Tools logoTools/
Math & Science/Z-Score Calculator

Z-Score Calculator

Standard score (z) plus the corresponding probability from a normal distribution.

Runs locally·Free, no signup·Updated May 6, 2026
Loading tool…
How it works

A walkthrough, end to end.

  1. 1

    Enter the value (x), the population mean (μ), and the standard deviation (σ).

  2. 2

    The calculator returns the z-score and the cumulative probability P(Z ≤ z).

  3. 3

    Useful for hypothesis testing, percentile calculation, and outlier detection.

Reference

Z-score formula

z = (x − μ) / σ. Standardizes any value into the standard-normal scale (mean 0, SD 1). The cumulative probability comes from the normal CDF Φ(z).

Use cases

What you can do with this.

Test score percentile

Score 85, mean 75, SD 10 → z = 1.0. Probability P(Z ≤ 1) = 0.8413, so the score is at the 84th percentile.

Detecting outliers

|z| > 2 typically marks moderate outlier (5% of data); |z| > 3 marks extreme outlier (0.3%). Use to flag unusual values in QC or screening.

Comparing different scales

Convert two scores from different distributions into z-scores → directly comparable. e.g., comparing SAT (mean 500, SD 100) vs ACT (mean 21, SD 5) — both reduce to z.

Six Sigma quality

Six Sigma target: defect rate at z = 6, meaning 99.99966% of products within spec. Industry-standard for high-precision manufacturing.

Heights / IQ scores

IQ standardized to mean 100, SD 15. IQ 130 → z = 2 → 97.7th percentile (top 2.3%). The calculator gives both z and percentile in one shot.

Z-score from raw data

If you have raw data, compute mean and SD first (Standard Deviation Calculator), then use the result here to standardize specific values.

Z vs t test

Z used when population SD known. t used when only sample SD available and n small. The z-score calculator is for the z-test side; use t-distribution tools for sample-only stats.

Z-score 2026 — what's current

Foundational stats. AI tools, spreadsheets, and dedicated stats software (R, Python) all compute reliably. Calculator wins for one-off speed.

FAQ

Frequently asked.

  • Depends on context. For percentile: higher = higher rank. For 'normal' values: |z| < 1 (within 1 SD of mean) is the typical 68%. |z| > 2 is 5% tails.

  • Z uses population parameters (μ, σ). T uses sample estimates (x̄, s) and a t-distribution that accounts for sampling error in small samples.

  • Yes — when the value is below the mean. Negative z means below; positive means above. Magnitude is what matters for outlier detection.

  • No. Calculations run entirely in your browser.