Account

What new tool do you want?

Math & Education

Z-Score & Percentile Calculator

Convert a value to a z-score and percentile from the mean and standard deviation — or a percentile back to a value — using the normal distribution.

How to use

  1. 1Enter the mean and standard deviation.
  2. 2Enter a value to get its z-score and percentile, or switch to find the value at a percentile.

How it's calculated

z = (x − μ) ÷ σ. Percentile = Φ(z), the standard normal CDF. Value at percentile p = μ + σ × Φ⁻¹(p).

Frequently asked questions

What does a z-score of 2 mean?

The value is 2 standard deviations above the mean — higher than about 97.7% of values in a normal distribution.

What's the 68-95-99.7 rule?

In a normal distribution about 68% of values fall within 1 SD of the mean, 95% within 2 and 99.7% within 3.

Does my data have to be normal?

Z-scores can always be calculated, but percentiles from Φ are only accurate if the data is roughly normally distributed.

A z-score is just a distance measured in standard deviations

Subtracting the mean and dividing by the standard deviation turns any value into a position on a common scale — a value with z = 2 is two standard deviations above the mean, regardless of whether you started with test scores out of 160 or heights in centimeters. The default example (mean 100, SD 15, value 130) gives z = 2 and a percentile of 97.72%, meaning that score beats roughly 97.7 out of 100 people in a normally distributed population.

Going the other direction takes an inverse function

Finding the value at a given percentile isn't the same calculation run backward in your head — it uses the inverse normal CDF (Φ⁻¹) to find the z-score for that percentile first, then converts back to raw units with x = μ + σz. Asking for the 90th percentile with mean 100 and SD 15 gives z ≈ 1.2816 and x ≈ 119.22, which is why the 90th percentile isn't simply '90% of the way between the extremes' — it's defined by the shape of the normal curve, which is denser near the mean.

The percentile is only as good as the normality assumption

The z-score itself is always valid — it's pure arithmetic. But converting that z-score into a percentile assumes the underlying data follows a normal (bell-curve) distribution, via the 68-95-99.7 rule: about 68% of values fall within 1 SD of the mean, 95% within 2, and 99.7% within 3. If your actual data is skewed — income or reaction times, for example — the percentile this tool reports can be noticeably off from the real rank in your data set.

Zero and one hundred percent are refused, and a zero spread is too

Finding a value from a percentile will not accept 0 or 100. The field has to be strictly between them, because those two ends have no finite z to invert. Finding a z from a value is the mode that opens, with a mean of 100, a spread of 15, and a value of 130, and that path does show the share of the curve below the value and the share above it. A standard deviation of zero or less is rejected on either path, since every distance would be undefined. Switching modes does not clear the numbers I already typed, but only the active path is used, so the percentile I left at 90 does nothing while I am still asking for a z from a value.

Related tools

More: Math & Education