I have a conference paper which is published at IEEE with some errors unfortunately. But I have submitted a version on ArXiv without the errors.

As a researcher, I honestly prefer reading versions on ArXiv because I always think that if such an error has been detected by the authors in the official version (i.e., IEEE, Springer, etc.), they will of course have tried to correct it in their ArXiv version. That is why I usually prefer to read ArXiv versions or that ones present on the authors’ personal page.

Some researchers told me that papers on ArXiv (or HAL for example) are more read than the official versions. Is this true?

  • 2
    Are you in math (or in a different field where practically everybody puts the paper into arxiv)?
    – user111388
    Mar 14, 2020 at 17:05
  • 1
    Are you just curious about this, or is there an action item behind it?
    – Buffy
    Mar 14, 2020 at 20:28
  • 3
    Most researchers aren't interested in any given paper. The few who are will go where they like. Your actions to produce a correct version are good, but you can't control what others do. Better to be more proactive before first publication. At this point, it is what it is and it can't be other than it is.
    – Buffy
    Mar 14, 2020 at 21:54
  • 1
    The field should be specified in the question as this will vary greatly by paper topic. Mar 15, 2020 at 0:20
  • 1
    A suggestion: if you do make any changes to the arXiv version after publication, make note of the differences in the arXiv document itself and in the arXiv's comments field for the paper (and also on your web page where you link to the paper) so that people can find out which version is more correct. (Many people will otherwise assume that the published version is more correct/final, since many authors fail to update their arXiv preprints when making revisions based on referees' comments.) Mar 15, 2020 at 13:36

1 Answer 1


Well, statistically speaking, ArXiv requires no reader subscription, so more people have access to it (regardless of what publications they or their institute pays for). Ergo publications on ArXiv have the potential receive views from a wider audience. Whether this translates to citations or not is another argument.

It’s relatively common in my field to see an author "park" some research on ArXiv before submitting it to official publications or conferences. The publication process can take a while (and conferences have fixed dates), but the paper might as well gather interest and citations during that time.

I’ve never heard of anyone publishing on ArXiv after publishing in a journal, but I have seen post-conference extended papers (which go into more detail without the page limit). I don’t know what copyright you signed over (or what access level your publications have) when you published but I’d be wary about infringing the official publisher's copyright by posting on ArXiv.


Well, according to comments, some journals explicitly allow publishing to ArXiv (or related sites) after publication. That might depend on gold access level, but it makes sense for journals to make ArXiv an ally rather than an enemy. Relevant xkcd.

  • Thank you for your answer. I have submitted to arXiv before my official publication. But when I saw my paper on IEEE, I detected some errors, and so I went directly and updated my arXiv version. So I want to know if most of researchers will make attention to my corrections on arXiv rather than just reading the official IEEE version that contains errors.
    – Christina
    Mar 14, 2020 at 21:22
  • 4
    I have published on arxiv after publishing in a journal - but my field is chemistry (on arxiv stat.ML) and [some] chemistry journals may make a fuss about previous publication if the manuscript is already on arxiv. Whereas the copyright transfer explicitly allowed me to put the authors' accepted manuscript to arxiv. Mar 15, 2020 at 0:05
  • 2
    "I’ve never heard of anyone publishing on ArXiv after publishing in a journal" I always do it that way. Mar 15, 2020 at 0:22
  • The first paragraph was confusing: it started "Well, statistically speaking" but then described an argument based on intuition, not statistics. Statistically speaking would be if there were some statistics available about how often people read arXiv and journals. Mar 15, 2020 at 17:21

Not the answer you're looking for? Browse other questions tagged .