerical calculations. The number of papers has some significance, but it’s very far from the only thing people care about.

16

First of all, in all of my experience (in pure mathematics), when evaluating candidates there is no distinction made between accepted papers and published papers. So I will speak about preprints versus accepted papers.

In all normal circumstances, having manuscript X as an accepted paper is strictly preferable to having X as an arxiv preprint. (What would be an abnormal circumstance? What I can come up with at the moment is: X is published in a shady journal, or X is published in a journal that is legitimate but of much lower quality than anyone would expect X to be published in. These are edge cases...)

So let us suppose that you have accepted papers X and Y and arxiv preprint Z. I think we can agree that the hypothetical candidate who has accepted papers X, Y and Z is strictly preferable to you....with the key understanding that X,Y and Z denote the same manuscripts in both cases! But now you see that this assertion is useless in practice. In actuality, you are going to be competing with people who have done different work from you. If a search committee views your manuscript Z_1 as being of roughly similar quality to someone else's manuscript Z_2, then it is a point in someone else's favor if their Z_2 has been accepted. But it is just a point, not all the points. If your recommenders talk about the significance of Z_1 and their confidence that it is correct and will be published in a very good journal, and if these recommenders have a strong reputation in the eyes of the search committee, then that is a point in your favor. Not that it is about points: please read on...

The phrase "all other things being equal" is often used in hypothetical discussions of job candidates, but those who have been on search committees are well aware of the irony: in practice, all other things are very rarely equal. People hire based not on rubrics or numerical schemes, but on partially verbalized subjective preferences. Which is not a criticism! We want to hire the person whom we sincerely and holistically believe is the best academic, not the one who has the most points according to some fixed system. If such "metrics" are used at all, they are probably used to sway the opinions of third parties on the search committee, i.e., faculty who are not going to be directly involved with either candidate. In my experience, such "metric differences" have to be rather stark in order for them to really matter: two accepted / one preprint vs. three accepted probably does not look like a significant difference, especially for young academics. But if you were competing with someone who had eight more accepted papers than you, then the person who is pulling for you probably has to convince the rest of the committee that your work is distinctly qualitatively better than that of the other candidate.

Finally, I certainly hope that as a young academic, when you post a paper to the arxiv you also submit it for publication shortly thereafter (if not concurrently) and that you are doing everything you can to make sure your papers are being processed in a timely manner. If so...okay, you are doing your best, and what remains is a lot of randomness that is entirely out of your control. I think search committees understand this and are not going to fault a candidate much for the vagaries of the refereeing process.

  • 1
    Great answer. It’s worth adding that one specific situation where the “all else being equal” caveat is actually satisfied is when comparing the same person to themselves at two points in time: that is, a person who has a submitted preprint on arXiv is slightly less attractive as a job candidate than themselves a while later when their paper has been accepted. – Dan Romik Jul 16 '18 at 9:54
  • 1
    What would be an abnormal circumstance? --- Posting in arXiv a correct solution to the Poincaré conjecture (and more) vs. pretty much anything else, although of course this is not a useful example for the OP. – Dave L Renfro Jul 16 '18 at 12:52
  • 1
    @Dave: Right. If you can believe it, when I was first typing my answer, I included this example specifically and even entertained the hypothetical situation in which Perelman had published his work in the Albanian Journal of Mathematics as an example of this not improving his case....and then I realized that this was not useful for the OP. – Pete L. Clark Jul 16 '18 at 16:04
6

This answer will be a bit orthogonal to the others. It may also apply to other job-seeking situations, not just post-docs. It is based on the observation that few hiring committees choose between candidates just by weighing piles of paper.

For many positions, especially post-docs and those seeking a first position, what you need to show is current and active research activity. The hiring institution will want to know that you have a lot of ideas and that you are exploring them and that you can contribute to a research program. Even that you have so many ideas that you don't mind sharing them, especially with students.

In that sense an old journal publication may not have as much value as a few recent papers in arXiv. Show, however you can, that you aren't resting on your old laurels, but are still totally fired up. Show them that you have work-in-progress that shows promise.

3

Both quality and quantity matter.

There are some sub-disciplines, such as high-energy physics and cosmology, that use arXiv as a means of establishing primacy and initial clearinghouse before eventual publication elsewhere. In those areas, it's harder to say what the impact is.

But in general, arXiv papers can be published without going through a peer-review process. Consequently, they are not viewed as highly as a peer-reviewed paper in a "known" journal in the field.

One should also not just look at raw number of publications. One paper in a major journal and two arXiv preprints may be better than three or four papers in low-impact journals.

  • 4
    in those disciplines, it’s hard to say. No it’s not. First of all, arXiv is not a “publication choice”, it’s a paper repository. Second of all, even “in those disciplines” authors will still submit their papers to a journal after uploading them to arXiv, and the prestige and credibility of the paper (and job prospects of its author) will be enhanced after the paper is published in a journal. I don’t believe that there is any discipline where arXiv has completely supplanted peer-reviewed publication, but please correct me if you know of one. – Dan Romik Jul 16 '18 at 5:13
  • 3
    @DanRomik Not entire disciplines, but my understanding from friends in cosmology and a few related subdisciplines primarily publish on arXiv. I’ll change “disciplines.” – aeismail Jul 16 '18 at 5:53
  • 1
    Again, your terminology is misleading. In academia “publish” generally means “publish in a peer-reviewed journal”, so it’s a category error to say someone “publishes in arXiv”, unless you are using “publish” in its other, non-academic meaning of “make available online”, in which case, since the discussion is about academic publications, it’s important that you make that clear (or better yet, not use the word “publish”, since it causes genuine and frequent confusion that we see here almost daily). – Dan Romik Jul 16 '18 at 9:43
  • As for your claim about cosmology etc, can you please clarify if you mean people in those disciplines only make their papers available through arXiv and do not submit them to peer-reviewed journals at all? I’d be skeptical that that’s the case, but if you can provide evidence (or even just check with our friends and confirm that that’s indeed the norm) I’m open to being convinced otherwise. – Dan Romik Jul 16 '18 at 9:46
  • 1
    Typo: “our friends” should be “your friends” of course. (Wish I had such cool friends...) – Dan Romik Jul 16 '18 at 9:55
2

The biggest advantage here with arXiv preprints is that it allows you to include papers in a citeable way when they are not yet published. You can link to the paper, which is available in a readable format, with a notation of, "Submitted for publication in ...", or better still, "Accepted for publication in ..." in advance of having the paper published, at which point you could then use the vol./issue reference as normal.

While it may be unlikely that a review panel will read your papers, they may look at the author list for collaborators and affiliations of co-authors.

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.