Skip to calculator
Make It Exact

Math

Z-Score Calculator

Enter a value with the mean and standard deviation of its distribution.

The link carries your figures, so it reopens on exactly these numbers.

The z values worth knowing

On a standard normal distribution — mean 0, standard deviation 1. The last two columns are the one-sided and two-sided readings of the same z, and they differ by a factor of two.

Z-scorePercentileShare above this valueShare at least this far from the mean
1Two thirds of values sit within ±1.84.1345%15.8655%31.7311%
1.645The 5% one-sided threshold.95.0015%4.9985%9.997%
1.96The 5% two-sided threshold — the one behind “p < 0.05”.97.5002%2.4998%4.9996%
297.725%2.275%4.55%
2.576The 1% two-sided threshold.99.5002%0.4998%0.9995%
3About 1 in 370 in either direction.99.865%0.135%0.27%

The formula

z = (value − mean) ÷ standard deviation
percentile = Φ(z)

Worked example

An IQ score of 130, where the mean is 100 and the standard deviation 15

  • (130 − 100) ÷ 15 = 2
  • Φ(2) = 97.72nd percentile
  • 2.28% score higher

Where this goes wrong

Reading a one-sided percentile as a two-sided result

A z of 2 puts you at the 97.7th percentile, so 2.3% score higher — but 4.6% are at least that far from the mean in either direction. Significance tests almost always want the two-sided figure, and using the one-sided one makes a result look twice as unusual as it is.

Questions

What does a z-score actually measure?
How many standard deviations a value sits from the mean. It strips the units away, so a z of 2 means the same thing whether you started with millimetres, dollars or exam marks.
Does this work for any data?
The z-score itself does. The percentile does not — it assumes the data follow a normal distribution. For skewed data, such as incomes, the z-score is still computable but the percentile it implies will be wrong.
What counts as an unusual z-score?
Around 68% of a normal distribution lies within one standard deviation, 95% within two and 99.7% within three. Beyond three is genuinely rare — about one observation in 370.

Related calculators