Is the rise of pre-prints lowering the quality and credibility of researcher and increasing the pressure to publish?

Question

From what I understand pre-prints have been common in some fields for many years because of some practical obvious reasons, for example very long lag time from submission to publication.

However, it is clear that publishing in pre-prints is becoming more and more popular across many fields with the main reason of calling yourself first to a particular research/method/idea/etc. The only problem is that these pre-prints are not peer-reviewed, and even though they require a certain standard, which is quite low, nobody is really checking the credibility and soundness of any of it. People might rush through the scientific process to publish a pre-print because the scrutiny level is almost non-existent. Some of this has been already observed with pre-prints about COVID https://science.slashdot.org/story/21/02/13/1558235/misleading-viral-claims-show-dangers-of-preprint-servers-researchers-warn, I think there's even no verification on affiliations.

The more pre-print papers are being published, the more these papers are being cited in peer-reviewed papers. They might be correct but nobody might check in detail as a reviewer or could even content since it's not "published". In addition, usually citations such as "in preparation" or "submitted" are not accepted by many journals. It just seems that this could trigger a vicious cycle in which more non-peer reviewed papers are being cited in permanent publications which then remains in print for years to come.

Does the rise of pre-prints make science and publishing less credible? Does it increase the pressure to publish because it adds an additional tier to the publishing process?

Answer 1

The people churning out garbage preprints are also churning out garbage papers in predatory journals which are "peer reviewed".

The preprints aren't why these people are doing it. Being shit scientists driven by perverse incentives is why they're doing it.

Also I cite preprints all the time from people I know (either directly or by reputation) and it's not a problem because I don't need reviewers to tell me if a paper is good. I can read.

Answer 2

Does the rise of pre-prints make science and publishing less credible?

No. Peer reviewed publications are still peer reviewed. People who do not know the difference between a preprint and a peer reviewed publication are not in a position to judge the credibility of science.

Does it increase the pressure to publish because it adds an additional tier to the publishing process?

No. "Pressure to publish" is about competition for jobs and funding. The number of publication options is not relevant.

Answer 3

I agree with the gist of the other answers, which by and large refute the major concerns postulated in the question (though I can see where those concerns are coming from).

I would add a major benefit coming from the increased prevalence of preprints, that of accelerating the open speedy dissemination of important research results.

I'm old enough to remember when arXiv (the former LANL archive) became more broadly used than in physics. In particular, over my very limited time as a research mathematician, it went from used rarely in (pure) mathematics, by next-generation scholars in physics-adjacent fields, to much broader but not universal applicability. As it hopped from one subfield to the next, it much expanded the speed at which new results were disseminated.

Prior, if you were well-connected in a field, you might hear gossip that so-and-so had achieved such-and-such, and very exceptionally you might receive a pre-publication fragment, which might be anything from a completed draft paper submitted, to a 3rd party's cryptic notes from a talk somewhere. There was generally a caution about over-sharing, to avoid being scooped as well as just for arcane cultural reasons. That also meant there was little incentive, in fact disincentive, to rush the word out about one's own achievements, other than to frankly claim turf. What was shared flowed really only to those well-connected (friend of a friend stuff), and there was no version control: "unpublished results" might float around in various contradictory incomplete versions and it was very hard to know what was accurate and complete.

While in that environment, preprints did technically exist, they were not as prominent part of academic discourse. arXiv and other centralized servers, clearinghouses, etc. played a huge role in encouraging early sharing of results. Rather than scooping, they provided a way to "postmark" your achievements earlier in a long publishing timeline. They allowed a much wider range of scholars to be aware, read, learn, and build on recent research. And via centralized repositories, they provided version control.

Of course, a "preprint infrastructure" like this did not and does not have all the elements of a full peer-reviewed publication pipeline, and is not a substitute. But it is an important, positive complement. Any effect in the direction of "decreasing quality" or "increasing pressure to publish" (as mentioned in the question) must also be complemented with considering the increase in quality (from more eyes, sooner, on important results), decreased pressure to rush out a poorly written paper (the preprint already announces it, so can take time to write "the paper" propertly), and faster and broader research collaboration.

Answer 4

Peer review has never been a 100% reliable gatekeeper and plenty of incorrect results get published anyway. I am sure every scientist is aware of X result that was later shown to be incorrect. Example from The Lancet on Covid.

Science gets bad PR when it turns out to be incorrect. Since peer reviewed articles can also be incorrect, I doubt that the rise of preprints is having much impact. If someone is going to disbelieve science as a whole based on incorrect results, they are probably not going to distinguish between whether the science has been peer reviewed or not.

Answer 5

No, some kind of concept of preprint has existed for a long time and has never devalued the concept of publishing research findings in research journals. Preprints are what they are. Perelman published his proof of the Poincaré conjecture in a set of three preprints. OK, exceptional example, but obviously this is not clear-cut.

It would be like asking ''Are journal articles causing X?' or something like that.

For example, Gödel originally wrote up his incompleteness theorems which we are all now familiar with and sent them to John von Neumann in the form of a preprint prior to publication of his findings in a journal. It's not a big deal, I think people are just overthinking it. I read preprints all the time and find a lot of them very useful and interesting. Some of them are bad, but some journal articles are bad as well.

As another example from pure mathematics, Bhatt and Scholze uploaded a preprint in 2019 which has not yet been accepted for publication due to the length and complexity of the document, but which has received 64 citations.

Answer 6

Absolutely! It definitely creates an incentive to rush out papers.

I worked in between applied math and engineering and I have noticed a trend in applied engineering journals in recent decade that these research papers will assume some abstract framework and do a bunch of math derivations and prove some type of convergence results in said abstract framework.

At the end, everyone is left just wondering if there exists a single scenario in the universe that can fall into their framework. Eventually one will be found, in a couple of years, usually not by these authors, but then they get to claim priority because they worked out the theory first, and then they say "oh your example just happens to fall into our abstract framework!"

It used to be, at least in my field, that demonstrating the applicability/usefulness was a priority. And it is usually very hard to show how math can be applied. Now it is about who can churn out proofs more in more and more esoteric setup. While this can be hard, I have noticed that they tend to be excruciatingly incremental.

Of course, you wouldn't expect these papers to go through the review process, so here is what they do: team up with a big name. I suspect these "big names" have never even read the paper they purport to have wrote (given I personally know that many are not even an expert in the subject they are writing in), but as long as their name is on there, no matter how marginal the work is, it will be heavily cited.

Answer 7

Preprints serve a very important issue of science, fast dissemination of results.

Peer reviews never guaranteed the correctness of the results. Preprints also never say anything about the correctness. But traditional peer review is slow and sometimes have preconceptions about the research.

First of all, slowness of peer review. Sometimes, your research is good if you are one the first in that area. For example, you applied an algorithm to a domain first time. preprint helps a lot in this case.

This preconceptions about the research is also very important. I put a lot of effort to a review article in my PhD thesis. It was rejected 3 times from different journals. After 3 rejections, I talked with my advisor and put my review article as preprint. After this preprint, it was again reject 3 more times and I decided to not submit it anymore. Well, it is cited 107 times in google scholar as of 2021 February. Most of the citing articles are from journals with impact factor.

A review of KDD99 dataset usage in intrusion detection and machine learning between 2010 and 2015 Authors Atilla Özgür, Hamit Erdem Publication date 2016/4/14 Journal PeerJ PrePrints

This means that:

1. You should never accept that article results are fully true even it is published in a very good journal.
2. Also, you should never accept that preprint results are fully true too.

My conclusion: preprints are not lowering the quality and credibility of the researchers. It is actually increasing it.