STANDARDIZE in Excel: Z-Scores, Reference Parameters and Positive Scale Checks

• Introduced: first introduction version is not established by the official history reviewed; the current product list is not a launch record.
• Availability: official help lists Excel for Microsoft 365, Excel for Microsoft 365 for Mac, Excel 2024, Excel 2024 for Mac, Excel 2021, Excel 2021 for Mac, Excel 2019, Excel 2016.
• Checked: October 14, 2026; official STANDARDIZE reference. Documentation and independently modeled examples; no native Excel execution or broader platform certification.

=STANDARDIZE(x, mean, standard_dev)

All three arguments are required. Microsoft documents #NUM! when standard_dev is zero or negative. The mean and standard deviation are parameters you supply; STANDARDIZE does not infer them from a dataset or choose a sampling model.

The cited equation is part of Microsoft’s dedicated reference and was visually checked. The independent examples use ordinary finite numbers and keep the reference population explicit.

=STANDARDIZE(70,50,10)
=STANDARDIZE(50,50,10)
=STANDARDIZE(35,50,10)

Expected values are 2, 0 and -1.5. The differences from the reference mean are 20,0,-15, and dividing each by 10 gives the standardized values.

A negative result means the value is below the supplied mean. It does not mean the original value is negative: the original 35 is positive, yet is below the reference mean of 50.

Put observed values 35, 50, 60 and 70 in A2:A5, reference mean 50 in H2 and reference standard deviation 10 in H3. Fill down:

=STANDARDIZE(A2,$H$2,$H$3)

Expected outputs are -1.5, 0, 1 and 2. The absolute references keep the same reference distribution for every row.

Those four observed values happen to have mean 53.75, not 50. That is not a formula error: the example deliberately uses an external reference mean rather than fitting the parameters to this small batch.

For the numeric dataset 2, 4 and 6, the mean is 4. Its sample standard deviation is 2, while its population standard deviation is SQRT(8/3), approximately 1.632993.

=STANDARDIZE(6,4,STDEV.S(2,4,6))
=STANDARDIZE(6,4,STDEV.P(2,4,6))

Expected results are 1 and approximately 1.224745. Both apply the stated formula correctly; they use different scale parameters.

Choose the reference measure according to the analysis, not according to which standardized value looks more convenient. STDEV.S and STDEV.P document their own denominator and input rules.

Using sample mean 4 and sample standard deviation 2, the values 2,4,6 become -1, 0 and 1. Their mean is 0 and their sample standard deviation is 1.

Using the corresponding population scale instead gives approximately -1.224745, 0 and 1.224745. Their population standard deviation is 1. The denominator convention must stay consistent between the parameter calculation and the check.

These transformations do not make a distribution normal. They change location and scale while leaving the shape of the data otherwise unchanged under a positive linear scale.

For already numeric standardized values, a policy that flags absolute values at least 2 can use:

=ABS(STANDARDIZE(A2,$H$2,$H$3))>=2

With reference mean 50 and scale 10, inputs 30, 35, 50, 70 and 71 give expected flags TRUE, FALSE, FALSE, TRUE and TRUE. The endpoints 30 and 70 are included because the comparison uses >=.

This is a chosen screening rule. A flagged row is not proof of an erroneous observation, and an unflagged row is not proof of business validity. Investigate the actual input and context before discarding or replacing it.

If z is a standardized value under known parameters, the reverse arithmetic is:

=50+10*2
=50+10*(-1.5)

Expected originals are 70 and 35. This round-trip control checks that the same mean and positive scale were used; a different parameter set would reconstruct different originals.

=STANDARDIZE(70,50,0)
=STANDARDIZE(70,50,-10)

Both have documented #NUM! expectations. A constant reference dataset has zero standard deviation, so it cannot supply a usable denominator. Replacing that scale with 1 would change the analysis rather than repair the original definition.

Text coercion, blank arguments, nonfinite values and every overflow or precision boundary remain unverified. A standardized score is not automatically a normal-distribution probability, confidence interval or percentile; those interpretations need additional assumptions and separately appropriate calculations.

• Microsoft STANDARDIZE reference.
• Editorial testing standards.

Expected results are independent statistical/arithmetic models or explicitly documented rules, not captured native output. Examples use English function names and comma separators. Datasets and reporting policies are fictional.