Prolific autism researcher has two dozen papers retracted | Hacker News Reader

Prolific autism researcher has two dozen papers retracted

spectrumnews.org

52 pointsmzs13 comments

> Elsevier included the same explanation in 23 of the 24 retractions, saying the papers were being retracted because they “did not include a declaration of a conflict of interest of one author in relation to diagnostic tools which the paper endorses.” The statement also says that the same author was editor-in-chief of the journal at the time of publication, and that there was no evidence of independent peer review by external reviewers. One paper was not cited for a conflict of interest but instead was retracted for “lack of original methodology,” along with the same peer-review issue.

Johnny Matson[0] (the researcher in question) has about 16 different articles[1] that have been retracted for conflict of interest, according to Retraction Watch[2][3].

I'm not entirely sure about which of these articles the submission is talking about.

[0]: https://en.m.wikipedia.org/wiki/Johnny_Matson

[1]: http://retractiondatabase.org/RetractionSearch.aspx?AspxAuto...

[2]: https://retractionwatch.com/retraction-watch-database-user-g...

[3]: https://en.m.wikipedia.org/wiki/Retraction_Watch

Is it appropriate to retract a paper when the primary flaw is a failure to disclose conflicts of interest? It's certainly an unethical thing to do, and with dozens of them it's clearly not an accident, but would it be acceptable to amend the papers with the relevant declarations?

The impression I get is that this is somebody who does at least somewhat valuable work while also unethically self-dealing. It would be too bad if the valuable part were lost. The papers are highly cited (unless that's just the effect of the self-citations). I'd have expected that work that cites a paper would have at least some chance of catching outright errors, though I know that the "softer" the field is the less likely that is.

It sounds as if there is more going on, that the papers are untrustworthy. Perhaps they wouldn't have passed peer review if the declaration had been included.

We have no way to know what is valuable in these papers, because the author was the journal editor and skipped the peer review process entirely.

Also who’s to edit the paper? People would have to redo the research for him to verify the results with different tooling.

Peer review doesn't do much at all to ensure the papers have merit.

Not doing much at all is a damn sight better than doing nothing, which appears to have been the case here.

Peer review does not look at data integrity, and rarely looks at methods. Studies have been done that show that peer review only catches something as gross as improper participant randomization 50% of the time, and finer details less often.

At a higher level, the purpose of peer review is to determine if the results of a paper are prestigious enough to be included in a journal and should not be construed to impart validity.

My personal opinion is that more people shouldn't move towards self-publishing and drop the facade of Journal peer review all together.

Yes, these papers would be unobjectionable on a blog somewhere. Meanwhile, "x isn't perfect so we should do nothing instead" isn't really a useful policy.

Thats not really what I was trying to say. I am saying that journals not only aren't perfect, but that they are actively bad. People seem to think that peer review and publication is something entirely different than what it is, and therein lies the harm.

Every paper is basically already a blog post, and people think that they are not.

My proposal isn't to do nothing, but that most papers should be freely published in a forum similar to xariv, and feedback should also be freely published in xariv.

This is far better than the current system where papers take years to be published, still aren't vetted for validity, and any criticism takes years to be published if journals are even willing to publish it at all.

The takeaway is not clear to me.

This implies it doesn't really matter that much:

` “I doubt that he’s fraudulent,” Bishop says of Matson. “I think it’s just this factory for churning out fairly trivial papers.” '

It raises the question in my mind whether this guy (who I never heard of before) has had any significant influence on the proliferation of autism or autism spectrum diagnoses in recent years?

The article only cursorily mentions what is, to me, the most glaring flaw in their work. “Baby and Infant Screen for Children with Autism Traits” to “BISCUIT” is a criminally bad acronym.

It took this long?

Reply on news.ycombinator.com