FISHERINV in Collabora Office: Return From the Fisher Scale
Use Collabora FISHERINV to reverse Fisher transforms, compare transformed and original scales, and audit round trips without premature rounding.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
Quick answer: =FISHERINV(0.5) has independently expected result approximately 0.462117157. It maps a Fisher-scale value back into the interval between −1 and 1.
Syntax and inverse definition
=FISHERINV(Number)
Number is the transformed value to reverse. It is the only required argument. The Collabora reference identifies the inverse Fisher transformation and rounds the result of its 0.5 example to 0.46.
The independent mathematical model is the hyperbolic tangent: (EXP(2*z)-1)/(EXP(2*z)+1). This expression explains the mapping, but it is not a recommended replacement for the built-in function at extreme magnitudes, because its intermediate exponential may overflow.
Input z is on the transformed scale and may be greater than one or less than negative one. Do not apply FISHER’s original-input restriction to this inverse function. For finite real z, the ideal mathematical result lies strictly between −1 and 1, though finite-precision calculations can approach or round to an endpoint.
Example 1: build an inverse table
Enter −1, −0.5, 0, 0.5, 1 in A2:A6. Put =FISHERINV(A2) in B2 and fill down. Independent expected results are approximately −0.761594156, −0.462117157, 0, 0.462117157, 0.761594156.
These are numbers on the original scale, not necessarily probabilities. Negative outputs are valid. A correlation-scale interpretation is appropriate only if the transformed inputs genuinely came from the relevant statistical context.
For the moderate numbers in this example, =(EXP(2*A2)-1)/(EXP(2*A2)+1) in C2 provides a mathematical cross-check. This illustrative expression is deliberately bounded to avoid claiming stability at extreme input values.
Example 2: distinguish input from output
=FISHERINV(0.8) targets approximately 0.664036770, not 0.8. In contrast, =FISHERINV(FISHER(0.8)) targets 0.8 because the inner function first transforms the original value.
Label the intermediate column clearly. A common modeling error is passing an original correlation directly into FISHERINV while expecting it to remain unchanged. The inverse expects a transformed value.
Advanced application: reverse transformed endpoints separately
Suppose a transformed-scale interval runs from −0.5 to 0.8. Put those bounds in A10:A11 and apply FISHERINV separately in B10:B11. The original-scale bounds are independently approximately −0.462117157 and 0.664036770.
Because the mapping is increasing, their order is preserved. Their spacing is nonlinear, so transforming only the interval’s midpoint and then adding the old half-width would not reproduce the endpoints. This is a general transformation illustration, not a claim that the chosen interval is a valid confidence interval.
Advanced application: keep intermediate precision
If an original value is 0.8, its Fisher transform is approximately 1.098612289. Store that full calculated value and reverse it using a cell reference. A display rounded to 1.10 can coexist with full stored precision; manually replacing the formula with 1.10 changes the value being reversed.
In a separate control column use =FISHER(FISHERINV(A2)) for the moderate z values above. The independent target is A2. At large magnitudes the inverse can become numerically close to ±1, so this round-trip check may lose accuracy even if the displayed output looks plausible.
Errors and troubleshooting
Check numeric input and the origin of its scale before assuming a strange output is an error. FISHERINV does not restore information lost by rounding, averaging or other operations between a forward and reverse transformation.
The dedicated help does not specify every invalid-input code, extreme-value boundary or numerical tolerance. This guide makes no native guarantee for them. Do not infer that a transformed dataset follows a normal distribution solely because this function has been applied.
Availability and verification
The exact function is documented in Collabora Office 26.04 desktop help. Its source snapshot was retrieved on 2026-10-13 and hash-verified for this guide on October 15, 2026. Desktop documentation is not proof of Online or mobile support.
First introduced release and last functioning version are unknown. No native Collabora or WPS run occurred. The numerical values are independent mathematical expectations; they do not certify application-specific numerical tolerances, error codes or every edge case. Formulas use English names and semicolon separators.