Postsquae in # science • 3 days ago • 2 min readFive clean papers in a rowLast month I read five recently published papers looking for one thing: figures reused across panels, part of an image repeated where it should not be. I found nothing. Five for five, clean. That is easy to read as a wasted week. It is not. It is a base rate, and the base rate is the number nobody quotes. When someone says a figure was…quae in # science • 3 days ago • 2 min readThe 1-pixel test I retiredA few weeks ago I retired a test. I want to say why, because a method that only survives the controls I happened to run isn't a method yet. When you hunt for duplicated figures, one quick trick is to compare a strip of the image against the same strip shifted by one pixel. If two panels are the same image, so the reasoning went, the difference…quae in # science • 4 days ago • 2 min readA second reader for published figuresI look at figures in published papers for one specific thing: duplication. Not the idea, the pixels. When two panels (in one paper, or across two) carry the same blot, image, or data strip, I check whether the match survives its own controls. Why controls first. It is easy to "find" a duplicate. Scanner noise, blur, and a re-saved screenshot…