Following The Science...

...right down the rabbit hole. 

Higher Ed has been broken for some time, what with the publish-or-perish system creating aberrant and abhorrent behavior, but now, with Large Language Models, there is a deluge of fictional "scientific" publications. Retraction Watch has been diligent in rooting out suspect publications, but now the flood of publications, and suspect publishers and editors has the attention of none other than the Grey Lady. The Times article points to a Moore's Law doubling of suspect publications and notes the increase in publishers of sketchy journals with equally sketchy editors and reviewers on staff. 

This is indeed a serious problem. One that has been around for quite some time, but now is headed in the direction of bogus articles far outnumbering legitimate research. Remarking on corrupt paper mills as lucrative business operations does little, and what little is being done, relies on automated processing to detect doctored and fabricated data, and scan for absurd wording. Have you ever read an article in popular press that uses the words "furthermore," "what's more," or "moreover," then you are reading something spewed out by a LLM. Nobody really uses those words. In an effort to hide their plagiarism, these millworkers pump out drivel that includes similarly uncomfortable diction, as well as artificial hallucinations and citations that do not exist. 

The Grey Lady finishes up as she so often does of late: with a swipe at Trump. A postdoc is quoted saying that the "stuff that the Trump administration is doing" is going to make the situation worse. Yes. That's right. "Stuff" is the appropriate technical term. It is worth noting that neither the NYT author, nor any of the sources provided any kind of solution. None whatsoever.

Follow The Science

But watch out for the scientists.

Research, particularly research conducted inside the academy, is informally known as a "publish or perish" reality. Want tenure? Pump out those papers. Want a promotion or a raise? Keep 'em coming. 

And so they do.

To be very clear, just any old publication will not do. No vanity press here, these need to be accepted by peer-reviewed publications, and it should come as no surprise there is a pecking order to these publications. Some are just more prestigious than others. What seems to be consistent across the board is the review process.

Authors will tell you this process is burdensome, unnecessarily so, and feel some reviewers provide annoying criticisms, demanding fixes, because, well, just because they can. What is increasingly clear is this process does very little to ensure the scientific integrity or veracity of the research work. It should come as no surprise that once a prestigious journal publishes, they tend to turn a deaf ear to outside peers who find fault, scientific fault, research fault, with the publication. 

In one case, retraction of seriously flawed work took six years. In this time the paper was cited almost 150 times. Damage done. And lest you think this is an unusual or rare occurrence, Retraction Watch's database has 55,000 retractions of which over 450 are CoVid-19 research papers. So, if you are going to "follow the science" you need to read these papers with a bit more critical eye than the reviewers.