June 13th, 2008


almost significant

I can't believe I found an article in which the abstract itself discusses an "almost significant" relationship that was found between two variables. If I were a reviewer, I would tell the author to go get more data until the relationship was actually significant. It might be okay to mention an almost significant relationship in the results or discussion and indicate that you intend to do further experiments, but it shouldn't be the focal point of the paper.

/statistics nerdery