Frequency of discrepancies in retracted clinical trial reports versus unretracted reports: blinded case-control studyBMJ 2015; 351 doi: https://doi.org/10.1136/bmj.h5134 (Published 25 September 2015) Cite this as: BMJ 2015;351:h5134
This Research paper by Cole and colleagues (BMJ 2015;351:h4708, doi:10.1136/bmj.h4708) contains a few errors. In figure 1 (Discrepancy counts), the bottom end of the vertical axis appeared empty in the authors’ submission because it contained four pairs of zero length bars, and so was cropped away by mistake. In addition, the order of the bars was reversed but the legend was not updated. Therefore, the final two sentences in the legend should read: “Pairs are ordered by total number of discrepancies in the pair, with those with the most discrepancies at the top. The four pairs of zero length bars are not shown”.
The last sentence in the footnote of table 2 should read: “Only retraction status was significantly associated in both the excess zero components (negatively, in that retracted trial reports are less likely to have zero discrepancies) and binomial components (positively, in that retracted trial reports contain more discrepancies).”
The authors apologise for any confusion and remind readers that the raw data and web appendices containing the data have been uploaded to thebmj.com.
Cite this as: BMJ 2015;351:h5134