Preprint Strategy 2026: medRxiv vs SSRN vs MetaArXiv vs OSF

After A2's fourth journal rejection, I posted to MetaArXiv. DOI live in 24 hours. A LessWrong crosspost picked up 11 karma. Not the outcome I was hoping for from a submission cycle — but the citation that landed three months later wouldn't exist without it. And it happened because MetaArXiv was the right server for that work, not just the nearest one.

Most researchers treat preprint servers as interchangeable. They're not.

Four Servers, Four Audiences

medRxiv is the standard for clinical and public health work — and it has a moderation step most researchers underestimate. Submissions go through a screening panel before posting, which adds 24-48 hours but carries a credibility signal that journalists and funding agencies notice. If your work is clinical and you need a citable DOI while under peer review, medRxiv is the default. The screening is a feature, not friction.

SSRN skips moderation entirely. Posts in hours, and Google Scholar indexes it fast — sometimes faster than PubMed-indexed journals get picked up. The tradeoff is optics: SSRN started in social science, and some clinical reviewers notice. But for researchers who need Google Scholar visibility during a review cycle, SSRN is consistently underused by clinicians. Fast indexing is the whole point.

MetaArXiv (hosted on OSF) is where methodological work, meta-research, and epistemics-adjacent papers belong. The DOI comes through OSF, the community actually reads preprints, and a crosspost to platforms like LessWrong or EA Forum is a genuine distribution strategy. A2 landed here because it argued about research methodology, not about a clinical population — that distinction determines who sees the work.

OSF Preprints is the catch-all, and "catch-all" undersells it. Protocols, data-sharing agreements, mixed-methods work, software documentation. If you're posting a pre-registration alongside study materials, OSF lets you attach analysis scripts and datasets to the same DOI record. For a clinical study making its protocol publicly citable, OSF is usually the most defensible option.

The Decision Tree

Three questions in order:

1. Clinical outcome data or public health? → medRxiv. Unless same-day Google Scholar indexing is the priority.
2. Research methodology, epistemics, or systematic review process? → MetaArXiv.
3. Protocol, materials package, or doesn't fit a disciplinary server? → OSF.
4. Need fast Google Scholar visibility, any topic, and medRxiv moderation delay matters? → SSRN.

The overlap case: a systematic review methodology paper. I'd use OSF for materials and medRxiv for the preprint manuscript, linked in the OSF record.

Rejection-Recovery Routing

When a paper hits its third desk rejection, preprint posting shifts from optional to strategic. A posted preprint gives you a citable DOI that accrues independently of journal outcomes. Some papers build 5-10 citations during the preprint window — that changes the resubmission context.

Before posting, check the sherpa/romeo policy for your target journal. Most Elsevier and Springer journals permit preprints. BMJ group is permissive. A few specialty journals prohibit preprints or require disclosure — find out before posting, not after.

One version-control trap: if you substantially revise for submission after posting, the preprint diverges from the submitted draft. Citations may land on the older version. Use OSF versioning, or at minimum note the revision date in the preprint record.

What Preprints Don't Fix

The publishing game's core dynamics — fit, framing, contribution clarity — don't change because a preprint exists. The reasons for desk rejection don't disappear once you have a DOI. A poorly framed preprint will sit at zero downloads regardless of which server hosts it.

What preprints do solve: the visibility gap between submissions, the citation timing problem, and the dead air while peer review runs. That's worth solving. One server choice makes it work or doesn't.

AI for Academic's peer review simulation at aiforacademic.world identifies the gaps that lead to rejection before you submit — useful to run before posting a preprint too. Free to start.