The probability of having a {{false negative}, i.e. how likely it is that if there is some real effect it shows up in the statistical analysis of the data you have measured. If an experiment has high power the statistics have fewer false negatives, and real effects are more likely to be seen. The power is a combination of the experiment we have performed and choosing the correct statistical analysis technique for the data. As an example of the latter nonparametric statistics usually have lower power than parametric statistics as they make fewer assumptions, so if you know that your data is Binomial distribution or approximately Normal it is better to use tests based on these distributions. The noise–effect–number triangle offers various tactics to increase the power of a study.

Defined on page 105

Used on pages 15, 51, 87, 93, 105, 106, 108

Also known as power (statistical)

### Links:

- Wikipedia: Power of a test
- GraphPad: Key concepts: Statistical Power
- emj.bmj.com: An introduction to power and sample size estimation
- statisticsteacher.org: What Is Power?