PlanMaker FISHER: Transformation, Input Limits and Logarithmic Checks

• Introduced: Documented by the PlanMaker 2010 manual; the first introduction release and date are not established.
• Availability: Documented in PlanMaker 2026; this is not a final working-version endpoint.
• Scope: The reference belongs to the 2026/NX manual family. FreeOffice and individual mobile-build availability remains unverified.
• Checked: October 14, 2026, against the official function reference and historical entry. Examples were independently calculated; no native PlanMaker test was performed.

=FISHER(Number)

The PlanMaker 2026 reference requires Number to be strictly between −1 and 1. Both endpoints are excluded. There is one argument and no optional sample size, significance level or logarithm base.

The mathematical Fisher transformation is:

=0.5*LN((1+Number)/(1-Number))

This definition is independently documented in SAS’s statistical reference. Here it supplies a mathematical cross-check, not evidence about PlanMaker’s internal numerical algorithm. LN uses the natural logarithm.

Create a blank sheet named FisherValues. Put Input, FISHER and Logarithmic check in A1:C1. Enter the six numeric inputs below in A2:A7.

Enter =FISHER(A2) in B2 and =0.5*LN((1+A2)/(1-A2)) in C2. Fill B2:C2 down through row 7. Keep these outputs outside the input column and retain full stored precision.

Positive inputs give positive transformed values; negative inputs give negative values. Opposite inputs have opposite outputs in the mathematical model. The zero input stays zero, while 0.95 transforms to a value greater than 1. The transformed scale does not share the original upper bound.

Put Recovered input in D1. Enter =FISHERINV(B2) in D2 and fill to D7. The mathematical targets are the original inputs: −0.6,−0.2,0,0.2,0.6,0.95. This is a round-trip check using the inverse function, not a reciprocal calculation.

Keep A:D intact. Put Input and Transformed value in F1:G1. Enter 0.2,0.4,0.6,0.8 in F2:F5. Put =FISHER(F2) in G2 and fill to G5.

In H3 enter =G3-G2; expect approximately 0.2209163761. In H5 enter =G5-G4; expect approximately 0.4054651081.

Both original changes are 0.2, but the second transformed change is larger. This explains why an arithmetic operation on transformed values cannot generally be replaced by the same operation on raw coefficients. It does not establish that a larger difference is statistically significant: the worksheet has no sample-size or uncertainty inputs.

Use a coefficient such as 0.6, not the number 60. A stored value of 0.6 may display as 60% under percentage formatting, but a plain numeric 60 is outside the documented domain.

Do not round a near-endpoint coefficient to exactly 1 or −1 before calculation. The rounded input would violate the stated range even if the original full-precision value was valid. A correlation of exactly 1 can be a legitimate result of a separate calculation while still being outside FISHER’s documented input interval.

The current page does not provide a complete error-token, numeric-text, array or extreme-precision specification. These examples use ordinary numeric scalar values. An imported range should first be checked for stored types and accidental labels; a shared function name does not guarantee identical coercion across programs.

A Fisher-transformed value is not itself a confidence level, p-value or combined-study estimate. Statistical uses require a justified model and information beyond this one-argument function. The worksheet demonstrates the transformation and its inverse only.

• FISHERINV reverses the transformation.
• CORREL calculates a correlation from paired observations.
• LN supplies the independent logarithmic check.
• PEARSON provides another documented paired-data correlation function.