Skip to content

Fisher

FISHER Function

The FISHER function in Excel calculates the Fisher transformation of a given number. This mathematical transformation is often used in statistics, particularly when working with correlation coefficients, to normalize data or make it more suitable for hypothesis testing and confidence interval estimation.

Key Features of FISHER:

  • Computes the Fisher transformation for a given input value.
  • Useful when analyzing correlation coefficients, as it stabilizes the variance and converts the correlation to a normal distribution for statistical analysis.
  • Commonly applied in hypothesis testing and regression analysis involving correlations.

Syntax:

FISHER(x)
  • x: The value to be transformed. This must be between -1 and 1 (exclusive) since it is often derived from a correlation coefficient.

Formula:

The FISHER function implements the following transformation:

FISHER(x) = 0.5 * LN((1 + x) / (1 - x))

Where LN is the natural logarithm.

Examples:

  1. Compute the Fisher transformation for a specific value:

    =FISHER(0.5)

This calculates the Fisher transformation for the correlation coefficient 0.5.

  1. Analyze a series of correlations:

Suppose you have multiple correlation coefficients in a dataset and want to normalize them for further statistical testing.

If a correlation coefficient is stored in cell A1:

```excel
=FISHER(A1)
```
  1. Using Fisher-transformed values in hypothesis testing:

Combine the FISHER function with other statistical functions to calculate confidence intervals or test hypotheses about correlation coefficients.

Notes:

  • If x is less than or equal to -1 or greater than or equal to 1, FISHER returns the #NUM! error, as the transformation is not defined for such values.
  • This function is often paired with the FISHERINV function to revert the Fisher-transformed value back into the original scale.

Applications:

  • Statistical Analysis of Correlations: Improve the properties of correlation coefficients for significance testing.
  • Hypothesis Testing: Test hypotheses about population correlations using normalized Fisher-transformed values.
  • Confidence Intervals for Correlation Studies: Calculate confidence intervals using Fisher-transformed values to ensure normality.

Tip: Use FISHER in combination with FISHERINV to easily switch between the transformed and original scale of correlation coefficients for statistical analysis.