TL;DR: thousands of different authors use the same template emails to communicate with journals. Why?
So, we’ve established that papermills like to use templates. We see templates in referee reports and in the text of cookie-cutter research papers. There’s an important insight here:
- Templates are used in legit academic behaviour as well as in industrial research fraud. So we need to be careful — we can’t assume that templates are necessarily a sign of research fraud.
- But if a signal is part of an industrial process, it will be much more common than a signal in research outputs because it is much cheaper and faster to produce fake research industrially than it is to do real research, so the volume is higher.
This wasn’t obvious to me at first, but it seems self-evident now and it applies to a lot of the templates I’ve picked up.
Templates in author correspondence
But what about author correspondence? Here’s where it gets interesting.
Around a year ago, I was looking at some author correspondence which had been published by journals using transparent peer-review. It happened that the data I was looking at was a set of papers that I already knew was likely to be milled.
I noticed in the author correspondence that thousands of different authors had used a phrase something like ‘thank you for your warm work’ in response to a round of referee reports.
It’s clearly a grammatical error, but for thousands of authors to make the same error seems highly unlikely. It must be a template.
Here are some examples in Google and Google Scholar. In fact, just try Googling:
site:somepublisher.domain “warm work earnestly”
…and you will see how common this template is.
So we know that papermills are using this template in their correspondence. (In fact, it’s a set of templates covering all author-journal interactions.) Smoking gun, right? All we need to do is search for this phrase and others from that set of templates and we’ll find loads of milled papers! Right….?
Wrong!
That’s actually exactly what I did. I searched for and then manually sifted through dozens of these papers and, besides the ones that I started with, I couldn’t find a lot of obvious signs of fraud.
With a little digging, I found the same templates being shared online as far back as 2011. So, it looks like these templates are actually in widespread use and, while mills do use them, the templates appear to also be used in the correspondence of legitimate papers.
Would I recommend that editors look out for these templates? Yes. I think editors should be aware that this pattern exists, but I’d caution against basing a decision on this particular signal alone. Most of the time, the ‘warm work’ papers seem to be fine and using a template isn’t a problematic thing to do in author correspondence. I would also recommend that authors use template correspondence if it helps them to get honest work published.
Going back to what I said at the start, it seems likely there is some industrial process involved, but that could still be a lot of innocent things (as well as papermills).
The most recent site that I found offering these templates is citexs.com. Opening up that site was like lifting up a rock, finding a swarm of bugs and then hastily putting the rock back. How big does the need for services like automated paraphrasing, plagiarism detection (for authors) and AI-weight-reduction have to be for a site like this to exist?
Food for thought.
I’d like to thank Sholto David for helping look into the warm work papers in recent weeks. I recommend his YouTube channel and this recent interview with him in the Guardian.
