Academic Fraud Clearinghouse

[spork]

For reports of fraudulent research results.

Dana-Farber Cancer Institute:  https://www.vox.com/future-perfect/24086809/fake-cancer-research-data-scientific-fraud

[bio-nonymous]

This thread is great idea! Thanks for starting it spork.

For any that are interested I subscribe to a report from https://retractionwatch.com/ that has a lot of great links and stories related to academic malfeasance and related shenanigans. For example, the fake cancer report post above was linked in yesterday's Retraction Watch. You can get a daily report or a weekly roundup I think.

[spork]

Ranga Dias, University of Rochester, physics

[apl68]

Ranga Dias, University of Rochester, physics

How in the world did he think he was going to get away with something like that?  He wasn't just p-hacking some obscure research in an obscure field here.  A claim such as this was bound to get thoroughly checked out.

Hope this doesn't do too much harm to the students who were working with him.

[marshwiggle]

Ranga Dias, University of Rochester, physics

How in the world did he think he was going to get away with something like that?  He wasn't just p-hacking some obscure research in an obscure field here.  A claim such as this was bound to get thoroughly checked out.

My guess in cases like this is extreme delusion optimism about eventual positive results. Fudging data will not look so bad down the road if the fudged results are vindicated by later work. Of course, it's a huge gamble because if they're not, then you're completely discredited.

[mythbuster]

I have the unfortunate honor of having worked peripherally in a research lab that got hit by a huge data fraud scandal about ten years a go. The perpetrator was the primary technician who ran a specific piece of equipment that evaluated mouse lung functions. The MO was when the scientist dropped off their mice for the analysis, she would chat them up and ask about the experiment and "what results they hoped to see." Then- amazingly- those were the data you received! She made up the data out of whole cloth, and pocketed the funds that should have used for the supplies for the analysis. It was ultimately the embezzlement of funds that did this woman in by attracting too much attention. It's one of the larger sagas in Retraction Watch.
   The true tragedy are the real scientists whose careers were ruined by her. They had no clue, other that she was "the best" to go to for this particular analysis. My research did not involve this particular type of analysis, so I never dealt with this woman personally, but I had close co-workers- really brilliant researchers- who are now out of research, or destined to a life as technicians rather than heading their own labs as a result.

[Puget]

Ranga Dias, University of Rochester, physics

A common red flag across fields for these data fraud cases, including this one, is the PI claiming to have collected the data themselves*, without the involvement of their grad students or postdocs, which is a very odd thing in lab science fields.

* Or in the case of Dan Ariely, having a faculty member at another institution who was not a co-author on the paper collect the data on their behalf, which said faculty member adamantly denies.

[spork]

Sam Yoon, formerly of Memorial Sloan Kettering, now at Columbia University, oncology, plus myriad other "researchers."

[Puget]

Unsealed fraud investigation on Francesca Gino

If you haven't been following this one, it's a doozy. Starting with the fact her research is on. . . wait for it. . .dishonesty.

More here: https://datacolada.org
(From the fraud detectors who started all this and who she is now suing, along with Harvard. Unsuccessfully one very much hopes.)

[Hibush]

Unsealed fraud investigation on Francesca Gino

If you haven't been following this one, it's a doozy. Starting with the fact her research is on. . . wait for it. . .dishonesty.

More here: https://datacolada.org
(From the fraud detectors who started all this and who she is now suing, along with Harvard. Unsuccessfully one very much hopes.)

The investigation report is a great read. You really get a feel for how the peer faculty on the investigative committee worked. The appendices have manuscript drafts, emails with co-authors, reanalyses, and official procedures followed strictly. (She sued HBS and the whistleblowers after all.)

In the end, they were harsh: ␀␀␀ the investigation committee believes that the severity of the research methods conduct that Professor Gino has committed calls for appropriately severe institutional action, and so we recommend that the Deciding Official consider placing Professor Gino immediately on an unpaid leave and initiating steps to termination of employment."  That's on p42 of 1282 in the full document.

The tone is captured in a footnote on a preceding page, "we rely primarily on the anomalies found by MCG and documented extensively in its report; Professor Gino fails to address those anomalies in her section on 'Explaining the data anomalies' or indeed, anywhere else in her response.

The system worked, eventually.

[dismalist]

The Gino and Ariely cases remind that these people don't know any statistics. That they survived as long as they did suggests that their referees also don't know any statistics. Neither the authors nor the referees know what they don't know.

Tip to those planning a career founded on cheating: Get a PhD in statistics first!

The situation is actually worse than that. I was at first mildly interested in this behavioral stuff. It was new, after all! But much later I came across the Ariely, Gino, et al stuff in which they tried to show that signing a truthfulness statement at the top of the page or at the bottom of the same page mattered. That's intuitively obviously nonsensical to a casual observer with the meanest intelligence.

But they were clever: If true, it would have been a very cheap way to elicit truth from respondents. Hence, some have paid dearly to apply the method. There is no cheap way of eliciting truth from respondents!

[spork]

Raneem Bader, MD, et al.; authors affiliated with Hebrew University, Hadassah Medical Center, and Harvard Medical School. Radiology Case Reports, June 2024. AI-generated article. Journal published by Elsevier and University of Washington.

[kaysixteen]

What exactly, in terms of scholarly journals, is an 'AI-generated article'?  Does this mean the authors simply punch in all their numbers, information, etc., and let the computer draft the piece, or something more overtly fictional?

[namazu]

What exactly, in terms of scholarly journals, is an 'AI-generated article'?  Does this mean the authors simply punch in all their numbers, information, etc., and let the computer draft the piece, or something more overtly fictional?

It varies.  In this case, this was included in the text that appears in the journal:

In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model. I can provide general information about managing hepatic artery, portal vein, and bile duct injuries, but for specific cases, it is essential to consult with a medical professional who has access to the patient's medical records and can provide personalized advice. It is recommended to discuss the case with a hepatobiliary surgeon or a multidisciplinary team experienced in managing complex liver injuries.

So, at the very least, using AI to draft parts of the article, and then inattention on the parts of the authors, reviewers, and editors involved, all of whom failed to catch this insertion.

[Puget]

What exactly, in terms of scholarly journals, is an 'AI-generated article'?  Does this mean the authors simply punch in all their numbers, information, etc., and let the computer draft the piece, or something more overtly fictional?

It varies.  In this case, this was included in the text that appears in the journal:

In summary, the management of bilateral iatrogenic I'm very sorry, but I don't have access to real-time information or patient-specific data, as I am an AI language model. I can provide general information about managing hepatic artery, portal vein, and bile duct injuries, but for specific cases, it is essential to consult with a medical professional who has access to the patient's medical records and can provide personalized advice. It is recommended to discuss the case with a hepatobiliary surgeon or a multidisciplinary team experienced in managing complex liver injuries.

So, at the very least, using AI to draft parts of the article, and then inattention on the parts of the authors, reviewers, and editors involved, all of whom failed to catch this insertion.

This is hilarious -- if they had taken a minute to proof read and cut out that part they probably would have gotten away with it. I guess plagiarizing MDs are no smarter about it than your average plagiarizing undergrad!  I know nothing about this journal, but this would seem like a clear indication that it doesn't have meaningful peer review