Comparing published scientific journal articles to their pre-print versions | Hacker News Reader

Comparing published scientific journal articles to their pre-print versions

link.springer.com

36 pointsrbanffy14 comments

Here is the pre-print version of the article: https://arxiv.org/abs/1604.05363

It is a kind of funny situation that springer has this published.

Pre-prints are extremely useful in fields where they are widely used. Physics, my original field, uses arxiv as well as self hosted locations. I know biology is trying to do more of this as well. I really don't understand why other fields just start pushing towards this more open model.

I think there is a difference between a preprint that has gone through peer-review but lacks a journal's formatting than an archive where you post your research prior to peer review.

In the latter, crummy articles are frequently posted for the mere purpose of planting a flag, and you have no basic assurances that they did things to a reasonable standard of quality. In the former, you just lack "publish formatting" which often just means "lowering the DPI on your carfully crafted figures with their crappy automatic article formatter."

In that article they make distinction among articles that were previously submited to arxiv and articles that were first published and then just added by the author to arxiv. The great majority was articles being pre-printed first and then being peer-reviewed.

I should have checked first. Good point, at least in regards to the article's subject matter. I would expect even less difference between the after peer review pre-print and the article.

A question about arxiv. Can the preprint be updated by the authors? I can imagine wanting to update a preprint after getting input from reviewers.

Second question. Is this research then suggesting that peer-review does little?

> Can the preprint be updated by the authors

Yes, it can, altough it's not consistently done in every discipline, so unless you are familiar with a discipline, you can't rely on the research being most up-to-date.

> Is this research then suggesting that peer-review does little?

I think it's important to make a distinction between the formal peer review process, and the concept of peer review. There's a lot of remarks to be made about the former. In some disciplines, however, it's common to share research on arXiv, collect feedback from readers there, and submit a version to a journal that incorporates that feedback - it's not a stretch to call that peer review as well.

But to somewhat answer the question: practices between disciplines on arXiv vary too much to draw conclusions about the formal peer review process based on this research, in my opinion.

Very true on both counts. I do enjoy a good ridiculous arxiv every once and awhile. I think that is partially where community comes in handy.

There's quite a push in many disciplines, but there's many different circumstances that make it difficult, such as monographs being more common than articles, research moving slower, fear of scooping, and just habits in general.

(Disclaimer: I have started a project aimed at getting more arxiv-like initiatives in other disciplines off the ground, but am pivoting a bit now as I've noticed that the barrier to running such initiatives is also rather high: https://medium.com/flockademic/announcing-flockademic-academ... )

Conclusion from the pre-print paper:

> We have shown that, within the boundaries of our corpus, there are no significant differences in aggregate between pre-prints and their corresponding final published versions. In addition, the vast majority of pre-prints (90% - 95%) are published by the open access pre-print service first and later by a commercial publisher. Given the fact of flat or even shrinking library, college, and university budgets, our findings provide empirical indicators that should inform discussions about commercial publishers’ value proposition in scholarly communication and have the potential to influence higher education and academic and research libraries’ economic decisions regarding access to scholarly publications.

> We have shown that, within the boundaries of the arXiv corpus, there are no significant differences in aggregate between pre-prints and their corresponding final published versions.

Considering that many articles were funded by public money, at least partially, knowing that pre-print and expensive peer-reviewed articles have almost the same content, I just praise arxiv and other initiatives.

While arXiv is great, the sad thing is that even in disciplines where researchers are used to reading research from there, researchers still feel the need to publish in regular journals (understandably [0]), which keeps us dependent on the traditional publishers and hence wastes a lot of public money...

[0] https://medium.com/flockademic/the-ridiculous-number-that-ca...

TLDR from the preprint: The hypothesis was that if publishers did anything meaningful there would be a significant difference between the preprint and published versions of a paper.

There were not.

"anything meaningful" isn't quite what they analyzed, they just analyzed the claim that they "contribute and enhance text" during publication. The fact that they didn't can't be used to jump to the conclusion that they didn't do anything meaningful. E.g. by determining a paper meets certain standards of accuracy, transparency and relevance would be some meaningful things that could still be done.

Quote from the bottom of the paper:

"This is a U.S. government work and its text is not subject to copyright protection in the United States; however, its text may be subject to foreign copyright protection 2018"

However, Springer is still saying that you have to pay them $39.95 in order to get a pdf of the paper that probably costs them a few cents to deliver.

Springer isn't your secretary, though. Feel free to get a free legal copy from anyone else .

Reply on news.ycombinator.com