Two months after asking the question of this forum and one month after starting my small investigation, I would like to take the opportunity to draw some preliminary conclusions:

(1) Time-loops can be considered as a serious and continuous problem in academic publishing. They potentially distort citation metrics and make it potentially difficult to appropriately cite papers.

(2) Time-loops are not a specific problem of Elsevier. Many other publishers that use “first-online” have significant time lags between the first online publication and the final assignment to a journal issue. Partly, the problem is related to print versions of the respective journals.

(3) Currently, no other publisher than Elsevier uses time-loops in such a systematic way and assigns papers published at a certain point in time to a journal issue far in the future “final and fully citable”.

(4) Time-loops create particular problems for citations and journal metrics when first-online and final issue publications are shifted between years.

(5) Using the year of the first-online publication is the most promising way to fix the problems with regard to publication metrics. This solution was already suggested by Tort et al. (2012). It’s not so sure how it should be handled in scientific referencing.

(6) Clarivate Analytics is aware of the problem and wants to change the calculation to the “early access” principle, i.e. the date of the first publication.

Overall, I think there might be progress and an increasing awareness. I also like to inform you about the extent and results my investigations and the responses so far:

- Clarivate Analytics: very prompt, kind and serious response, informative and to the point

- Elsevier: no response after more than a month

- Journal of Cleaner Production: no response after more than a month

- Science of the Total Environment: no response after more than a month

- Ecological Economics: prompt, kind and serious response

- Land Use Policy: very prompt, kind and serious response

I am disappointed that the Journal of the Total Environment engages in these practices. I did not have much regard for the Journal of Cleaner Production. I have to confess I know very little about how would go about "gaming" Impact Factors.

But I see IFs as an example of a heuristic that allows people to make judgments about the quality of work without knowing very much about bodies of work.. I have reviewed some disappointing work based on electronic analysis of citations that purported to measure which authors were central in my subfield. It was evident that the authors were relying on IFs and similar measures without knowing enough about foundational works and theoretical debates in the subfield.

I did not call this time loop. Many journals put articles "online first", woth DOI and these are fully citable. Yes, they earn citations for the future volume and sometimes for the future year. But for the author, this practice is OK - would you like to wait a year bofore your accepted paper will be published in a printed volume?

On-line journals and paper versions of same journals usually have different ISSNs as well. Therefore, there is no time-loop.

I agree to Linas Balciauskas and Witold Bogusz . Allocating "online first"/"early view" articles to future issues is a good way to make quality research available faster and gain citations (with final references) to articles early on. This is particularly important in quality journals where the peer-review process can already take 1 to 2 years - it is then unaccaptable to further delay the final allocation to journal issues.

As said by others it is a good thing that a paper is available online ASAP once accepted for publication.

However, the point Volker Beckmann is trying to make is whether it is almost intentional to make papers citable while the actual assignment of issue number etc. takes a ridiculous long time. See for example:

https://www.sciencedirect.com/science/article/abs/pii/S0959652620330286

The details for this paper are: Received 23 August 2019, Revised 16 June 2020, Accepted 19 June 2020, Available online 18 July 2020. It will appear in true volume 20 December 2020! This by itself is indeed a consequence of holding on to a system of a century ago, but still the side-effect is that papers are already citable and might influence impact factor (in a positive way).

What about: https://www.sciencedirect.com/journal/current-opinion-in-behavioral-sciences/issues See for example this paper: https://www.sciencedirect.com/science/article/pii/S2352154620300930

It is online 22 July 2020 and will assign an issue number in April 2021. This definitely will influence the number of possible citations since the paper is already out there for 9 months before it is ‘officially indexed’. I am not sure but I think that the number of citations in 2020 will be added to the citations in 2021 and 2022 used for the impact factor calculation.

It looks to say the least unnecessary, even a journal that suffers from way more submissions than they can actually publish like "Biophysical journal" goes no further than their current issue August 2020. So, subscription based journals that still work with issue numbers can function properly, one might wonder if having such a gap between fist online publication and actual issue release is somehow beneficial for the impact factor.

Best regards.

Anyway, I agree that it is a form of not quite honest manipulation. It is true that waiting one or two years for the true publication is frustrating, but a solution should be reduction of the publication time - why in XIX century a paper submitted in, say, August might have appeared in print in September or October, and modern editors and printers, equipped with sophisticated computer-programs and so many other facilities available in XXI c., need many months or even years? Moreover, the custom of "publishing" "online first" versions cause a chaos: it is not always evident (although at least in my speciality, taxonomy, often critically important!) if the text before me is an online-only publication, online preprint of a paper to be really (in print) published later, or pdf of the final, printed version!

Volker Beckmann that's an interesting question. I keep wondering about this phenomenon for many years. The situation is the same for most Elsevier chemistry journals. As can be seen in the attached screenshot, the November issue of Coordination Chemistry Reviews has already "appeared". However, I don't see no reason to worry about this.

Thank you all very much for your thoughtful answers.

I think there could be systematic manipulation which increases the IF of Elsevier journals using time-loops and reduces the IF of other journals or publishers which don't use time-loops. There are two mechanisms at work which I like to reflect on.

Let me take Elsevier’s Journal of Cleaner Production and the MDPI’s journal Sustainability as an example. The Journal of Cleaner Production (like many other Elsevier Journals) publish papers fully citable four months before the official publication (time-loop). Sustainability (like all other MDPI journals and journals of many other publishers) publishes papers immediately when they are ready (without any time-loop).

Let’s further take the example of IF 2021. The IF in 2021 is calculated as the citations in 2021 of papers published in 2020 and 2019 divided by the number of citable papers in 2020 and 2019.

Citations need time. Researchers need to recognise the publication and find them relevant. Those papers need to be published which takes time again. Time-loops may then create advantages.

If I write a paper now in August 2020, I can cite the publication of the Journal of Cleaner Production of September, October, November and December 2020 for instance and my paper might get published in 2021. My citation contributes to an increase of the IF 2021 of the Journal of Cleaner Production. Of course, I can’t cite a paper now which occurs in September, October, November and December 2020 in Sustainability for instance, simply because I don’t know the paper yet. It will be published in the respective future, and not in four months in advance.

This potentially creates a bias in the IF in favor of the Journal of Cleaner Production (and all other journals using time-loops) and disfavoring Sustainability (and all other journals without time-loops).

A paper which is published in the Journal of Cleaner Production (or in any other journal using four months time-loops) in January, February or March 2021 can’t cite a paper which was published in Sustainability (or in any other journal without time-loops) in December 2020, simply because they were completed in September, October or November 2020. However, it can cite a paper which was published in the Journal of Cleaner Production (or in any other journal using four months time-loops) in December 2020 because it was already published in August 2020.

This potentially creates a second bias in the IF in favor of the Journal of Cleaner Production (and all other journals using time-loops) and disfavoring Sustainability (and all other journals without time-loops).

Is four month a long or short period? It's a third of a year! I think a time-loop of four months is quite substantial and might well be considered as systematic manipulation of IF favoring major Elsevier journals and disfavoring all other journals and publishers not applying time loops!

I’m looking forward to your opinion.

Very good question. This practice can affect the Impact Factor.

I disapprove of this practice. They could have put the published papers of December issue in the August one. It deserves further investigation.

.

Please have a look at the Elsevier Journal Land Use Policy. Already the first papers of 2021 are published now: https://www.sciencedirect.com/journal/land-use-policy/vol/100/suppl/C

Volker Beckmann it's still not clear to me if this is a real problem or not. Isn't it just a time displacement of the IF calculation? Do we need to worship impact factors at all?

Yes, it deserves further investigation. I had the impression that in the years after my article in Learned Publishing, the practice of stating a publication date in the future diminished. Of course, this can be coincidence, although I wish to think that my little paper had some positive consequences. Now, eight years later, we are "back to normal" it seems, and false publication dates become ubiquitous again. It can only have positive effects for publishers in terms of impact factor increase, but the publication dates are false. Shouldn't we try to have those metadata correct? And, no, we should not worship impact factors, but it is widely done, explicitly by administrators in many countries, implicitly by most of us. Don't we admire a colleague who has managed to get a paper accepted in Nature or Science? - So for the time being, I collect such discrepancies between stated and real publication date and prepare another paper. But it will take a while. For Christmas, please buy some newly published academic books, and if you do, check out the publication date. There is a big chance that you find many or most of them published in 2021. No impact factor here, but the book stays new for longer which helps sales.

The IF is calculated based on citations an article receives within 2 years of publishing (attaching to an issue). The CiteScore is determined based on citations an article receives within 3 years of publishing (attaching to an issue).

Witold Bogusz: I think I am also a co-author on the same paper. It was an Elsevier Journal. The paper was published online in 2018 and attached to an issue only in 2020 after 10+ citations. Indeed the Journal is manipulating their IFs and CiteScores. How and why the paper was accepted is another scandal.

There are several other ways Elsevier Journals manipulate their IFs. A top Elsevier Energy journal with an IF of 8+ and CiteScore of 16+ desk rejected one of our articles giving the following reason. Obviously the Editor wanted us to cite the recent articles published in their Journal to increase their IFs and CiteScores. We had to oblige and the article was published As-Is.

"You should carry out a thorough literature survey of papers published in a range of top energy journals in the last three/four years so as to fully appreciate the latest findings and key challenges relating to the topic addressed in your manuscript and to allow you to more clearly present your contributions to the pool of existing knowledge. "

Time-loops might manipulate Elsevier’s new CiteSore indicator at a large scale as well.

Elsevier recently introduced a new CiteScore metrics dividing the citation to publications of four years by the publications of the respective four years: https://blog.scopus.com/posts/citescore-2019-now-live

Journals have a big advantage if the official publications are always four to six months ahead. They are more likely to be cited in the same year or the year after.

Maybe the next Elsevier innovation is that citation to future publications will be counted as well. Sure, it's possible to cite Elsevier papers now which will be published officially in 2021. Land Use Policy for instance already published 53 papers in 2021: https://www.scopus.com/sourceid/14500#tabs=2 (Click on Scopus content coverage)

Funny, if it weren't so serious.

Thanks to Eric Fong, University of Alabama in Huntsville, I learned about some more interesting papers on the issue. The phenomenon is sometimes also called "online queue stratagem" or just “publication delays” in the literature, as some of your responses also indicated. To me it seems, however, that major Elsevier journals systematically apply time-loops.

Let me share the recommended literature with you.

B. Bjork, D. Solomon. 2013. The publishing in scholarly peer-reviewed journals. Journal of Infometrics, 7(4): 914-923.

C. Hopp, G.A. Hoover. 2017. How prevalent is academic misconduct in management research? Journal of Business Research, 80: 73-81.

B.R. Martin. 2016. Editors' JIF-boosting stratagems -- which are appropriate and which are not? Research Policy, 45(1): 1-7.

A.B.L. Tort, Z.H. Torgino, O.B. Amaral. 2012. Rising publication delays inflate journal impact factors. PLoS One, 7(12): e53374.

A. Wilhite, E.A. Fong, S. Wilhite. 2019. The influence of editorial decisions and the academic network on self-citations and journal impact factors. Research Policy, 48(6): 1513-1522.

In particular, I recommend Tort et al. (2012) because they analyzed the impact for the journals applying time-loops very clearly. See:

https://journals.plos.org/plosone/article?id=10.1371/journal.pone.0053374

To the best of my knowledge the possible adverse effects of this practice on the IF of journals that don't apply time-loops were not estimated so far.

I also recommend reading:

Hickman, C. F., Fong, E. A., Wilhite, A. W., & Lee, Y. (2019). Academic misconduct and criminal liability: Manipulating academic journal impact factors. Science and Public Policy, 46(5), 661-667.

https://academic.oup.com/spp/article-abstract/46/5/661/5488509

Where to find papers of Elsevier journals that will be published officially in 2021 (final and fully citable) now?

Here are the results of my small research.

- You will find them on ScienceDirect (with the year of publication and the hint that they are final and fully citable)!

- You will find them on Scopus (with the year of publication).

- You will find them also on Google Scholar, but without indicating the year of publication!

- You won’t find them on WoS!

- You won’t find them on SciLit!

Check it!

With Scopus Elsevier has an analytical tool that potentially allows to immediately assess the possible impacts of delaying or speeding up publications on IF and CiteScore! Not only for the individual journal but for all journals in their database.

Please have a look at the journal Science of the Total Environment. Nine volumes in progress ranging from 10 November 2020 until 15 January 2021. Lots of options to manipulate, alone by simply shifting papers between 2020 and 2021.

https://www.sciencedirect.com/journal/science-of-the-total-environment/issues

Same for the Journal of Cleaner Production:

https://www.sciencedirect.com/journal/journal-of-cleaner-production/issues

And others more.

This is not an exclusive strategy of Elsevier. It is a common practice among scientific journals from other publishers such as Springer, Wiley, Taylor & Francis among others that publish citable articles online, before being assigned to an issue.

Dear Nuno Guiomar ,

yes, I agree. However, with major Elsevier journals the practice has reached a new quality, I think. I first recognized the dimension when I visited SciLit in January 2020. Hundred thousands of papers of the year 2020 published already in January!

SciLit provides very up to date information on the number of published articles per publisher (currently as of 1 September 2020. See: https://www.scilit.net/rankings

You need to click on the year and choose "2020 until now". You may want to check it in January 2021.

Another new quality is that the papers are fully and final assigned today to issues in the far future! As far as I can see from other publishers they publish "online first" but don't assign the paper to a future issue (final and fully citable).

I just checked the Francis & Taylor journal "Human Dimension of Wildlife".

Yes, there are time loops, but "online first" papers are not assigned to a far future issue. See: https://www.tandfonline.com/loi/uhdw20

The effects of these different practices on IF might also be different, but I am not sure. I would cite a paper under the rubric “Latest articles” of “"Human Dimension of Wildlife" as an article of 2020, whereas for the Journal “Science of the Total Environment” I would cite a paper that is official published in 2021 (final and fully citable) as a paper of 2021.

However, if you (or anybody else) know of any non-Elsevier journal that now assigns papers final and fully citable to issues of 2021, please let me and others know. Thanks!

Dear Volker Beckmann

If the Impact Factor determination did not consider citations prior to their formal publication in an issue, this problem would be resolved, right? If so, why not demand this "innovation" in the calculation, in the same way that the exclusion of self-citations has already been introduced.

All the best

Nuno

Dear Nuno Guiomar ,

yes, an "innovation" in IF calculation could be a solution. However, in order to get the incentives right, I would rather follow the suggestion of Tort et al. (2012): "Not only for calculation of impact factors, but for any measure that includes citations within a specific time window, we believe that it is more appropriate to use databases that index articles on the basis of the date when they are first available online. ... In this sense, a simple measure to avoid distortions such as the one described here is the indexing of articles by scientific databases on the date of their online appearance, rather than on that of their publication in print."

See: Article Rising Publication Delays Inflate Journal Impact Factors

I think that such a calculation of IF and CiteScore would fix the problem.

Dear Volker Beckmann,

I agree with you, it seems to be a more viable and easier solution to implement than the one that occurred to me.

Best regards from Portugal

Nuno

Thank you all for your informative and encouraging answers. I came to the conclusion that the issue needs to be investigated and decided to take action. I started to inform Elsevier, Editors-in-Chief of major Elsevier Journals and Clarivate Analytics about the case. Let's see if the publication practice and the metrics calculation can be changed to the better!

Let me share with you the preliminary results of my investigation.

First and most important, I was informed by Clarivate Analytics that they are aware of the problem of time-loops and already plan to fix it. They will change (probably next year) the calculation of IF based on the "early access" principle. Thus, the year of a paper's first online publication will count in the nominator and the denominator of the IF calculation. This solution was already suggested by Tort et al. (2012). If this principle will also be implemented by CiteScore, the problem of potentially manipulative time-loops on IF and CiteScore metrics is fixed.

Second, I also got kind responses from two Editors-in-Chief and a production editor of two major Elsevier journals. They mainly argue that they face problems with the increasing number of submissions and the constraints of the printed version of the journal. Further it was argued that the time-lag is necessary for producing the printed version and that it's difficult for print-journals to change the practice of "online first" without harming authors. They all took the concerns very serious.

Third, I did not get any response from Elsevier, except an automated email informing me that I will get a response within 48 hours. The 48 hours have passed by without any response so far.

Speculation with academic publications...it's like the futures market in financial stock markets...

Two months after asking the question of this forum and one month after starting my small investigation, I would like to take the opportunity to draw some preliminary conclusions:

(1) Time-loops can be considered as a serious and continuous problem in academic publishing. They potentially distort citation metrics and make it potentially difficult to appropriately cite papers.

(2) Time-loops are not a specific problem of Elsevier. Many other publishers that use “first-online” have significant time lags between the first online publication and the final assignment to a journal issue. Partly, the problem is related to print versions of the respective journals.

(3) Currently, no other publisher than Elsevier uses time-loops in such a systematic way and assigns papers published at a certain point in time to a journal issue far in the future “final and fully citable”.

(4) Time-loops create particular problems for citations and journal metrics when first-online and final issue publications are shifted between years.

(5) Using the year of the first-online publication is the most promising way to fix the problems with regard to publication metrics. This solution was already suggested by Tort et al. (2012). It’s not so sure how it should be handled in scientific referencing.

(6) Clarivate Analytics is aware of the problem and wants to change the calculation to the “early access” principle, i.e. the date of the first publication.

Overall, I think there might be progress and an increasing awareness. I also like to inform you about the extent and results my investigations and the responses so far:

- Clarivate Analytics: very prompt, kind and serious response, informative and to the point

- Elsevier: no response after more than a month

- Journal of Cleaner Production: no response after more than a month

- Science of the Total Environment: no response after more than a month

- Ecological Economics: prompt, kind and serious response

- Land Use Policy: very prompt, kind and serious response

Thanks for a timely questions/ raising an important issue. Here is another good finding.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3697906

How should we cite time-loop papers? Here a random set of papers recently published in Science of the Total Environment.

This is the citation according to Scopus:

Feng Zhao, Sabine Filker, Chunsheng Wang, & Kuidong Xu (2021). Bathymetric gradient shapes the community composition rather than the species richness of deep-sea benthic ciliates. Science of the Total Environment, 755, 142623. https://doi.org/10.1016/j.scitotenv.2020.142623

Roya Pishgar, Dean Morin, Shane J. Young, Jon Schwartz, & Angus Chu (2021). Characterization of domestic wastewater released from ‘green’ households and field study of the performance of onsite septic tanks retrofitted into aerobic bioreactors in cold climate. Science of the Total Environment, 755, 142446. https://doi.org/10.1016/j.scitotenv.2020.142446

Now check Google Scholar. You will get the following results:

Zhao, F., Filker, S., Wang, C., & Xu, K. (2020). Bathymetric gradient shapes the community composition rather than the species richness of deep-sea benthic ciliates. Science of The Total Environment, 142623.

Pishgar, R., Morin, D., Young, S. J., Schwartz, J., & Chu, A. (2020). Characterization of domestic wastewater released from ‘green’ households and field study of the performance of onsite septic tanks retrofitted into aerobic bioreactors in cold climate. Science of The Total Environment, 142446.

The time-loops as applied by Elsevier create a big mess for proper and ethical referencing. I think that as scientists we might have an obligation to cite these (and other time-loop) papers as of 2020 (more general the date of first access), ignoring the official assignment to an issue!

Please be aware that up to now Science of the Total Environment has published 872 articles of 2021 (according to Scopus as of October 2020)..

Dear Volker Beckmann The promise Clarivate made, seems to come true:

https://scholarlykitchen.sspnet.org/2020/12/07/jif-calculation-2021/

They are changing basically the moment from where they will count. From moment of the printed publication to the moment of the electronic publication.

A move in the right direction.

Best regards.