Statistics
Z-Score Calculator
Enter your details
Runs in your browser
How to use it
Using the z-score calculator
- 01
Enter your raw score
The individual value you want to place in context.
- 02
Add the population parameters
Use published or historical μ and σ for whatever reference group applies.
- 03
Read z and percentile
Positive z means above average; the percentile converts it into “beats X% of the group”.
Good to know
Why standardize?
Raw scores are only meaningful relative to their distribution. Being +1.5 SD above the mean on height and +0.2 SD above on a fitness test tells you immediately where someone truly stands; something raw units could never reveal across different measures.
Rules of thumb worth memorizing
- |z| < 2 covers ~95% of normal data.
- z > 2 or z < −2 is conventionally “unusual”.
- Percentile 84 ≈ z = 1; percentile 97.7 ≈ z = 2.
How it's calculated
The math behind this calculator
z = (x − μ) / σ Percentile = Φ(z)A z-score re-expresses a raw value as the number of standard deviations it sits above or below the mean. This strips away units and scale, letting you compare a SAT score against a height against a lab measurement on one common ruler.
The percentile comes from the standard normal CDF Φ(z), evaluated numerically with the Abramowitz–Stegun approximation (error below 7.5×10⁻⁸); accurate enough for any reporting purpose.
Assumptions & limitations
- You supply the true population mean and SD, not estimates.
- Percentiles assume the underlying data is approximately normally distributed.
- σ must be greater than zero.
Worked example
Scoring 85 on a test with μ = 70 and σ = 10 gives z = 1.5; about the 93.32nd percentile, meaning roughly 93% of test takers scored lower.
FAQ
Frequently asked questions
- Can a z-score be negative?
- Yes; negative z simply means the value sits below the mean. The percentile row handles interpretation automatically.
- What if I only have sample mean and SD?
- Results are still informative but technically a t-statistic; with samples under ~30, use t-distribution percentiles instead of the normal.
- How accurate is the percentile?
- The numerical approximation errs by less than 7.5×10⁻⁸ in CDF terms; effectively exact to the displayed decimals.
- Does the percentile assume normality?
- Yes. For heavily skewed data, report the empirical percentile from raw data instead of the normal-based one.
Keep exploring