Talk:MDPI

Proliferation of Special Issue section might not be appropriately placed

This is about what is currently the "Proliferation of special issues" section. I think the section is inappropriate in the current article because it deals with the special issue model for managing a journal (also called the "Guest editor model" by Scholarly Kitchen [2]. The special issue model is a model, first and foremost. It is independent of MDPI. Per Scholarly Kitchen, the same model is used by Frontiers, Hindawi, and Springer Nature.

Because the special issue model is a business model, it is not directly relevant to MDPI, and I think the section is inappropriate. It's like writing a section saying that MDPI uses the open access model - it's not wrong, it's just inappropriate. If we have the section at all, it should be in the history section and say that MDPI started adopting the model on ___ date, and that it drove significant growth at potential reputational cost. For the last statement we can cite Crosseto/Scholarly Kitchen, but not Ovideo-Garcia, since the latter did not suggest anything wrong with the special issues. (Truth be told Crosseto is not a very good source either, because he only suggests there could be something wrong; he does not provide any evidence. If there's something similar for MDPI to Retraction Watch on paper mills @ Hindawi, then we could add it.) Banedon (talk) 03:40, 2 August 2023 (UTC)

"Because the special issue model is a business model, it is not directly relevant to MDPI"

It is directly relevant to MDPI because they (ab)use this model, discussed in multiple independent reliable sources. This is not a section about special issues in general, it's a section about special issues as used by MDPI. Headbomb {t · c · p · b} 07:01, 2 August 2023 (UTC)

I don't see anything in the section right now that applies to MDPI as opposed to the special issue model as a whole. Everything there could equally apply (with different numbers, of course) to the other three publishers listed above. Can you point something out? Banedon (talk) 07:05, 2 August 2023 (UTC)

3000+ special issues per journal, vs 3-4 special issue per journal is a major difference. It's three orders of magnitude greater. Headbomb {t · c · p · b} 07:11, 2 August 2023 (UTC)

Did you read what I wrote above in the Oviedo-Garcia section? The special issue model is applicable only to OA journals. It is not possible to run 3000+ special issues in a hybrid or subscription journal. Saying "there is a difference of three orders of magnitude" is like saying "there is a difference of an infinity orders of magnitude in the fees required to publish in these journals". It doesn't seem like you understand the topic. I won't be responding further until Crawdaunt (or someone else) responds. Banedon (talk) 07:33, 2 August 2023 (UTC)

This is nowhere near unique to open access publishers. Nearly every journal publisher out there publishes special/thematic issues. Headbomb {t · c · p · b} 08:03, 2 August 2023 (UTC)

On this, I agree with @Headbomb: it's simply not true that MDPI's use of special issues is similar to other publishers. There is a clear extreme use of Special Issues by MDPI. Those data reported in Times Higher Education and Science Magazine (news dept) originally come from the Crosetto analysis. To emphasize the notion that there genuinely is controversy about the sheer magnitude of special issues, see:

"Carlos Peixeira Marques of the University of Trás-os-Montes and Alto Douro, for example, says MDPI sent him multiple invitations to serve as a guest editor, in agriculture, animal science, and engineering—but never in his field of business and tourism. “The absolutely insane number of [MDPI] special issues has made it impossible to guarantee minimum peer-review standards,” he says.

The speed with which MDPI’s special issue manuscripts are reviewed and published is also a concern, Crosetto says. In 2022, MDPI’s median time from submission to acceptance was 37 days, well below the 200 days at the PLOS family of journals, another large, open-access publisher he examined for comparison. For about one in three papers, MDPI’s turnaround was 1 month or less. Considering the time it can take to recruit reviewers and revise manuscripts, “this looks just impossible,” he says."

https://www.science.org/content/article/fast-growing-open-access-journals-stripped-coveted-impact-factors

My concern is just a WP:DUE thing. As @Banedon says, this is not 100% unique to MDPI (although as above, MDPI is a flagbearer in growth through special issues, not without criticism from significant sources). So to have its own defined section, when it's just one among a long list of controversies, is unjustified to me. I don't grant the idea that, because it's longer-running, and not a single event, it somehow deserves its own section. MDPI's high rates of self-citation highlighted in the Oviedo article are long-running, but a specific source reported them, and so made it into a public controversy. I don't see anything different here.

So IMO it shouldn't be longer than one paragraph, and it shouldn't be in its own special section.

-- Crawdaunt (talk) 08:25, 2 August 2023 (UTC)

No, the point is that you must compare special issues from OA publishers using the guest editor model to other OA publishers using the guest editor model. The sheer number of MDPI special issues might sound unique, but it is not when you compare with Hindawi/Frontiers/Springer Nature. The Scholarly Kitchen article even gives links - here are 52 pages of "collections" (aka special issues) for Scientific Reports: [3].

The sources on there being a controversy because of the number of SIs are also very weak. E.g. Crosseto's quote basically boils down to "this looks impossible", but it looks to me that it's clearly possible and the evidence is staring him in the face, he just refuses to believe it. (See Argument from incredulity.) I for example have seen a lot of complaints about how MDPI asks for reviews in 10 days and also requests revisions from authors in 5-10 days, implying these things are happening, they're just happening at a very fast pace. If Crosseto has got any actual evidence that the papers were not properly reviewed, he hasn't produced it. He could go around emailing the authors of all the papers (not just by MDPI, but also Hindawi/Frontiers/Springer Nature) asking if they actually received reviews, and to share them. He could also approach the editors listed and ask if they actually made the decision. Absent this kind of evidence, I'd actually favor not using the source at all. Banedon (talk) 08:48, 2 August 2023 (UTC)

I don't agree that they're weak. Dozens of scientists have voiced concerns over the rigour of the MDPI peer review process, cited in the Science Magazine article (beyond just Crosetto, who contributed data), and also the Times Higher Education article published around the same time.

I don't think your notion of having to request info from every publisher under the sun stands... Wikipedia is a place to report what is part of the public conversation, not to question how valid one side of that conversation is. Followers, not leaders, and all that jazz.

There is a widespread view that MDPI's peer review is too lax, and its proliferation of special issues (possible thanks to relying on variable guest editors, not consistent staff editors), is part of that. Analyses by Petrou, an ardent defender of MDPI, nevertheless find MDPI's review process is bizarrely fast, requiring him to censor MDPI from the dataset to get a meaningful interpretation of the data (ref4 of the present page, publishing fast and slow (Petrou)).

...

Interesting position we all find ourselves in though... @Headbomb is pushing for a stronger focus on criticism of the SI proliferation, I am pushing for some middle ground, @Banedon is pushing for this to be dropped entirely(?).

@Randykitty or @Karlaz1 do you want to comment? Crawdaunt (talk) 19:17, 2 August 2023 (UTC)

Dozens of scientists have voiced concerns over the rigour of the MDPI peer review process Right, but these are concerns, not evidence. An analogous coincern like "I don't see how Wikipedia can maintain editorial standards over millions of articles when their decision-making council (Arbcom) is only 10ish editors" does not show that Wikipedia has lax editorial standards. Sources like the Oviedo-Garcia article talk about the number of special issues because that data is easy to acquire, compared to data on the rigour of the peer review process, which is much harder (aside from emailing the authors/editors I don't see how one can acquire the data, unless MDPI publishes it). Absent that data I am not even sure if we should imply a link. It would surely fail WP:BLP, but we're not a BLP.

I prefer adding information about the SIs to the history section. Something to the tune of "MDPI adopted the Guest editor model in ___, and the model drove rapid growth, but there are a lot of SIs that never published anything" would be appropriate. One could add "dozens of scientists have voiced concerns over the rigour of the peer review of SIs", but that's a general critique of the model. If we include that, then we could also reasonably include "dozens of scientists have voiced concerns over the gold OA model which incentivizes publishers to accept articles regardless of their quality" into this article, but then one could go around adding this claim to all OA publishers (which is pretty much all publishers at this point, there aren't many journals that don't have an OA option), and then it becomes silly. I would favor leaving it out; if kept it shouldn't be more than a couple of lines.

Your preference of a section under controversies would be my second choice. Banedon (talk) 01:08, 3 August 2023 (UTC)

I see your point. Again I think I differ in that MDPI is clearly a special case in use of SIs, and also anomalous growth (alongside this use of SIs), and has a peer review turaround half or less than literally all other publishers. There is no objective definition of "good" peer review, only a definition of peer review. So when it becomes a public conversation about the quality of MDPI's science, its peer review, by virtue of over-reliance on the special issue model, I do see that as a controversy of MDPI, not of special issues. The controversy is over the frequency of special issues, not the fact that they are used.

-- Crawdaunt (talk) 06:47, 3 August 2023 (UTC)

I'm not sure if MDPI pioneered SIs (i.e. the Guest editor model). It's possible they did. If they did, then yeah it's clearly a special case, but only because they're the pioneer. I'm not convinced MDPI shows anomalous growth due to the model. Hindawi, Frontiers and Scientific Reports all use the model and show comparable growth. In fact, the evidence seems to suggest that the SI model is very good at driving growth, which is why all these publishers/journals rely on it. Finally, given the model involves making as many SIs as possible, if there is a controversy over the frequency of the SIs, then that's a controversy over the model and not the publisher.

(Also not convinced that there's a peer review turnaround time of half or less than all other publishers, since IEEE Access is even faster, but this does not seem relevant to the SIs.) Banedon (talk) 05:41, 4 August 2023 (UTC)

Pioneered, no of course not. SIs have been around for over a century. What they pioneered was transforming to make the majority of their publications, across all their journals, published through SIs. This is what makes MDPI unique among all publishers.

It's not true 'all these publishers rely on SIs for growth. Definitely Frontiers and Hindawi have begun using SIs (not to the same extent at MDPI). Sci Reports is nothing like MDPI re: SIs, it is the Nature dumping ground journal like PLOS ONE but Nature portfolio. That overwhelming majority of its articles come from non-invited routes.

The 'model' here is business-oriented, not peer review-oriented. The controversy is over the practice of making SIs the majority way articles are published. One ought to appreciate that making "special issues" the default route for publication is an oxymoron.

The IEEE point is fair. Although I think IEEE notes their acceptance rate is ~30% whereas MDPI journals publisher their acceptance rates, and they're more aking to 50-60% iirc. Finding the blog that analysed this now... IEEE link and MDPI reject rate chart included below.

https://ieeeaccess.ieee.org/about-ieee-access/frequently-asked-questions/#:~:text=On%20average%2C%20the%20IEEE%20Access,they%20receive%20the%20acceptance%20notification.

https://danielbrockington.files.wordpress.com/2022/11/screenshot-2022-11-10-at-22.41.30.png

In general, there are a ton of factors going on here. What is true through the murk, to me, is that there is a controversy quite specifically about MDPI's use of SIs, and that is the point we should be debating, not its merit. The tone of the section is determined by the validity of the content, but the proposal of a section itself is certainly justified in my view.

-- Crawdaunt (talk) 08:07, 4 August 2023 (UTC)

It sounds like we're not thinking of the same thing. When I think of "SI model", I am thinking of a publishing model where you try to get as many guest editors to lead as many special issues as possible. This is not doable by subscription journals. The reason is because the prices of subscription journals is intimately related to the number of issues they publish - you cannot suddenly have an extra 100 special issues in your 8 issues/year journal without throwing your subscription prices out of whack. In the same way, a subscription journal cannot have an "issue" with no papers, but OA journals have no trouble (although it's still not ideal, of course). Because only OA journals can implement the SI model, it means the SI model is recent.

If Frontiers and Hindawi have only just begun to start using SIs (I'm pretty sure Hindawi is a latecomer to the model, haven't checked if Frontiers started before MDPI though), then MDPI is the pioneer. That should be worth including in the article. Still, if MDPI is not the only publisher using the model, then I still don't see the controversy over the model as a controversy about MDPI. As for Sci Reports - what makes you say it's nothing like MDPI? It looks very similar, except the special issues are called "collections". BioMed Central are doing the same thing, as well. Rejection rate is a fair thing to mention in the article in my opinion, but that would be a separate topic to SIs, and any comparison would have to involve more publishers than just IEEE Access vs. MDPI. Banedon (talk) 09:51, 4 August 2023 (UTC)

Topical special issues existed in other journals before MDPI was founded, I think more generally focused on bringing in high-quality content than on boosting author-fee revenues. An example: [4] —David Eppstein (talk) 12:18, 4 August 2023 (UTC)

I'm beginning to wonder how many times I have to repeat myself. When I think of "SI model", I am thinking of a publishing model where you try to get as many guest editors to lead as many special issues as possible. This is NOT doable by subscription journals, and NOT the same as the (few) special issues that subscription journals can do. Banedon (talk) 12:34, 4 August 2023 (UTC)

Yeah, I'm sorry to say the feeling's mutual. I'm not convinced by your arguments, you're not convinced by mine (or @Headbomb's). I really have understood your idea. Agree to disagree and just leave a vote open?

1. per Headbomb: a separate section devoted to the SI controversy.

2. per me: the SI controversy is folded in as a subsection of controversies, and reduced to similar length as other controversies.

3. per Banedon: the SI controversy is not an MDPI-specific controversy and doesn't deserve special mention on the page.

Does that seem like a fair summary?

-- Crawdaunt (talk) 16:25, 4 August 2023 (UTC)

I really don't understand this need to cull the length. It's 300 words, all backed by relevant sources. It's about 5% of the article lengthwise. That is not excessive. Headbomb {t · c · p · b} 16:44, 4 August 2023 (UTC)

Looks reasonable to me, too. XOR'easter (talk) 17:39, 4 August 2023 (UTC)

Same here. --Randykitty (talk) 08:07, 5 August 2023 (UTC)

That's 3:1:1, think we both just got overruled @Banedon.

Thanks for input @Headbomb, @XOR'easter, @Randykitty.

-- Crawdaunt (talk) 06:03, 6 August 2023 (UTC)

I have to say it's refreshing to have a civil disagreement, compared to the rest of the often inane discussions that surround MDPI. Headbomb {t · c · p · b} 07:59, 6 August 2023 (UTC)

Paper that analyzes individual MDPI journals for signs of predatory publishing

I notice that while this article covers many controversies, we have quite a few articles on individual MDPI journals which don't contain any negative coverage, apparently a similar problem to one discussed in the recent Frontiers RfC.

Editors here may be very interested in this 2021 paper, which discusses individual MDPI journals, not just MDPI as a whole, and which can be used in our subarticles on individual MDPI journals while sidestepping any SYNTH or NPOV problems. DFlhb (talk) 10:12, 1 September 2023 (UTC)

Sorry, didn't notice we already cite that paper here; still, it can be of use in child-articles. DFlhb (talk) 11:49, 1 September 2023 (UTC)

Predatory again

The site PredatoryReports.org has included the MDPI journals on their predatory publications list. Is the site reliable? --Julius Senegal (talk) 07:10, 18 September 2023 (UTC)

Short answer: no. This website is an anonymous blog with no accountability that often copy/pastes (i.e. plagiarizes) text from other blogs, and even this Wikipedia article, to their own site.

-- Crawdaunt (talk) 09:58, 18 September 2023 (UTC)

• I'm going to echo Crawdaunt and say no, with the caveat that many things about MDPI are controversial and depends on your definition of 'predatory', to which there's no universally-accepted definition. This is plagiarized from our article, appears to equate success with predatory, appears to equate fast publication time with predatory, alleges MDPI articles have overly-high self-citations, and literally quotes ChatGPT as evidence that MDPI is predatory (seriously). None of the arguments are convincing. "Fast publication time = predatory" is especially silly, since to slow down all one needs to do is sit on the article for a few weeks before proceeding as normal. If the author genuinely thinks that is a good idea, they should post that on the blog.

Then there's Oviedo-Garcia's article [5], which has multiple problems. First it suggests that if any publisher owned 100% of the market, it would have 100% self-citations, and therefore be predatory. Second it compares MDPI vs. leading journals, which is comparing a publisher to journals, which is clearly not an accurate comparison. Finally it doesn't present any evidence that MDPI authors are coerced into citing other MDPI journals. I suppose one can't fault Predatory Reports for this post - they are after all citing Oviedo-Garcia's article, which did get published in the literature - but Oviedo-Garcia's article is poor enough that I wonder how it passed peer review.

Finally there's [this https://predatoryreports.org/news/f/mdpi-peer-review-problem?blogcategory=MDPI], which actually would be concerning, but is missing several key details. First Sergey Gromov could simply be wrong about the scientific content of the paper. Second there's no evidence that the journal accepted the article over the objections of its editors. If the editors decided to accept (wouldn't be the first time editors decided to accept a paper which some reviewer recommended reject for), then the journal can scarcely overrule the editors. In this scenario Sergey Gromov ought to criticize the editors, not the publisher. Finally even if Sergey Gromov is correct, retractions happen all the time, even in the top journals, so a single error is at best suggestive and not nearly conclusive.

One could press an argument that MDPI is predatory (under certain definitions of 'predatory') but this website is not it. Banedon (talk) 13:13, 18 September 2023 (UTC)

The PredatoryReports article is problematic for many reasons, but it doesn't suggest anything like what you say it suggests. It's not that fast = predatory. It's that anomalously fast = likely problematic. Likewise a publisher having 100% of the market leading to 100% self citations is again ridiculous. The issue is anomalously high self citations. Etc. etc. etc. Headbomb {t · c · p · b} 21:56, 18 September 2023 (UTC)

I don't see how that addresses my criticisms. Anomalously fast = likely problematic? So sit on the manuscript for a couple of weeks before starting to work on it, and you are magically no longer anomalously fast. Claiming that is desirable is ridiculously silly. As for "anomalously high self-citations", that's not actually in Oviedo-Garcia's article, because it doesn't have data from any other publisher. If it wants to claim anomalously high self-citations, it needs to provide data for other publishers against leading journals. Banedon (talk) 23:06, 18 September 2023 (UTC)

Declaring COI

Hi all,

I believe I have developed a formal conflict of interest to declare, and will not make any further direct edits to this or related pages for the time. Reading the WP:COI page, I was surprised to see that it mostly addressed financial COI, which I do not have. However, I am lead author on a recent preprint that is very closely tied to the page. Again I have no financial interest involved (and no funding to declare for that work, which was voluntary).

My coauthors on this study, which began informally around April 2023, include Paolo Crosetto and Dan Brockington, both bloggers on publishing/MDPI. At one time, I recommended edits to the page that cited Paolo's work, which I did so prior to developing a close professional relationship with him. I think the various 3rd-party validations of his work's relevance (articles by Times Higher Ed, Science Magazine) also justified citing his blog anyways, and the blog was cited strictly as the 1st party source of that 3rd party popular media information.

You can find the preprint at: https://arxiv.org/abs/2309.15884

Hope it's interesting for the folks here also because... well... it was a lot of work to take a very thorough review of the state of publishing! If an author summary is more your style, I'd encourage reading the thread on Mastodon for highlights.

Cheers,

Mark -- Crawdaunt (talk) 06:44, 30 September 2023 (UTC)

I wouldn't call this COI, unless you added it to this article, but it deserves to be added in any case (once peer-reviewed). Here's my not-exactly-peer review as well:

• " which has outpaced the limited growth – if any – in the number of practising scientists" -- this needs to be backed up. The number of PhDs awarded is not an accurate indicator here, since a lot of PhDs don't go into research (at least the kind that produces journal articles). I actually think it is probable that the number of practising scientists has increased, especially in developing countries, where new universities are being established.
• Table 1 - some of the "publishers" are owned by the same publisher. That may jeopardize the conclusions.
• "“Special issues” are distinct from standard articles because they are invited by journals or editors, rather than submitted independently by authors. They also delegate responsibilities to guest editors, whereas editors for normal issues are formal staff of the publisher." This is not true. You can submit to a special issue if you so desire (e.g. this MDPI special issue has a clear "Manuscript Submission Information" section. Also the editors for normal issues are scarcely the formal staff of the publisher (they are the formal editorial board of the journal, but not staff of the publisher).
• "This behaviour encourages researchers to generate articles specifically for special issues, raising concerns that publishers could abuse this model for profit" - I don't have access to the cited source, but I'd check carefully that it actually does say this, because a priori I don't see how writing articles specifically for special issues can lead to the SI model being abused for profit. The statement actually makes it sound like writing articles can potentially be abused (in which case why did you write this article?). Furthermore, even for APC-charging journals, invited articles often have waived APCs.
• "Turnaround times also reflect a trade-off between rigour and efficiency: longer timeframes can allow greater rigour, but they delay publication." -- this is also in Oviedo-Garcia's article and I criticized it for the same reason. Unless you are seriously suggesting that it's a good idea to wait a year before starting to process any paper (and that somehow gives you the best peer review"), this just misses the point. The argument that there is a minimum time required to do rigorous peer review is surely correct, but the way to establish that would be to first discern how long it takes to do rigorous peer review. I'd be surprised if it took me more than two days to review your article in detail, for example. It might take me a week to find the time to review your article, but it does not take me a week to review it (and even if it did take me a week, that implies there's plenty of room to push down turnaround time even further). Pushing down the peer review time increases the average number of reviewers invited, but does not intrinsically jeopardize the rigour. (Although it can. An example is if the most-suitable reviewer is not available to review quickly.)
• "This implies these articles, regardless of initial quality or field of research, and despite the expectation of heterogeneity, are all accepted in an increasingly similar timeframe" - this conclusion is likely too strong. An alternative explanation is that these publishers have been focusing on decreasing their publication times. One way they can do this without affecting peer review is speeding up the time taken from acceptance to publication (this does not look analyzed at all in your article). Another way is by asking for reviews in a shorter time. As long as this time is longer than the necessary time it takes to do rigorous peer review, there should be no issue, beyond them needing to invite more reviewers.
• Re rejection rate - one thing worth pointing out is that if articles are invited, then they are expected to be high quality (or why would you invite the author to submit?). This should imply lower rejection rates. The different trend you found between Frontiers, Hindawi and MDPI is still unusual, though.
• Re impact factor inflation - if this is supposed to show that there could be some kind of citation cartel in MDPI's special issues, have you tried using Web of Science excluding self-citation data?

Banedon (talk) 14:13, 30 September 2023 (UTC)

Thanks for comments! Will note these for colleagues. The wording of "staff" should probably change, just key note being that editorial board members and guest editors are a different tier is all it wants. On the turnaround times and such: you're free to have your own take on Figure 3!

Yes ref says this (it directly says the line on potential motivation by profit, and is commenting on the rise of mega journals, which we show is driven by special issue model). On all other comments, I think you'll find additional considerations that directly answer, or provide more full context in supp. :)

Cheers -- Crawdaunt (talk) 21:14, 30 September 2023 (UTC)

I did admittedly only skim through the supplemental material. I might take another look later when I have more time. On another note, I think it's advisable for you to consult someone with experience in publishing (i.e., has worked in academic publishing before) - otherwise the lack of experience with how academic publishing works is very apparent. For example I pointed out several issues above that should have been apparent to someone who's had an insider's view of journal operations before. Ideally that person will have worked for both subscription & open access publishers. I'd nominate myself, but I'm currently in academic publishing. Things might change in a few months, but as of right now, I'm COIed. Banedon (talk) 01:38, 2 October 2023 (UTC)

I assure you we are well-versed, with decades of experience on our team (which you are free to look up), and your comments are well considered in the supp and explained (e.g. why Nature or BMC are not collected under "Springer"). The typo-esque "staff" phrasing hopefully doesn't invalidate the point for you, which remains relevant. We also had comments from numerous colleagues, and were in touch with the publishers themselves (see Acknowledgements), and again you're free to look up who those people are, what sorts of (extensive) editorial experience they have, who they work for (COPE), etc... We literally talked to the MDPI CEO, their Chief Scientific Officer, and got feedback (and thumbs up) from a former MDPI EiC. Hope that helps assuages concerns. -- Crawdaunt (talk) 10:16, 2 October 2023 (UTC)

I want to be critical, but ultimately it's your article, so good luck with it =) Banedon (talk) 11:37, 2 October 2023 (UTC)

By the way I think we should add results from this paper to the article now, without having to wait for peer review, especially since MDPI senior staff OKed the article. If there are no objections I'll go ahead and do it. Banedon (talk) 12:18, 2 October 2023 (UTC)

To clarify: MDPI senior staff were sent the document, and they provided comments, and we revised the text to acknowledge some of the points raised by their comments. We also reached out to Gemma Derick (this is noted in acknowledgements), who is the former EiC of "MDPI Publications." But MDPI senior staff did not "OK the article" in any official way. They did thank us for the opportunity to comment. Gemma Derrick, who is no longer with MDPI, "OK'd the article" so to speak, providing comments that were largely in agreement and having only minor points to raise - Although I still wouldn't want to put words in her mouth.

But yes, MDPI were alerted weeks in advance of our article, offered the opportunity to provide commentary, and we took their commentary, as well as many others, and made changes as appropriate. In fact, our contact with MDPI alerted them to the fact that they were still shipping their rejection rate data to the public, which the CEO and Chief Scientific Officer were apparently not aware of.

If you check MDPI journal stats pages now, the rejection rate data are no longer shipped in the html. However you can still see its former presence in the html using the Wayback Machine up until early September 2023, and until mid-2021, MDPI even included plots of their rejection rates on their journal summary stats pages. So I guess we managed to inspire some change at MDPI - although that change was to reduce their transparency...

-- Crawdaunt (talk) 18:59, 2 October 2023 (UTC)

Also, thanks! We might differ in opinion on some minor things, but the bulk of the work hopefully speaks for itself (or at least provides some data/evidence to inform some of the conversations that have been had on social media, or even this Talk page :P ).

-- Crawdaunt (talk) 19:17, 2 October 2023 (UTC)

Yes, your paper looks significantly more honest than Oviedo-Garcia's paper, which is why I'm in favor of adding the results to the paper before peer review. I made several edits, let me know if there's anything missing. Banedon (talk) 03:44, 3 October 2023 (UTC)

Thanks! Minor note for the page: Our article's "total articles" reflects only articles indexed in both WoS and Scopus, and not total articles period. So I removed this line on overtaking Springer etc... AFAIK, presently MDPI is 3rd largest behind Elsevier and Springer Nature.

- Also worth pointing out our choice to not aggregate all journals per publisher is per publisher licensing brand in Scimago, which we deemed was a reflection of corporate/editorial size independence. But of course Nature portfolio, BMC, etc... are part of Springer, Cell Press is part of Elsevier, etc... so the total journals/articles per publisher in Hanson et al. could be tallied differently given a different question/focus.

Cheers -- Crawdaunt (talk) 07:25, 3 October 2023 (UTC)

Hi there,

While we normally refrain from getting involved in the Talk page to provide space and respect for the editors working on our Wikipedia page, it is important for us to state a few things in this matter.

Thank you, Banedon, for your critical approach to this preprint. Your quick critique, along with that from others we have seen shared on different platforms online, presents valid points.

Also, thank you, Mark, for updating the preprint based on the feedback MDPI provided. We appreciate you taking some of the feedback into account. However, this does not mean that MDPI endorses or supports the article in any way. In your communications, you mentioned, "please note we are not seeking commentator's agreement on the final text," indicating you are not seeking approval from MDPI on your preprint.

With that said, it is important that we clarify MDPI does not approve or support the preprint. In fact, we still have issues with the narrative this preprint presents around academic publishing.

Based on our review of Wikipedia policies, we believe that the content of the MDPI Wikipedia article should be sourced from credible secondary sources such as peer-reviewed articles, not from preprints or blogs.

“Preprints – Preprints, such as those available on repositories like arXiv, medRxiv, bioRxiv, or Zenodo are not reliable sources. Research that has not been peer-reviewed is akin to a blog, as anybody can post it online (...).” https://en.wikipedia.org/wiki/Wikipedia:Reliable_sources

It's difficult to believe that one author of a preprint with a clear COI and another anonymous user can add a preprint, which is generally considered against Wikipedia's policy. While it has been far from the case we hope for a fair and neutral evaluation of the MDPI Wikipedia page and thank the editors who truly make an effort to keep their bias in check and take this approach.

MDPI Mdpi comms (talk) 12:55, 3 October 2023 (UTC)

The part of that quote which is obscured by the ellipses is unless they meet the criteria for acceptable use of self-published sources, which seems to be the case here. MrOllie (talk) 13:15, 3 October 2023 (UTC)

To be clear as well: at no point did I claim MDPI approved the article, and in fact I corrected this mistaken notion immediately. I also did not add the preprint to the main page, but rather drew attention to it in the Talk page alongside declaring my COI.

I will be impartial on when, or whether, the article merits any additions to any Wikipedia page in its preprint state. If another editor wishes to add the article to the page, it is not my place to argue for or against this. My only contribution to this was to first raise its awareness (in order to declare my COI), and then I made edits recently to correct that our preprint makes no comment on total articles published by any publisher, but rather, total articles indexed in both WoS and Scopus, per the Scopus database.

Best,

Mark -- Crawdaunt (talk) 19:25, 3 October 2023 (UTC)

I can see the argument that 1) I'm adding my OR (or at least my interpretation) on top of Crawdaunt's OR and 2) I'm also COI'ed, so if someone reverts the edits I made I won't revert back. Banedon (talk) 00:47, 4 October 2023 (UTC)

@Banedon can I ask what your COI is? Just curious. -- Crawdaunt (talk) 09:33, 4 October 2023 (UTC)

I currently work in academic publishing, which means I either work for MDPI or work for a direct competitor. In both cases I'm COI'ed. Not comfortable saying more than that, unfortunately. Banedon (talk) 10:36, 4 October 2023 (UTC)

It feels relevant to the MDPI page to know if you work for/with MDPI, to be fair. A confirmation of yes/no financial connection to MDPI greatly changes the context of your COI, and seems appropriate if you'd oblige. -- Crawdaunt (talk) 12:20, 4 October 2023 (UTC)

I'm really not comfortable saying more than what I've already said (which might even be too much). If there's an editing dispute that ends before Arbcom, I might say more in private evidence. Banedon (talk) 02:23, 9 October 2023 (UTC)

If you're not comfortable talking about your COI why did you choose to edit a topic which you have a conflict of interest with? When someone doesn't want to disclose details they simply don't edit the topic. I'd also strongly advise you to cease editing the page directly, you can propose changes on the talk page instead. Horse Eye's Back (talk) 05:34, 9 October 2023 (UTC)

You'd rather I not have disclosed COI? Truly?

Right now, academic publishing as a whole is giving me considerable grief, and comments like yours don't help - so I'm going to stop watching this page. Have fun. Banedon (talk) 01:48, 10 October 2023 (UTC)

I'd rather you'd never edited the page at all and confined yourself to the talk page as is best practice when you have a COI. It would be an excellent idea if you stepped away from the page, either temporarily or permanently. Horse Eye's Back (talk) 16:32, 10 October 2023 (UTC)

Thank you for your comments, Banedon and Mark. What exactly meets the "criteria for acceptable use of self-published sources"? In this instance, it seems to be an agreement between two individuals, both of whom have a COI. Banedon (who likely works for our competitor, as adding a preprint to only our article and not those of our competitors doesn't benefit us but could benefit them) wrote: "Yes, your paper looks significantly more honest than Oviedo-Garcia's paper, which is why I'm in favor of adding the results to the paper before peer review."

Is the subjective opinion of one Wikipedian with a COI, who may work for our direct competitor, enough to replace formal peer review and, more importantly, the general rules of Wikipedia regarding reliable sources (which should be even more relevant in sensitive matters like this one)?

It appears that nearly any negative news about MDPI is quickly added to our article, and Wikipedia policies seem to be supplanted by the subjective opinions of editors with COIs, as in this case. However, it's not surprising if this has been going on for a long time. Nonetheless, it does seem unusual that the primary source of "objective information" about our company is edited mainly by Mark, who, along with a co-author of this preprint and supporter Crosetto, have been quite critical of MDPI for years (and now we're seeing that the MDPI article is serving as a platform for their self-published works: blogs and preprints), and Banedon, who may work for our competitor.

@Randykitty Mdpi comms (talk) 11:14, 5 October 2023 (UTC)

I'll chime in briefly to say: I actually agree with @Mdpi comms that this is a preprint, and so it is not clear that it merits inclusion in its current state per WP:RS.

I can see a potential for WP:RS based on WP:USINGSPS as Paolo Crossetto is already cited as an expert source (per expert consult status from THE and Science articles). If the editor community here feels that it provides a better clarity of existing text(?), then I'm happy if we can improve the clarity of what is discussed. For instance, @Banedon's opinion that perhaps our article discusses things in a 'more honest' way. However I also agree with @Mdpi comms that this clearly needs a 3rd party to mediate its merit, as I am obviously COI'd and Banedon has said they are as well.

Cheers -- Crawdaunt (talk) 12:15, 5 October 2023 (UTC)

@Mdpi comms: I don't have a COI, and I think the citation is fine. Self published sources are usable when they are authored by subject matter experts. I get that you disagree with what the paper says, but casting aspersions as you have been on this talk page is itself a violation of Wikipedia policy. You should stop. MrOllie (talk) 12:27, 5 October 2023 (UTC)

I can see the argument that this is premature as well, hence I wrote "I'm adding my OR (or at least my interpretation) on top of Crawdaunt's OR" above. But it does look like a more honest paper than Ovideo-Garcia's, which had a strong smell of confirmation bias, plus many misleading statements to boot. Also not everything I added is negative on MDPI. But if someone thinks including this source now is inappropriate and reverts, I won't revert back. Banedon (talk) 02:23, 9 October 2023 (UTC)

Thank you for your comments. We apologize if any of our remarks can be perceived as aspersions. However, we have sought to clarify the issue by referencing clear facts about COI, as highlighted by Mark (aKa @Crawdaunt) and Banedon.

@MrOllie, could you please share with us the Wikipedia policy that states: “Self-published sources are usable when they are authored by subject matter experts”? We have found just an explanatory essay that declares this: “nor whether the author is a famous expert, makes any difference.” (https://en.wikipedia.org/wiki/Wikipedia:Identifying_and_using_self-published_works)

__

Nevertheless, before Mark started editing this article, even Crosetto wasn't considered an expert on this subject by some Wikipedia editors of this page. We notice that Crosetto has been active since then, but by adding his articles to the Wikipedia page could help Crosetto become “a subject matter expert,” as this page has a significant impact. So we would like to ask if adding these works by his collaborator and coauthor could be considered as COI. You can see the previous discussion on this Talk page before:

"Is there any reason to believe that this source rises above our usual prohibition against blog sources? Crosetto appears to be an economist, with no special expertise in bibliometrics or academic publishing beyond (like all academics) as a participant, so I don't think the "established subject-matter expert" escape clause of WP:SPS applies. —David Eppstein (talk) 18:54, 21 August 2021 (UTC)

Yes, I think we'd really have to stretch the "established subject-matter expert" escape clause to allow that here. XOR'easter (talk) 22:55, 21 August 2021 (UTC)

https://en.wikipedia.org/wiki/Talk:MDPI/Archive_3

__

Also, Mark, you mentioned here about your COI with the statement: “My coauthors and I began this study informally around April 2023.” However, your coauthor on Twitter stated that you started your collaboration “Back in 2022.” We are curious about which of these dates is correct and whether the COI would actually be considered for a longer period.

https://twitter.com/pagomba/status/1709835364301443388

__

Thank you, Mr. Ollie, for your declaration that you are not in a COI. However, as it now appears that some of the most active editors of this article just haven't declared their COI before, and have been editing it for a long time, we are not sure who else here may have a COI and who does not.

As the preprint states, MDPI's growth is most significant. Because of this growth, we have heard rumors that some of our competitors from academic publishing support initiatives that are against MDPI, and if one of the main editors of this article has declared that he works for our direct competitor and has now added a preprint which is highly critical of us and to our brand, we are giving it some more thought and investigation.

Is there an option for an independent review of this article by Wikipedians who are not directly involved in its editing to ensure it complies with Wikipedia policies? Or should we contact the Wikimedia Foundation to describe the case and ask for help?

@WhatamIdoing Mdpi comms (talk) 18:47, 9 October 2023 (UTC)

You have editors uninvolved in editing the page in the process of reviewing it. Note that account sharing is not allowed, if you want to be "we" you will need to create an account for each of "we." Horse Eye's Back (talk) 18:52, 9 October 2023 (UTC)

We apologize if any of our remarks can be perceived as aspersions. A classic Non-apology apology followed up with more aspersions. Disappointing. The policy I was referring to was the same one you initially quoted. The bit about subject matter experts is the bit you elided with an ellipsis. MrOllie (talk) 19:01, 9 October 2023 (UTC)

Hi, happy to respond per my own note, as @Mdpi comms brings up a fair question on when my collaboration with the group started as public messages of 2022/2023 might be confused: No that is not correct. Pablo and Paolo were working together (unbeknownst to me) back in late 2022, which was the work that was featured by the Science magazine news department and Times Higher Education in March 2023 (links below). This is what Pablo means (I think) when he says he started on this back in 2022.

Science: https://www.science.org/content/article/fast-growing-open-access-journals-stripped-coveted-impact-factors

THE: https://www.timeshighereducation.com/news/quality-questions-publishers-growth-challenges-big-players

The 'expert' status of Paolo Crosetto would be established beyond reasonable doubt around this time per WP:SPS, given he was being consulted as an expert by reputable third parties (Science Magazine News department, Times Higher Education). As mentioned above, I think I added Paolo's 2020 blog to the page in the wake of this as a WP:SPS because it was the primary data source behind the articles written, and could be used finally given these articles' establishing Crosetto's expert status on the subject. For further context, see: https://en.wikipedia.org/wiki/Wikipedia:Identifying_and_using_self-published_works#Self-published_doesn't_mean_bad

"Self-published sources can be reliable, and they can be used (but not for third-party claims about living people). Sometimes, a self-published source is even the best possible source or among the best sources."

Again, there was never any formal (or even informal) collaboration between myself and the others until April 2023. I've just checked my old Twitter messages, so I'll revise and confirm the exact date even: I reached out to Paolo and Dan March 24th/25th 2023, following the Clarivate delistings and following the articles in Science/THE. In fact, at this time I also reached out to Christos Petrou, hoping to bring in an author who might push the collective on some opinions, with the hopes of being as unbiased as possible. Christos declined for personal reasons. It was only starting around early April 2023 that things formalized into a collaboration (at least one that included myself). I had hardly spoken with any of my co-authors prior to this, and indeed those were my first ever Twitter private messages to Paolo and Dan on March 24th/25th 2023. I firmly confirm that I had no collaboration with my co-authors until ~April 2023.

-- Crawdaunt (talk) 19:20, 9 October 2023 (UTC)

@Mdpi comms, Wikipedia:Identifying and using self-published works says that works self-published by a famous expert are still self-published. Wikipedia:Verifiability#Self-published sources says that works self-published by a published subject-matter expert can be cited in Wikipedia articles (and that self-published works by non-experts generally cannot be cited in Wikipedia articles). WhatamIdoing (talk) 01:13, 10 October 2023 (UTC)

See also WP:PREPRINT, which may be more relevant.

@Crawdaunt, if there's still a chance of changing the text, then the line due to copyright concerns over our web scraping of information in the public domain needs a re-write. There cannot be any copyright concerns over works in the public domain. WhatamIdoing (talk) 16:03, 10 October 2023 (UTC)

Some sources (e.g. Scimago) have legalese in their terms of use that we cannot republish their data as an independent database (those data, in that format, are property of Scopus/Scimago). The scraped article data are also public domain, but not in the format we put together, etc... and similar concerns apply in a company-by-company fashion looking at their terms and legal rights to text mining across our institution countries. We had proper legal council on this. Ideally, when we do update this text, it will be with a confirmation that we can share the data freely!

We expect to be able to, but are following the advice we received for now. We also want input from our eventual journal to be sure we have full compliance and agreement from the journal on data availability since it will be tied to them (even if ultimately indirectly by hosting on an external site). Best -- Crawdaunt (talk) 17:12, 10 October 2023 (UTC)

MDPI, Hindawi, Frontiers blacklisted in Malaysia

Bringing to attention of page. Follows from similar blacklists of universities in Czechia and China (already on page).

https://tuoitre-vn.translate.goog/3-nha-xuat-ban-lon-vao-danh-sach-den-20230927113809178.htm?_x_tr_sl=auto&_x_tr_tl=en&_x_tr_hl=en&_x_tr_pto=wapp

Cheers -- Crawdaunt (talk) 12:43, 1 October 2023 (UTC)

Interesting — thanks for the link. —David Eppstein (talk) 18:54, 1 October 2023 (UTC)

How reliable is this source? I can't seem to find a primary reference. The closest is this tweet, which is still a Twitter post and not very reliable. I'd certainly add this to the article if a primary reference can be found, but as it is I am holding off. Banedon (talk) 01:21, 2 October 2023 (UTC)

I was just looking into this out of curiousity. It's strange that I can't find any official post by the Ministry of Higher Education Malaysia (maybe my searches are not sensitive enough?). But it is now circulated by other Malaysian news sites, and also English sites including librarylearningspace.com. It would be scholarly/journalistic for the original source to at least link to the official announcement... But looking up the original article source it seems like this is the largest newspaper in Malaysia (Tuoi Tre), so it is somewhat a WP:RS itself. Would be good to watch for any firm verification of this from the Malaysian Ministry of Higher Education, or even a Malaysian university making a statement in response (verifying its legitimacy). As it stands, in terms of available paper trail, it seems like a generally reputable source has broken news on something with no official memo.

-- Crawdaunt (talk) 07:55, 9 October 2023 (UTC)

None of these are Malaysian sources. Vietnam and Malaysia are different countries. I'm not convinced the sources are reliable. Banedon (talk) 01:43, 10 October 2023 (UTC)

Ah, you're correct, I was confused. Tuoi Tre is Vietnamese. So yeah, this is now picking up steam, but I agree there is no confirmatory source that would qualify as a WP:RS. Even if Tuoi Tre is a major paper, it is odd they do not link out to a statement, or otherwise provide evidence of this supposed blacklisting. -- Crawdaunt (talk) 11:31, 10 October 2023 (UTC)

Redoing the search three months later, there is this source by the ACS and this one by what looks like a Malaysian source. They are somewhat contradictory though. The first one makes it seem like the ban is because of high publication fees, while the second one says it's about academic integrity. The quotes by the publishers in the first source does suggest that they perceive the decision as one about academic integrity, however. I won't add either to the article, but someone else reasonably might. Banedon (talk) 04:02, 1 January 2024 (UTC)