June 13, 2020
Issues with Null Hypothesis Significance Testing
According to Andrew Gelman:
- Overestimation of effect sizes. The “statistical significance filter,” by which estimates are much more likely to be reported if they are two standard errors from zero, introduces a bias which can be huge.
- Estimates in the wrong direction.
- Extraction of statistically significant patterns from noise.
- Incentives for small samples and noisy measures.
Updated Jul, 02 2020