Statistical Power

Statistical Power

Subject: Survey Research

• Entry
• Entries A-Z
• Subject Index

• The probability of correctly rejecting a null hypothesis that is false is called the statistical power (or simply, power) of the test. A related quantity is the Type II error rate (β) of the test, defined as the probability of not rejecting a false null hypothesis. Because power is based on the assumption that the null hypothesis is actually false, the computations of statistical power are conditional probabilities based on specific alternative values of the parameter(s) being tested. As a probability, power will range from 0 to 1 with larger values being more desirable; numerically, power is equal to 1 − β.

The statistical power is also related implicitly to the Type I error rate (∞), or significance level, of a hypothesis test. If ∞ is ...

• [0-9]
• A
• B
• C
• D
• E
• F
• G
• H
• I
• J
• K
• L
• M
• N
• O
• P
• Q
• R
• S
• T
• U
• V
• W
• X
• Y
• Z