Fisher's z'
The
sampling distribution
of
Pearson's r
is not
normally distributed
(click here for an illustration).
Fisher developed a transformation now called "Fisher's z' transformation" that converts Pearson's r's to the normally distributed variable z'. The formula for the transformation is:
z' = .5[ln(1+r) - ln(1-r)]
where ln is the natural logarithm. It is not important to understand how Fisher came up with this formula. What is important are two attributes of the distribution of the z' statistic: (1) It is normal and (2) it has a known
standard error
of:
Fisher's z' is used for computing
confidence intervals on Pearson's correlation
and for
confidence intervals on the difference between correlations.
You can use the
r to z' table.
to convert from r to z' and back.