Publishing in arXiv and other open access repository

[PI]

It seems that cutting edge results are encouraged to be submitted to arXiv to retain the credit for first publication and avoiding reviewing delays. Can anyone comment on what happens to doi numbers when the work is later published in a peer reviewed journal?
We have never done this, but considering for some of our new work.
Any good and bad experience in using such repositories before sending for peer review to journals? What are some of the popular repositories for STEM areas?

[ribekle]

Published article will have its own doi number (attached to the journal). arxiv allows you to add a journal reference/DOI number after publication. biorxiv does this automatically.

Preprint customs are specific to the field and area. For example, nowadays high-impact journals in biology expect biorxiv link in their submission system. In CS, for machine learning and theory, posting on arxiv is the norm.

[Kron3007]

I am in biology and started posting my work on biorxiv or preprint earlier this year.  Once the article is published, the preprint server includes a link to the published version.  I dont know too much more about the DOI etc.

Overall, my experience has been good and I very much plan to do this as standard practice moving forward.  One of the pre-prints had been downloaded about 3000 times by the time it made it through the review process, meaning that people were aware of the work months before they would have been otherwise.  It has also been interesting to review papers that cite the pre-print before the published version is even out.  To me, that really speaks to the advantage of doing this and I really feel that the traditional publication system with the long delays for publication is bad for science.  This seems like a good compromise, where people have access to the results instantly, but it still goes through the process of peer review for quality control.  Perhaps if you are in a slow moving field none of this would matter, but I am working in a rapidly growing field where it is important to get results out quickly.

I also came across a paper that found articles that are posted as pre-prints receive more citations than those that do not.  This makes sense given what I mentioned above and supports my decision to use this approach.

[mamselle]

So there's no way it counts against you in tenure submissions? I.e., it's going to be peer-reviewed and published by a "regular" publisher, it just gets out sooner?

The green-eyed monster may arise among your humanities colleagues...or maybe not.

M.

[PI]

Thanks for these info. Once the journal paper is out, does the open access repository get deleted or people continue to cite both versions? Do you retain the same title etc. or treat these as two different papers?

[Kron3007]

Thanks for these info. Once the journal paper is out, does the open access repository get deleted or people continue to cite both versions? Do you retain the same title etc. or treat these as two different papers?

My understanding is that ok nice it is published, the preprint server will include a Link to the published version, but the preprint remains as well.  I guess people could end up citing either, but most people would cite the published version.

[fizzycist]

Thanks for these info. Once the journal paper is out, does the open access repository get deleted or people continue to cite both versions? Do you retain the same title etc. or treat these as two different papers?

At least for arXiv, the preprint will always remain, you are not able to remove it. But you can put DOI and link to journal version on the arXiv abstract page so people can find the published version.

The best part of arXiv is that if the published version is behind a paywall, readers can just read your arXiv version instead. You can update your arXiv version as many times as you want so you could even fix things after publication of you wanted.

Some journals ask that you wait X months before updating your arXiv version with any revisions that came from their review process. But I don't know if they can legally force you to comply.

Vast majority will cite your published version, not arXiv, with possible exception of some math fields. However it doesn't matter much because your arXiv version will have doi and journal link so readers can choose what they read.

And Google Scholar will figure out that arXiv and journal version are the same paper and combine cites to either into one entry. Almost always happens automatically but if it doesn't you can merge manually in google scholar.

[namazu]

So there's no way it counts against you in tenure submissions? I.e., it's going to be peer-reviewed and published by a "regular" publisher, it just gets out sooner?

Right.  In your CV you treat it as a preprint, i.e. an unpublished/non-peer-reviewed (but demonstrably completed) manuscript until such time as it is published through ordinary channels.  Meanwhile, you stake your claim to the problem, you can get potentially-useful feedback (unofficial peer review), and people can make use of your findings if they deem them worthy.

[Kron3007]

So there's no way it counts against you in tenure submissions? I.e., it's going to be peer-reviewed and published by a "regular" publisher, it just gets out sooner?

Right.  In your CV you treat it as a preprint, i.e. an unpublished/non-peer-reviewed (but demonstrably completed) manuscript until such time as it is published through ordinary channels.  Meanwhile, you stake your claim to the problem, you can get potentially-useful feedback (unofficial peer review), and people can make use of your findings if they deem them worthy.

Yeah, I actually think it is helpful for a CV as you can show that you are actively writing papers and people can see what you have in the pipeline.  Seems better than just listing papers as in progress.

The caveat is that it should be clearly marked on the CV to avoid confusion or the appearance of padding.  Once it is published, I would remove the preprint and perhaps make note of any paper that has also appeared in this fashion in the list of pubs.

It may also depend on your field as this practice is well established in some but quite foreign in others.  In my specific area of biology it is not common but is starting to gain traction.  So, I could see some of the old guard not loving it.

[Puget]

I've been thinking about starting to do this as well-- it's a pretty new practice in psychology (psyarxiv was only launched at the end of 2016) but I've started seeing a lot of big names doing it so I'm about ready to get on board. You upload the preprints through the Open Science Framework (OSF), which we already use for preregistrations and  sharing our data and analysis scripts, so it looks like it will be pretty seamless for us to start adding preprints to our projects there.

[Kron3007]

I've been thinking about starting to do this as well-- it's a pretty new practice in psychology (psyarxiv was only launched at the end of 2016) but I've started seeing a lot of big names doing it so I'm about ready to get on board. You upload the preprints through the Open Science Framework (OSF), which we already use for preregistrations and  sharing our data and analysis scripts, so it looks like it will be pretty seamless for us to start adding preprints to our projects there.

From my experience there is very little extra work to do this.  In fact, biorxiv has agreements with a number of journals to allow the pre-print to be submitted automatically to the journal, meaning that the only extra work is pretty much to press the button.

For anyone considering this, I would recommend trying it out with one paper and see how it works for you.  As I said, I am sold and will probably do this routinely from now on. 

One other benefit I forgot to mention is that doing this enabled me to include my most recent work in a review paper that I wrote.  The review was very timely, and if it were not for posting pre-prints, I would not have been able to include the results.  These results had a big impact on the review, so it would have been flawed or needed to wait for several more months otherwise.   

[PI]

Thanks everyone. This is a great discussion. I am inspired to try. Do you typically keep the exact same title for the paper. Fizzycist mentioned that you can update content, in which case you can match the title as well I guess.

[Kron3007]

Thanks everyone. This is a great discussion. I am inspired to try. Do you typically keep the exact same title for the paper. Fizzycist mentioned that you can update content, in which case you can match the title as well I guess.

I've always kept the same title, but you could change it.

One thing to note is that you can update the preprint, but all versions will be kept and are accessable.  Essentially, since they can be cited, they need to be  a permanent record.

[fizzycist]

Thanks everyone. This is a great discussion. I am inspired to try. Do you typically keep the exact same title for the paper. Fizzycist mentioned that you can update content, in which case you can match the title as well I guess.

I've always kept the same title, but you could change it.

One thing to note is that you can update the preprint, but all versions will be kept and are accessable.  Essentially, since they can be cited, they need to be  a permanent record.

True. And if you are using latex and uploading to arxiv, beware that you must use arxiv's compiler and deposit your raw .Tex file. And others can download the source files. So be careful about what you put in the source file, even if it is commented out!

But really that is a very minor nuisance for a really great resource. And if you write your paper in Word or whatever then you can just upload the pdf, latex is not required.

I'm in a field where preprints have been standard for decades, so pretty much every paper I've published in 15 yrs first appeared on arxiv. Which means there are no paywalls that can prevent ppl to access some version of any of my work.

And, while I'm an arxiv loyalist for my own work, I love these new preprint servers. bioarxiv has a very attractive interface imo.

[polly_mer]

Is this really a question in 2021?  Huh.  Well, welcome aboard the bus that has been rolling for decades now (https://arxiv.org/archive/physics started Oct 1996).