Statistical

FISHER Formula

FISHER returns the Fisher transformation of a value x, computed as 0.5 * ln((1+x)/(1-x)). It converts a correlation coefficient (bounded between -1 and 1) into a value on the real number line, which is approximately normally distributed. This is essential for performing hypothesis tests on correlation coefficients and constructing confidence intervals.

Syntax

FISHER(x)
ParameterDescription
x Parameter of the FISHER function.
Try FISHER in Viztab — free, no signup

Examples

Transform a correlation

Formula
=FISHER(0.75)
Returns 0.973 — the Fisher z-transformation of a correlation coefficient of 0.75.

Confidence interval for correlation

Formula
=FISHER(CORREL(A2:A50, B2:B50))
Transforms the sample correlation to Fisher z-space, where you can add/subtract 1/SQRT(n-3) to build a confidence interval.

Comparing two correlations

Formula
=(FISHER(0.85) - FISHER(0.70)) / SQRT(1/47 + 1/47)
Tests whether two correlations (0.85 and 0.70 from samples of 50) are significantly different using the Fisher z-test.

Common Errors

#NUM!

Returned when x is less than or equal to -1, or greater than or equal to 1. The input must be strictly between -1 and 1.

#VALUE!

Non-numeric input.

Tips

Input must be between -1 and 1 exclusive

FISHER is undefined at exactly -1 and 1. Valid inputs are strictly -1 < x < 1, which matches the range of correlation coefficients (excluding perfect correlation).

Inverse is FISHERINV

After performing calculations in Fisher z-space, use FISHERINV to convert back to the correlation scale.

Why transform?

Correlation coefficients are not normally distributed (especially near -1 or 1). The Fisher transformation makes them approximately normal so standard statistical tests apply.

Try FISHER in Viztab

Import your data and use FISHER with 370+ other formulas. No signup required.

Open Viztab

Related Formulas