Move between Z-scores and percentile ranks
A Z-score says how many standard deviations a value sits from the mean of a normal distribution, while a percentile says what fraction of the distribution falls below it. They are two views of the same point on the standard normal curve. This calculator converts in both directions: Z to percentile using the normal CDF, and percentile to Z using the inverse normal.
How it works
To convert a Z-score to a percentile, integrate the standard normal density up to z — the cumulative distribution function:
percentile = 100 * Phi(z)
Phi(z) = 0.5 * (1 + erf(z / sqrt(2)))erf is the error function, computed here with a high-accuracy rational approximation. To go back the other way, the tool applies the inverse normal (the probit function) using Acklam’s rational approximation to solve Phi(z) = p for z.
Worked examples
Z to percentile:
A Z-score of 1.5 gives Phi(1.5) ≈ 0.9332, so the percentile is about 93.3 — roughly 93 percent of values fall below this point. If a student scores 1.5 standard deviations above the class mean, they outperformed about 93% of classmates.
Percentile to Z: The 90th percentile corresponds to a Z-score of about 1.2816. If a standardised test reports a score at the 90th percentile, the raw score is approximately 1.28 standard deviations above the mean.
Symmetry check: Z-scores that are equal and opposite (for example, -1 and +1) give percentiles that sum to 100. A Z of -1 is at about the 15.9th percentile; a Z of +1 is at about the 84.1st percentile. They add to 100, which is a quick sanity-check for both conversion directions.
Common reference points
| Z-score | Percentile (approx.) | Interpretation |
|---|---|---|
| -2.0 | 2.3% | Well below average |
| -1.0 | 15.9% | Below average |
| 0 | 50.0% | Exactly average |
| +1.0 | 84.1% | Above average |
| +1.645 | 95.0% | Top 5% |
| +1.96 | 97.5% | Used in 95% confidence intervals |
| +2.0 | 97.7% | Top ~2.5% |
| +2.576 | 99.5% | Used in 99% confidence intervals |
| +3.0 | 99.9% | Exceptional outlier |
Where this comes up in practice
- Test score interpretation — SAT, GRE, IQ, and many other standardised tests publish norm-referenced percentile ranks alongside raw or scaled scores. Use this tool to convert between the reported scale and a percentile for comparison.
- Quality control — Z-scores describe how many standard deviations a measurement sits from a process mean. A Z of 3 (the “three-sigma” rule) marks a value so rare it usually signals a real defect rather than normal variation.
- Clinical reference ranges — growth charts and lab reference ranges are often expressed as standard deviations from the mean. A child’s height at Z = -2 is at approximately the 2nd percentile for their age group.
- Research and statistics — hypothesis testing uses Z-scores to locate a sample statistic in the distribution under the null hypothesis; the corresponding percentile is one minus the p-value (one-tailed) or twice that distance from 50% (two-tailed).
Everything runs in your browser using a numerical approximation accurate to about 1 part in 10 million — more than sufficient for all practical uses.