Last updated: 2026-05-19

What This Calculator Does

This calculator computes the Z-score for any data point and converts it to an approximate percentile rank, assuming a normal distribution. Enter the observed value, the mean, and the standard deviation — no Z-table needed.

Definition

A Z-score (also called a standard score) measures how many standard deviations a value lies above or below the mean of a distribution:

• x — the observed value
• μ — the population (or sample) mean
• σ — the standard deviation

A Z-score of 0 means x is exactly at the mean. Z = +1 means one standard deviation above. Z = −2 means two standard deviations below. The sign tells you direction; the magnitude tells you how far.

Standardization across distributions

Z-scores place values from different distributions on a single common scale, which makes comparisons meaningful:

• A student who scored 78 on one exam (mean 70, σ 8) and 55 on another (mean 45, σ 8) performed equally well in absolute standard-deviation terms — both have z = 1.0.
• A height of 180 cm is far above average for women (μ ≈ 163 cm, σ ≈ 7 cm, z ≈ 2.4) but only just above average for men (μ ≈ 175 cm, σ ≈ 7 cm, z ≈ 0.7).

Without standardization, values from distributions with different means and spreads cannot be compared directly; the Z-score removes that obstacle by expressing every value in the same units.

The 68-95-99.7 rule

For any normal distribution:

A Z-score of |z| ≥ 2 therefore occurs less than 5% of the time in a normal population, and values with |z| ≥ 3 are rare — less than 0.3%.

From Z-score to percentile

If the underlying distribution is normal, the percentile equals the cumulative distribution function (CDF) of the standard normal, multiplied by 100:

This calculator evaluates Φ(z) using the standard normal CDF, implemented via the error function erf — a special function defined as $\text{erf}(t) = \frac{2}{\sqrt{\pi}} \int_0^t e^{-s^2}\,ds$. Key reference points:

Population vs. sample standard deviation

The formula is the same either way, but the choice of standard deviation matters:

• Population SD (σ) — used when data cover every member of the group (e.g., the complete test scores for a class).
• Sample SD (s) — used when the data are a subset drawn from a larger population (e.g., a random sample of heights).

Sample SD is slightly larger than population SD because it uses n − 1 in the denominator, known as Bessel's correction. Choosing the wrong one shifts both the Z-score and the percentile. By convention, the population SD applies to Z-score standardization problems in textbooks, while the sample SD is more appropriate in inferential statistics.

Worked examples

Example 1 — Exam score

A student scores 85 on an exam where the class mean is 70 and the standard deviation is 10.

Φ(1.5) ≈ 0.933, so the student is at approximately the 93rd percentile — better than 93% of the class.

Example 2 — Height comparison

A woman is 175 cm tall. Reference values for adult women: μ = 163 cm, σ = 7 cm.

Φ(1.71) ≈ 0.956, placing her at roughly the 96th percentile for women's height.

Example 3 — Below-average value

A product batch has an average weight of 500 g with σ = 20 g. A package weighs 455 g.

Φ(−2.25) ≈ 0.012, so only about 1.2% of packages from this batch weigh as little or less — likely a quality-control concern.

Limits of the percentile conversion

The percentile conversion assumes the data follow a normal distribution. When the data are:

• Skewed (e.g., incomes, reaction times) — the normal-CDF percentile overstates or understates the true rank.
• Heavy-tailed (e.g., financial returns) — extreme Z-scores correspond to far higher percentiles than the normal model predicts.
• Discrete (e.g., counts, dice rolls) — treat the percentile as a rough approximation only.

In these cases, the Z-score is still a useful comparison metric; it is only the percentile label that becomes misleading.

Relationship to the normal-distribution calculator

The Normal Distribution Calculator covers the same Z-score formula but adds the probability density f(x) and the complementary tail P(X > x). That calculator suits a full distributional picture, while this one standardizes a single value and looks up its percentile.