NORMSINV Function (Gnumeric)

• Introduced: Documented by 1.12.62; original introduction not established.
• Availability: Supported in Gnumeric 1.12.62, based on the function-specific registration.
• Scope: Desktop; the fn-stat function module must be included and enabled.
• Checked: October 14, 2026. Official source and independent mathematical examples checked; Gnumeric 1.12.62 was not run.

NORMSINV returns a standard-normal z-score from a cumulative probability. The result expresses distance from the mean in standard-deviation units. It is the inverse CDF, not the reciprocal of NORMSDIST. Official function reference.

z=Φ⁻¹(p)=√2×erf⁻¹(2p−1). The median probability 0.5 gives zero; probabilities below 0.5 give negative z-scores, and probabilities above 0.5 give positive ones.

=NORMSINV(p)

Every argument shown is required. Use numeric inputs or cell references. The examples follow English Gnumeric’s comma-separated notation and decimal points; localized separators can differ. Release-pinned argument help and evaluator.

Enter each formula in an unused worksheet cell. These are independently calculated mathematical reference values rounded to twelve decimal places. Keep full stored precision in later formulas rather than retyping a rounded display.

Select the intended tail

For a one-sided upper 5% cutoff, use =NORMSINV(0.95), approximately 1.644853626951. For a central 95% interval, split the remaining 5% between the two tails: =NORMSINV(0.975) gives approximately 1.959963984540. Supplying 0.05 instead asks for a negative lower-tail cutoff.

Convert a z-score into measurement units

For mean 10 and SD 2, =10+2*NORMSINV(0.9) has expected value 12.563103131089. The equivalent general quantile is =NORMINV(0.9,10,2). Keep the probability separate from the modeled measurement units.

The wrapper rejects p outside [0,1]. Its qnorm helper produces infinities at the two endpoints, which the shared result constructor maps to #NUM!. Thus use 0<p<1 when you need a finite z-score; do not interpret 0 or 1 as ordinary finite cutoffs. This is release-source reasoning, not an observed cell error. Pinned evaluator. Quantile endpoints and boundary macro. Normal inverse helper. Nonfinite numeric result handling.

Put 0.975 in A2 and =NORMSINV(A2) in B2; =NORMSDIST(B2) should recover approximately 0.975. Sensitivity grows in the tails, so use the original cell rather than a displayed probability rounded to too few places. Such a round trip checks numerical consistency, not the appropriateness of a statistical test.

NORMSDIST; NORMINV.

• NORMSINV manual entry
• Gnumeric 1.12.62 argument contract and calculation
• Quantile endpoints and boundary macro
• Normal inverse helper
• Nonfinite numeric result handling