I recently found a massive error in one of my published papers. The main finding was the result of a programming bug and was, in fact, completely untrue. THREAD with the very short version below, essay with the full description here: https://medium.com/@jstrand_76194/when-science-needs-self-correcting-a130eacb4235

The bug had introduced a systematic difference between the critical conditions and was entirely responsible for the observed effect. The published finding was totally wrong. 2/5

Revealing this error to my co-authors, the journal, my dean and tenure review committee, my NIH program officer, and my research students was the hardest thing I’ve done in my career. 3/5

Part of me wanted to tell as few people as possible. But, in the interests of normalizing that mistakes happen and showing that revealing mistakes doesn’t have to sink you, I’m sharing this experience (even though there are bigger things going on in the world right now). 4/5

We need systems to avoid mistakes and transparency to air them when they happen, and we also need methods for people to reveal honest errors without the stigma of retraction. TL;DR: Admitting scientific errors is hard. It’s also important. END #AcademicChatter

If you want to hear more details, I talk about how it all went down with @jpeelle in Episode 17 of the juice and the squeeze: http://www.juiceandsqueeze.net/17