A paper published in Measurement in Physical Education and Exercise Science has been retracted for statistical and typographical mistakes.
Here’s the notice for “Comparing Measurement Error Between Two Different Methods of Measurement of Various Magnitudes”:
Due to errors, the statistical analyses of the manuscript titled “Comparing Measurement Error Between Two Different Methods of Measurement of Various Magnitudes” by Gerald S. Zavorsky (volume 14, issue 4, pages 265–274) are incorrect. For this reason, the aforementioned published article is officially retracted from this journal. Readers who are interested in a corrected version of the manuscript can contact the author directly (e-mail: [email protected]).The retracted article will remain online to maintain the scholarly record, but it will be digitally watermarked on each page as RETRACTED.
We got in touch with author Gerald Zavorsky for more information:
There were too many typographical errors…and when I asked the journal to published a corrected version, it was decided that a corrected version did not warrant space in the journal. So they retracted it…It was a paper on a novel way to do statistical analyses on measurement error. Too many stupid mistakes.
Zavorsky is not sure if he’ll republish his updated manuscript anywhere.
Hat tip: Rolf Degen
“Due to errors, the statistical analyses of the manuscript titled “Comparing Measurement Error Between Two Different Methods of Measurement of Various Magnitudes” by Gerald S. Zavorsky (volume 14, issue 4, pages 265–274) are incorrect.”
Are the errors typographical errors that should have been noticed and corrected in proof, or errors in the statistics or math that should have been noticed during peer review?
The explanation provided by G Zavorsky makes the journal’s decision to retract look ill-advised if the only things that needed correction were typos but the scientific and technical content was sound.
Couldn’t the journal have published a brief correction if they didn’t have enough space to republish the whole corrected article?
Just trying to figure it out here based on incomplete information.
I could not get to the paper due to a paywall, but the author was so kind to share it on ResearchGate…
I had a quick look, although it isn’t my field. I did not notice “typographical errors”, as in typesetting, or other minor spelling mistakes. I think it would be weird to retract for that reason anyway, as Karen says. A correction would be better. Plus, isn’t a journal contractually required to fulfill its end of the bargain once there is an agreement to publish (especially if there are fees, which in this case I didn’t find info about)? Surely, a journal cannot decide to retract single-sidedly just because they change their mind on whether it deserves space? So I suspect that “typographical errors” is a euphemism for “methodological errors”.
I can’t judge the content definitively, but I noticed that the principal thing that seems to be new in the paper (the use of a log-transform to compare squared-differences, according to the author’s conclusion) is simply postulated to be valid; nowhere is there any substantiation, in the form of theoretical arguments or derivations. Another weird practice was that a value that equaled zero was replaced by an arbitrary positive value in order to allow the logarithm to be taken. Not sure, but perhaps that is where the mistakes lie. In either case, those would be methodological rather than typographical errors. Anyone?
Karen, good questions. The notice does say “the statistical analyses…are incorrect”. So, that pretty much indicates that the results seem to be invalid. The numerous typographical errors seems to be related to the actual poor editorial work by the publisher, but this aspect is not clear at all. As for so many retractions, it is unclear who detected the error, why it was detected only 4 years later, and not during peer review, and if such statistical flaws exist in any more of Zavorsky’s other papers [1]. Other minor queries I have: a) who is the publisher? b) how many times has this paper been cited and will the down-stream papers that reference this paper now issue errata?
[1] http://louisville.edu/education/faculty/zavorsky/zavorsky-cv13.pdf
a) Follow the link to the notice
b1) Enter the title of the paper in Google Scholar
b1 also answers your b2