Plagiarism in that terrible BMJ Public Health article by Mostert et al

Much has already been written in blogs, newspapers and on X about the article “Excess mortality across countries in the Western World since the COVID-19 pandemic: ‘Our World in Data’ estimates from January 2020 to December 2022” that appeared in BMJ Public Health on 3 June. I can summarise the criticism with that this article adds nothing to what we already know about excess mortality, that it ignores many analyses carried out by others into the possible causes, and that the Introduction and Discussion do strongly suggest that it could have something to do with the mRNA vaccines after all. One aspect has been somewhat underexposed, namely that what is supposed to pass for ‘original research’ is for quite a bit rather silly plagiarism.

Update 26 November 2024: Princess Máxima Center has completed its investigation.

Mostert et al use data and a model developed by Ariel Karlinsky and Dmitry Kobak. They do refer to their work repeatedly, but the way they use their 2021 eLife article can only be considered plagiarism.

A superficial comparison of Mostert’s article with that of Karlinsky & Kobak, tells you that at least they took the mathematical description of the model almost verbatim. At first, this did not strike me as odd in itself; it is not that easy to paraphrase mathematical text really well. But why would you? If it is so important to include in your article, you could also have done with a literal quote. But if you don’t do anything at all with the specific properties of the model (the variances e.g.), then why not refer the interested reader, for those details, simply to the original article?

I noticed that Mostert et al even made some small mistakes in copying the text. Small textual mistakes, but which someone who actually understood the mathematics of the model would not make, in my opinion.

But when I looked at other parts in which the authors refer to Karlinksy & Kobak, I noticed more remarkable things. The most striking example is this one:

I don’t know about the reader, but I first had to look up “chasteness” in the dictionary to understand what they could mean by it. However, the comparison with the original text made it clear: “flexibility and simplicity” had been transformed into “suppleness and chasteness”…
Another example is when they ‘transformed’ the sentence “Note that our projection for 2020 uses a linear trend (see below) and so can implicitly account for improvements in death registration over the recent years.” from Karlinsky & Kobak into “Because the projected baseline uses a linear trend, the model can also reckon for ameliorations in death registration across recent years.”

Here Mostert et al have tried to disguise their plagiarism by using synonyms, like high school students do when they copy-paste from Wikipedia. These can be considered ‘tortured phrases‘, but I don’t think it points to the use of an LLM like ChatGPT as Professor Peter Openshaw suggested in the Daily Mail.
But how could these experienced scholars not have realised that this text did turn out very odd? I guess that this has to do with the mysterious second author: Marcel Hoogland.

It was quite a puzzle to find out more about Marcel Hoogland, but I am now certain that he is the partner of first author Saskia Mostert (Maarten Keulemans found the first clue in Mostert’s thesis). His background is in finance, and he doesn’t seem to have any experience in science.
At the end of the article in BMJ Public Health, we can read which of the authors contributed what. I think that at least the second mention of ‘MH’ refers to Hoogland (and not to Minke Huibers) and that in fact, the whole idea for this article arose at the kitchen table of the Mostert-Hoogland household.

NB At this moment I don’t think it would add anything to explicitly explain how I found out which Marcel Hoogland is the Marcel Hoogland of this article, as it would involve showing material that was probably not intended for a wide audience. But I’m willing to share this with serious journalists.

For the other issues with the article, I refer to the blogs I linked to before. Science journalist Maarten Keulemans broke the news that the Princess Máxima Center, where three of the four authors work, distances itself from publication in de Volkskrant. He has provided an English translation of his article on his personal website. Retraction Watch mentions the coming Expression of Concern.

Update: 26 November 2024
Princess Máxima Center has completed its investigation.

The medical center issued a news release this morning reporting that the investigation has been completed:

The CWI concluded that while there was no breach of scientific integrity, several principles of the Netherlands Code of Conduct for Research Integrity were violated during the publication process. However, the CWI found no evidence of intent, plagiarism, or data falsification by the researchers at the Princess Máxima Center. As a result, the conduct of these researchers does not qualify as a violation of scientific integrity. The investigation is complete but not yet finalized, as those involved may still request the National Board on Research Integrity to review the findings. Once final, the Princess Máxima Center will make the findings and conclusions publicly available in an anonymized format.

Presumably, this news is brought at this time because science journalist Maarten Keulemans is reporting on the case today in de Volkskrant based on the confidential report of the Committee for Scientific Integrity (CWI). This is because the procedure is not yet fully completed, the two employees on whom the Board of Governors is passing judgment can still request a second opinion from LOWI. I will probably return to the committee’s study in more detail later (which I also have in my possession), for now, a few comments.

Only the two least involved authors participated in the investigation, Gertjan Kaspers and Minke Huibers. The Mostert-Hoogland couple did not respond. Marcel Hoogland could not be forced to do so because Princess Máxima Center did not employ him. And Saskia Mostert is no longer employed there; she resigned shortly after all the fuss broke loose.

I am somewhat surprised that the committee says it found no evidence of plagiarism. I think my piece above clearly shows that there is. No doubt the committee read this blog, as it is mentioned in the report. However, I suspect that the outside expert who reviewed the content of Mostert et al.’s article did not and that the committee members adopted her opinion and did not investigate this issue further.