If science is to be both honest and healthy, we must accept that statistically non-significant results are part of reality. The SAMPL guidelines, if adopted widely by scholarly publishers and journal editors, hold a solution for authors who worry their results are not “significant.”
A new paper finds unexpected disturbances around p-value ranges approaching 0.05. Is there something going on beyond mere science?
I knew there was something they weren’t telling me!
Like ice-cream and murder, there is no causal relationship between reference length and citations. Now go tell Nature.
A recent article about statistics started a useful discussion in the blogosphere. And I was left wondering: Are open data dreams built on statistical sand?