The Mystery of a Retracted Study That Came Back to Life
A new paper in the February issue of Pediatric Obesity probes an important question. Can a gardening, cooking, and nutrition program exert an effect on obesity risk for Latino youth? At first glance, the results are encouraging. Right there in the title, the authors answer the question. The LA Sprouts program “reduces obesity and metabolic risk.” But look a little closer and you’ll find a mystery. Is this a retracted study that has come back to life?
Back in May of 2015, the same four authors of the new study published a seemingly identical study in Obesity. It was an evaluation of the same program. The data in this study seem to be the same. Their conclusion was the same:
LA Sprouts was effective in reducing obesity and metabolic risk; however, additional larger and longer-term studies are warranted.
Seven months later, the authors and the journal editors agreed to retract this study from publication. The retraction notice explained that the data did not support the original conclusions:
The retraction has been agreed to because the statistical analysis was not correct given the cluster-randomized design, and the wrong degrees of freedom were used. The conclusion that the original paper drew about having demonstrated treatment efficacy was not supported in the corrected analysis.
In the new paper, you will find no explanation. No mention of a revised analysis that somehow justifies another twist in the story of this dataset. No reference to the retracted paper at all. All we have is a subtle suggestion in the new paper that the reader should be skeptical of these conclusions:
Because this was a small study of four schools, analysis of change measures between clusters of students in schools would be underpowered, and the current approach may lead to false positive results.
Retractions are an awkward, but important aspect of peer-reviewed science. Particularly in behavioral and biological sciences, fresh eyes can reveal flaws that escape the notice of researchers who are really close to a study. People pay close attention to retractions, because the truth matters in science. Honest mistakes account for most retractions. Finding and correcting honest mistakes marks a success for everyone, including the original researchers. All can take pride that peer-reviewed science is self-correcting.
So how did this retracted study come back to life in a new paper that’s very similar to the original one? That’s a mystery that remains to be solved.
Click here for the new paper, here for the retracted paper, and here for the explanation of the retraction.
Undiscovered Mysteries, photograph © Chris Chabot / flickr
Subscribe by email to follow the accumulating evidence and observations that shape our view of health, obesity, and policy.
February 13, 2017
February 17, 2017 at 8:19 am, Kb said:
Yes! I came across these papers about a year ago (the Pediatric Obesity one had just come out as ePub) and was baffled how the retracted paper was re-published in a new journal without having corrected the analysis.
February 17, 2017 at 11:45 am, Asheley said:
I think perhaps the question is not “How did it come back to life?”, but rather “How would a reviewer or journal ever know it had been previously retracted?”
Unless someone who knew of the previous paper happened to be involved in the review, it could easily slip by.
February 17, 2017 at 1:02 pm, Ted said:
Good point, Asheley. A Google Scholar search for LA Sprouts yields the retraction as the third result.
February 18, 2017 at 10:24 am, Kb said:
But as reviewers, are we responsible for checking every manuscript we review for a prior retraction? Perhaps that should fall on the Editor?
Also when you submit your manuscript and state that it is not published or under review elsewhere, it seems like ethically that statement includes if the science has been published elsewhere and retracted.
February 18, 2017 at 3:17 pm, Ted said:
I agree that if a paper is retracted and subsequently resubmitted for publication, best practice is to include a disclosure of the retraction and an explanation of the re-analysis. Also, the prior publication and retraction would seem to be a relevant reference. Otherwise, the authors create the appearance of hiding something.
You can find a thorough discussion here: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4524344/