Seeing #medtwitter #academictwitter talk on errors after publication. Following in this researcher's example, I'd like to recall an error that changed my results from (+) to (-). Never published, but a discussion about the experience is still valuable to the academic process. 1/ https://twitter.com/bnallamo/status/1022541335646142464

I was an MS2 and it was my 1st exp with medical research. I didn't realize STATA coded missing values as "99". My code was heavily based on Boolean statements (e.g if > then TRUE). As a result, missing values were misclassified as TRUE. 2/

As I responded to reviewer comments, I discovered the error, which changed the conclusion. I was so embarassed, I didn't publish the paper and stored it away as a learning experience for myself. Now I think it is important to share as a learning experience for others. 3/

So kudos to @bnallamo @Jama_current and the authors of this paper for your scientific integrity. Remember that trainees like myself are looking to senior academics to model scientifically ethical behavior. 4/ https://jamanetwork.com/journals/jama/article-abstract/2764714

I'll conclude with my comment from another relevant thread: https://twitter.com/CrystalZhengMD/status/1252815771291381774?s=20