Z-Score Standardization Calculator

StatisticsLast updated: August 22, 2026

Standardization (the z-score transformation) re-expresses every observation of a variable as the number of standard deviations it sits from the mean. After the transformation the series has a mean of exactly 0 and a standard deviation of exactly 1; the values become unitless, and variables measured on different scales become directly comparable.

This calculator standardizes not a single score but the whole data set: it lists each observation's deviation from the mean, its z-score, and the T score used in educational measurement (50 + 10z), flags observations beyond the threshold you choose, and draws before-and-after histograms. This is the preprocessing step that lets variables contribute with equal weight in regression, principal component analysis, clustering, and neural networks.

You can separate the numbers with commas, spaces, or line breaks. A decimal comma is also accepted (3,14).
Determines the denominator of the z formula. The SPSS option Descriptives > Save standardized values uses n−1.
In a normal distribution, 95% of observations lie within ±2 and 99.7% within ±3.

Standardization Formulas

z-score:             zᵢ = (xᵢ − x̄) / s
Population form:     zᵢ = (xᵢ − μ) / σ
Sum of deviations:   Σ(xᵢ − x̄) = 0  →  z̄ = 0
Transformed variance: s²_z = Σ(zᵢ − z̄)² / (n − 1) = 1
T score:             T = 50 + 10 · z
IQ scale:            IQ = 100 + 15 · z
Inverse transform:   x = x̄ + z · s

The identities z̄ = 0 and s_z = 1 hold algebraically for every data set; if the computation is correct, those two numbers cannot take any other value. The very small deviations you may see on screen come only from floating-point rounding.

How to Calculate

  1. Paste every value you want to standardize into the box; commas, spaces, and line breaks are all accepted.
  2. Choose whether the data is a sample or the entire population — this decides whether the standard deviation in the denominator of the z formula uses n−1 or n.
  3. Choose the outlier threshold: |z| > 3 is common in the social sciences, while |z| > 2 is often used in small samples.
  4. Read each observation's z-score and T score from the table; rows beyond the threshold are flagged as 'Beyond threshold'.
  5. Confirm in the results card that the transformed series has a mean of 0 and a standard deviation of 1 — that is the check that the transformation was applied correctly.
  6. Compare the before and after histograms: the shape staying the same shows that standardization does not make the distribution normal.

Worked Examples

Standardizing exam scores

The 15 scores have a mean of 73.20 and a sample standard deviation of 14.18; the score of 40 is the only observation beyond the |z| > 2 threshold, with z = -2.341 and a T score of 26.6. The transformed series has a mean of 0.0000 and a standard deviation of 1.0000, while the skewness stays at -0.589.

Sample size (n): 15 · Raw mean (x̄): 73.20 · Raw standard deviation (s): 14.18

Standardizing with the population formula

The five values have a mean of 14.00 and a population standard deviation of √(40.00 / 5) = 2.83. The two ends of the series take z = 1.414 for 18 and z = -1.414 for 10. If the same data were treated as a sample, the denominator would be n − 1 instead of n and the z-scores at the ends would shrink slightly.

Sample size (n): 5 · Raw mean (x̄): 14.00 · Raw standard deviation (σ): 2.83

Flagging an outlier

The mean is 31.07 and the standard deviation is 13.53; the value 78 is the only observation beyond the |z| > 3 threshold, with z = 3.469. Because the outlier inflates the standard deviation, the z-scores of the remaining 14 observations are squeezed between -0.744 and 0.217.

Sample size (n): 15 · Raw mean (x̄): 31.07 · Raw standard deviation (s): 13.53

Frequently Asked Questions

Does standardization turn data into a normal distribution?
No. Standardization is a linear transformation; it changes only the scale and the center, not the shape of the distribution. Skewness and kurtosis are exactly what they were before the transformation. To move closer to normality you need a nonlinear transformation such as a logarithm, a square root, or Box-Cox.
Why is the mean of the z-scores exactly 0?
The sum of deviations Σ(xᵢ − x̄) is zero for every data set; because the z-scores are those deviations divided by a constant (s), their sum is zero as well. The same algebraic necessity makes the standard deviation of the z-scores equal to 1. If you see a very small number instead of 0.0000 on screen, that is only floating-point rounding.
Should I use n−1 or n?
If your data is a sample drawn from a larger population, use n−1 (the sample standard deviation); it is the unbiased estimate of the variance and the default in SPSS, R, and the Excel function STDEV.S. If you measured the entire population, the n denominator is the correct one. In small samples the difference between the two choices is noticeable; it fades away as n grows.
Are standardization and normalization the same thing?
No, they are two different operations that are often confused. Standardization sets the mean to 0 and the standard deviation to 1, and the values have no upper bound. Min-max normalization squeezes the data into a fixed range such as 0–1 and takes its scale only from the smallest and largest observation. Standardization is preferred when outliers are present, normalization when the application requires a bounded range.
Is the z-score a reliable way to identify outliers?
Only up to a point. Because the outlier itself inflates the mean and the standard deviation, it shrinks its own z-score — this is called masking, and in a small sample such as n = 10 the absolute z-score cannot mathematically exceed 2.85. Tukey's interquartile range rule, or a robust z-score based on the median absolute deviation (MAD), gives a more reliable answer.
How do I do this in SPSS and Excel?
In SPSS it is enough to tick the 'Save standardized values as variables' box under Analyze > Descriptive Statistics > Descriptives; the program creates a new column named Zvariable and uses the n−1 denominator. In Excel, =STANDARDIZE(A2,AVERAGE($A$2:$A$100),STDEV.S($A$2:$A$100)) gives the same result. In R, the function scale(x) is used.
What does standardizing the variables buy you in regression?
Standardized coefficients (betas) make it possible to compare the relative importance of predictors measured in different units; you can weigh age against income on the same scale. They also reduce multicollinearity in models with interaction and quadratic terms. The trade-off is that the direct unit interpretation of the coefficients (how much one extra year is worth) is lost.