News for hackers. Stuff that matters.
Re:The Statistical Crisis in Science by cja23 on Friday January 16, 2015 @09:34AM (#8899097)
Yes! I think any good statistician would agree with all of that and emphasize the importance of how you present your conclusions. Forming "data-dependent hypotheses" is part of what most statistician's call "exploratory data analysis" (EDA). When presenting EDA findings, we should use terms like "association", "relationship", "correlation", "possible", and not words like "cause", "effect", "p-value", "test", etc.

There are no comments on this comment yet.