Fisher transformation
|
|
In statistics, hypotheses about the value of r, the correlation coefficient between variables x and y of the underlying population, can be tested using the Fisher transformation applied to r.
Definition of the Fisher transformation
Let N be the sample size. The transformation is defined by:
- z = ½ log((1+r)/(1-r)
Here z is approximately normally distributed with mean r, and standard error
- 1/((N-3)^0.5).
This is a common way of testing whether a correlation coefficient is significantly different from 0, and hence ascribing a p-value.