NORMINV Function (OpenOffice Calc)

• Introduced: First release not established. This name is documented in the OpenOffice.org 3.3 function appendix; that establishes historical presence, not introduction.
• Supported in: Apache OpenOffice 4.1.16. No final supported release has been established.
• Scope: Desktop Calc. These examples use English function names, semicolon separators and decimal points.
• Checked: October 14, 2026. Official source/documentation and independent mathematical examples checked; Apache OpenOffice Calc was not run for this guide.

NORMINV reverses the cumulative form of NORMDIST. With probability p, mean μ and positive SD σ, the result is μ+σΦ⁻¹(p). It has the same units as μ and σ. A larger p produces a higher cutoff for fixed model parameters. Official reference.

An inverse CDF is different from a reciprocal. Dividing 1 by a normal probability does not find its percentile. Nor does NORMINV fit a distribution to data: the mean, SD and intended tail must already be specified.

=NORMINV(Probability; Mean; StDev)

All three arguments are required. Use 0<Probability<1 and StDev>0 for a finite answer. A probability of 0.9 or a cell holding 90% means 90%; entering 90 means a value outside the probability domain. Mean may be negative, zero or positive.

The release evaluator treats probabilities outside [0,1] or a nonpositive SD as invalid arguments. The endpoint values 0 and 1 follow a separate no-value error path because the normal model has no finite cutoff there. This is source-based behavior; no native error display is asserted here.

Enter the formulas in separate cells. These independent mathematical values are rounded to twelve decimal places; retain unrounded stored results in subsequent calculations.

Put probability 0.9 in A2, mean 10 in B2 and SD 2 in C2. In D2 enter =NORMINV(A2;B2;C2). The expected cutoff is 12.563103131089. Format A2 as a percentage if useful, but format D2 as a measurement. Showing D2 as a percentage would obscure its meaning.

In E2 enter =NORMDIST(D2;B2;C2;1). It should return approximately 0.9. If it differs, check whether D2 was manually rounded, whether the same parameters were used, and whether NORMDIST’s last argument is cumulative rather than density mode.

A central 95% interval leaves 2.5% in each tail. For this model, use =NORMINV(0.025;10;2) and =NORMINV(0.975;10;2), giving approximately 6.080072030920 and 13.919927969080. Supplying 0.95 for the upper endpoint alone instead leaves 5% in the upper tail and does not define the same central interval.

Check the band with =NORMDIST(13.91992796908;10;2;1)-NORMDIST(6.08007203092;10;2;1), which is approximately 0.95. In an actual worksheet, reference the unrounded endpoint cells rather than these displayed decimal values.

This is an interval containing a specified portion of an assumed distribution. It is not automatically a confidence interval for an estimated population mean. The latter requires an appropriate inferential model, sample size and uncertainty treatment.

Keep the original probability cell. A value rounded to a displayed 100% might still be less than 1 internally, while replacing it with the literal number 1 changes it to an invalid finite-cutoff request. Quantiles become increasingly sensitive near the endpoints.

If the modeled variable cannot be negative, a normal model may still produce negative lower percentiles. The calculation is not necessarily wrong; the chosen distribution may be unsuitable. Review that assumption rather than clipping the result silently. Cross-check extreme probabilities and imported input types in the actual installed release when those cases matter.

• NORMINV function documentation
• Apache OpenOffice 4.1.16 argument handling and calculation