What this tool calculates
The calculator expresses how far a value lies from a supplied mean in units of the supplied standard deviation. The result is a dimensionless z-score. Positive values lie above the mean, negative values below it, and zero equals the mean.
This page performs standardization only. It does not estimate the mean or standard deviation from raw data, test normality, or convert a z-score to a probability.
Inputs
- Value x is the observation to standardize.
- Mean is the reference center from the same variable and population or sample definition.
- Standard deviation is the positive reference scale and must be greater than zero.
x and the mean must use the same unit, and the standard deviation must be expressed in that unit. Mixing a sample mean with an unrelated population standard deviation produces a number but not a coherent comparison.
Method and formula
The calculation is:
z = (x − mean) ÷ standard deviation
Subtracting the mean centers the observation. Dividing by the standard deviation removes the original measurement unit. A standard deviation of zero is rejected because division by zero is undefined and a no-spread reference distribution cannot provide this scale.
The sign comes entirely from x − mean when the scale is positive. Reversing that subtraction would reverse the interpretation and is not the convention used here.
Worked example
Let x be 85, the reference mean be 70, and the standard deviation be 10.
z = (85 − 70) ÷ 10 = 15 ÷ 10 = 1.5.
The value is therefore 1.5 supplied standard deviations above the mean. If x were 55 with the same reference values, z would be −1.5. If the standard deviation were 0, the calculator would return invalid input rather than infinity.
How to interpret the result
A z-score measures relative position on the given scale, not importance or quality. Values from different unit systems can be compared after valid standardization, but only when their reference distributions and definitions are meaningful.
Do not automatically translate 1.5 into a tail probability. That step requires a distributional model, often but not always a normal distribution. Even under normality, one-sided and two-sided questions differ.
Accuracy and limitations
Z-scores are sensitive to the supplied mean and standard deviation. Outliers, skewness, subgroup mixing, measurement changes, or using sample estimates can alter interpretation. A large absolute value can indicate rarity under a model, a data error, or simply a heavy-tailed distribution.
Floating-point formatting can round the result. The tool does not calculate robust scores, percentiles, p-values, confidence intervals, sampling corrections, or clinical reference ranges. It should not be used alone for diagnosis, admissions, employment, or other high-impact decisions.
Sources
- NIST/SEMATECH e-Handbook, Normal Distribution describes standardized normal-variable notation and its distributional context.
- NIST/SEMATECH e-Handbook, Measures of Scale explains the standard-deviation scale used in the denominator.
Editorial record
Author: SoupCalc Editorial Team
Last reviewed: August 14, 2026
Review scope: Center-and-scale formula, positive-denominator validation, dimensionless interpretation, normality cautions, NIST sources, and the 85/70/10 example were checked.