Some journals rules prohibit submitting to arxiv. Researchers still want to submit their research to these journals because they are prestigious. One of the main goals of researchers is prestige (for their work to be widely read, for getting grants, for attracting grad students and post doc laborers, etc.).

Are there any efforts to work against journals doing that?

TFA links to http://thecostofknowledge.com/

which is an attempt to organize researchers to boycott Elsevier due to its practices. Elsevier is one of the journal publishers that has (or had) issues with preprints like arvix. I do think efforts like this are gaining traction, such as the agreement between the UC system and Elsevier mentioned at https://www.sciencemag.org/news/2021/03/california-universit...

to not publish, review for them or be part of their editorial board anymore. really easy. and there are always other venues.

Speaking as a biologist working in Australia:

1) Older collaborators with more power are being 'scared' of getting scooped when they post preprints. They're slowly coming around as the argument makes no real sense, preprint = published to the younger generation.

2) Funding bodies, i.e. the people who pay our bills and have all the power. The Australian Research Council, the biggest non-medical funding body here, does not allow mentions of preprints in grant applications. They give no reason for that rule.

3) Again, money. You want your research to go into the journal with the highest impact factor since that mostly decides whether you get to keep your job, get that new position, or get that grant. Some journals still don't allow preprints. You also get no feedback when a top journal rejects your paper, could the rejection be because others have seen your preprint, published papers built on top of the preprint, and now your paper submitted to the journal is not so novel any more? You have no way of telling.

4) Fear. Sometimes, after your paper has undergone peer review it's a little bit better. A few select people see the embarrassing errors in your paper and help you fix them before the paper is widely accessible. If you post a preprint the world sees your glaring errors.

When you say journals don't allow preprint, do you mean they don't let you publish if you've published on Arxiv? Can you not post the preprint and submit to a journal for peer review? Isn't that the best of both worlds?

If there are changes to your paper due to peer review, can you post the changes to Arxiv too later?

Yes, some journals will desk-reject your submission if it's on arxiv/biorxiv.

If you not post the preprint and submit, you're defeating the purpose of preprints, isn't it? We're back to square one!

Normally yeah, people submit an updated preprint based on peer reviewers' comments. Biorxiv has a nice 'see the newest version of this paper' button for that.

In medicine/public health/etc. there are two reasons:

1) medRxiv was relatively new, and while the COVID-19 pandemic was its moment, there are a lot of concerns about pre-peer review stuff, and the public, press, etc. not treating it with appropriate caution.

2) Some journals heavily penalize or won't consider preprints, and some of these journals are quite good. This has gotten a little pushback recently as well.

FWIW, I once backed off publishing on Arxiv because their requirements for the exact format of my PDF were too restrictive. At the time I didn't feel stronly about it one way or the other, and if I was in the same situation maybe I'd persevere. But it would have taken an enormous amount of time to reformat my large and complicated PDF. I've heard the same complaint from others.

You need to upload the original tex and somehow it always took some tweaking to get it to compile relative to my local setup. But yes, it’s a low barrier.