Statistics and science
- Twenty tips for interpreting scientific claims (Nature Publ Co)
- Broad cautions and advice on statistical validation for scientific inference.
- Weak statistical standards implicated in scientific irreproducibility
- One-quarter of studies that meet commonly used statistical cutoff may be false. From Nature, Nov 11 2013. Note that this refers to social and biological science studies that commonly use p=0.05 as a criterion for significance. In contrast, most astronomers use p=0.003 (3-sigma), and should be less vulnerable to these problems.