Research paper vs survey vs position paper — and the CS peer-review rule

Not every paper is the same kind of paper. For most of arXiv, the difference between a research article, a survey and an opinion piece mostly affects how readers use it. In computer science, though, the difference now changes whether arXiv will accept it at all. This article explains the three types and the CS rule you need to know about.

ResearchGuild is independent and not affiliated with arXiv. Please read the linked arXiv blog post yourself before deciding what to do.

Three kinds of paper

Research paper

A research paper reports something new: a method, a result, a proof, an experiment, a dataset, a system. It usually has a clear contribution statement, a description of what you did, evidence, and a discussion of limitations. This is the kind of paper arXiv's moderationModerationarXiv's check of every submission for fit and basic quality before it is announced. It is separate from endorsement and from peer review. pages have in mind when they describe "refereeable scientific contributions that follow accepted standards of scholarly communication."

Review (survey) article

A review or survey organises existing work. It might compare methods, build a taxonomy, or summarise the state of a field. Good surveys are genuinely valuable, but their contribution is synthesis rather than new results.

Position paper

A position paperPosition paperA paper that argues for a viewpoint or direction for a field rather than reporting new results. Like surveys, arXiv's CS section only accepts these after peer review elsewhere. argues for a point of view: where a field should go, what it is getting wrong, what a community should prioritise. It may use evidence, but its main contribution is the argument.

A quick self-test: if you removed the related-work section, would your paper still have a core contribution of its own? If yes, it is probably a research paper. If most of the paper is related work, it is probably a survey.

The CS rule for reviews and position papers

On 31 October 2025, arXiv announced an updated practice for its CS categoryCategory (subject class)arXiv's topic label for a paper, such as cs.AI (Artificial Intelligence) or cs.CL (Computation and Language). Moderators and readers use it to route and find papers.. In summary (as of October 2025, per arXiv):

• Review articles and position papers must be accepted at a journal or a conference and complete successful peer reviewPeer reviewFormal evaluation of a paper by independent experts, usually organised by a journal or conference, before it is accepted for publication. arXiv moderation and endorsement are not peer review. before being considered for arXiv CS.
• When submitting, authors should include documentation of successful peer review. The announcement specifically asks for the peer-reviewed journal reference and DOIDOIDigital Object Identifier: a permanent ID for a published work, such as 10.1145/1234567. A DOI can serve as proof that a paper was published after peer review. metadata.
• Review/survey articles or position papers submitted without this documentation are likely to be rejected and will not appear on arXiv.

arXiv explained that the change was made because of an unmanageable influx of review articles and position papers to arXiv CS.

The announcement also notes that arXiv has always released scientific papers studying the impact of science and technology on society, for example in cs.CY or physics.soc-ph. If your paper is a research study in that space, it is not automatically a "position paper" just because the topic is societal. Read the post carefully and decide honestly which type your paper is.

What counts as "successful peer review"?

We have not found an official arXiv list of exactly which venues or review processes qualify. If you are unsure whether your venue's review counts (for example, a workshop or a non-archival track), check the arXiv announcement and, if still unclear, ask arXiv through its user support channels rather than guessing.

If your paper was already rejected

The announcement says that if your position paper or review articleSurvey or review paperA paper that summarises and organises existing work on a topic rather than reporting new results. arXiv's CS section only accepts these after they have passed peer review elsewhere. was rejected because it had not completed a successful peer review process, you can submit an appeal request to resubmit once it has. arXiv's appeals page says appeals must go through the user support portal and that you should not contact moderators directly.

What this means for first-time authors

• Writing a survey as your first paper? That is fine and common, but in CS plan to submit it to a peer-reviewed venue first, then post to arXiv with the reference and DOI once it is accepted.
• Mixing types? Many papers include a survey section plus a new method or experiment. If your novel contribution is real and central, present it clearly in the title and abstractAbstractA short summary of a paper: the problem, what you did and what you found. On ResearchGuild, only the abstract is shown until you choose who reads the full paper. so a moderator can see it quickly.
• Other subject areas may treat reviews differently. arXiv's content-type policy page is the best starting point outside CS.

How ResearchGuild can help

Reviewers on ResearchGuild can give you an outside view on whether your manuscript reads as a research paper, a survey, or a position paper. That honest feedback before you submit can save you a rejection. When you're ready, start a paper or run the readiness check.