Through a “combination of masochism and stubbornness,” Edward Raff, chief scientist at Booz Allen, spent eight years trying to reproduce the results from 255 papers (with only 162 successful reproductions). This paper distils the 26 key features of reproducibility.
It's really important, that if you want others to believe your results, that they can analyse and reproduce them. It's a founding principle of all medical research and analysis that is presented to peers for review and that they are able to reproduce it. Yes, this paper is aimed at machine language, but I believe you can apply some of these for anything. Read it, bookmark it, and keep a copy on your noticeboard.
No comments:
Post a Comment
Note: only a member of this blog may post a comment.