NORM.S.INV in WPS Spreadsheet: Convert Probability to a Z-Score

=NORM.S.INV(0.975) has independently expected output 1.959963985. It converts a lower-tail probability into a standard-normal z-score. The model has mean 0 and standard deviation 1.

NORM.S.INV(probability)

probability is required and must be strictly between 0 and 1 according to the formal lesson. It is the probability to the left of the required z-score. This one-argument function does not take a mean or standard deviation.

Enter 0.95 in A2 and =NORM.S.INV(A2) in B2. Expected B2: 1.644853627. Under the standard normal model, 95% lies below this cutoff and 5% above it.

For reference, enter these additional probabilities down A3:A5 and copy the formula down:

The sign identifies which side of the mean the cutoff falls on. A negative result is valid and is not a negative probability.

A central 95% band leaves 2.5% in each tail. Calculate:

=NORM.S.INV(0.025)
=NORM.S.INV(0.975)

Expected results: -1.959963985 and 1.959963985. For a model with mean 50 and standard deviation 8, transform the upper cutoff with =50+8*NORM.S.INV(0.975). Expected result: 65.67971188. The lower transformed cutoff is 34.32028812.

These are distribution cutoffs, not automatically a confidence interval for a sample mean. The appropriate sampling model and standard error would need to be established separately.

The standard-normal cumulative integral evaluated at 1.95996398454 gives 0.975 to the shown precision. Symmetry gives opposite signs for probabilities p and 1-p. The examples were checked with an independent inverse-normal implementation; they were not run in WPS.

The official NORM.S.INV lesson excludes probabilities at or outside 0 and 1 and lists #NUM!; it lists #VALUE! for nonnumeric probability. Its error bullets misspell the name as NORMS.INV, while the title, syntax and workflow consistently use NORM.S.INV. Enter the formal syntax shown above.

The separate historical NORMSINV lesson uses different endpoint wording. Do not transfer that wording here, collapse the identities, or claim the two names have identical edge behavior. Use strict interior probabilities for these examples.

If the formula is rejected, check the two periods in the function name and the separator conventions in your installed editor. The source describes an iterative search but supplies no current numerical tolerance or extreme-tail accuracy guarantee. Rounding a very small probability to 0 before evaluation destroys the input.

The official WPS NORM.S.INV lesson was fetched afresh on October 19, 2026. Its edition banner says WPS Office for Windows Personal. The lesson dates from 2019; it does not establish introduction or retirement dates, exact build coverage, or current Mac, Linux, web and mobile behavior. Examples are independently calculated expectations, not observed WPS outputs. Native execution and genuine current-application screenshots remain unverified.

NORMINV, NORMDIST, STANDARDIZE. Browse the WPS function library.