BALLS: A REBUTTAL
in episode #11, the error bar identified the worst brain imaging study ever published. comments placed on pubpeer.com have prompted the authors to rebut that claim. the error bar takes another look.
this story was in episode 19 #penalty #choke #football #FNIRS #QRP
the error bar says
back in June, i reviewed what i called the worst brain imaging study ever published. in the study, researchers put brain imaging devices on footballers' heads & tried to measure brain activation data in the football field. this, for information, is not possible.
in my review, i pointed out that 86% of the data was removed, in various different ways. i also pointed out the very large number of statistical tests that were done & the lack of overall clarity in the analytic approach. the brain responses in the paper look nothing like what we should expect.
i concluded that this work was balls: impossible, unpublishable, pseudoscience. so bad indeed that the authors, reviewers & editors should all have been sent off.
this week, i received a rebuttal by email & on pubpeer.com. the authors pointed out, first, that it was only 59 percent of data that were discarded (not 86 percent); second that, yes, there was no clear brain response in the data as a whole, but this is because all of the individual data was averaged together & the effects were smoothed out.
even under the most generous interpretation of this rebuttal, the authors discarded 59% of the data & did not find any consistent responses.