1

I know people who set a timer for 2 pm eastern time to make sure their paper is the first one to appear on arXiv the next day. I guess the assumption is that the ordering is first-in-first-listed, based on submission order, but I couldn't find any confirmation of this policy. Is this indeed how the ordering works?

Assuming it is the policy, is there a good justification for it?

Note: this isn't to say I endorse this kind of micro-career-optimization, I was just curious to know if this is a documented and justified somewhere.

Improve this question
1
  • 2
    If you are the last to submit before the deadline, say on a Tuesday, your paper will be first in the Wednesday mailing for its subcategory (e.g. astro-ph.CO). If you are the first to submit after the Tuesday deadline, your paper will be first in its category (e.g. astro-ph) in the Thursday mailing (but last in its subcategory).
    – astronat supports the strike
    Nov 28, 2022 at 21:02

1 Answer 1

Reset to default
2

There are two different orderings, so you have to work out which one you care about.

There is one ordering for people who receive a mailing for a specific category. This will normally be in order of submission time (first submitted at the top), except that crosslisted items (those that have a different primary category) are moved to the end.

There is a different order for people who go to the website and look at the latest articles for a particular category. This is in reverse order of submission (last submitted at the top), and crosslisted items are kept in submission order.

I assume the main justification for this is simply that there are good reasons not to do anything else. What would you sort by, if not by submission time? Anything I can think of would either be patently unfair or have unintended consequences.

I think it does make sense to sort the list on the website in reverse order - after all, if you continue long enough you will get to the previous day's articles, so sorting in reverse order makes this consistent.

Improve this answer
5
  • 1
    You could make it random, which would make it impossible for people to optimize their submission time. That might be useful if their optimization was causing issues (i.e. abnormal load on the server).
    – Shep
    Dec 2, 2022 at 19:21
  • 1
    @Shep we have considered this in the past, and may do so in the future. There's been studies that show that the papers at the top of any given category mailing receive more citations, and it's been people's theory that a randomization to that list could influence future citations.
    – Jake
    Dec 5, 2022 at 19:25
  • I would assume that the influence on citations would be a good thing: if people are spending time optimizing their submission times on arXiv that's time that could be better spent working out a more merit based ranking. (I'm not saying arXiv should do that, just that academics who are worried about this should probably be working to curate their own lists of arXiv links)
    – Shep
    Dec 6, 2022 at 10:54
  • For extra chaos you could also randomize the website listings more regularly.
    – Shep
    Dec 6, 2022 at 10:55
  • 1
    I suspect that effect is at least partly a self-selection issue. In my experience not many people care about optimising their position, and those who do care tend to write papers that would be likely to be cited anyway.
    – Especially Lime
    Dec 6, 2022 at 12:27

You must log in to answer this question.

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