Power, represented by β

Power, represented by β

Power

Power: The probability of correctly identifying a difference between the two groups in the study sample when one genuinely exists.

Lowest acceptable is usually 80% (meaning 8/10 chance of detecting the difference you are trying to detect, and 2/10 chance of missing a true difference.)

In general, power is increased by having larger sample size. To calculate sample size required for your study, you need:

  1. an estimate of what entails a true difference between your groups (from prior studies)
  2. standard deviation of outcomes (from prior studies)
  3. power level you want
  4. p-value you want