Z-Score Calculator
A z-score answers the question "how unusual is this, really?" by expressing any raw number as how many standard deviations it sits from the average. An 85 on a test means nothing until you know the class averaged 75 with a standard deviation of 10 — then 85 is exactly one standard deviation above, putting it in the top ~16%. Enter your value, the mean, and the standard deviation to get the z-score and its true percentile, or flip it: enter a target percentile to find the cutoff score.
Score to Z-Score
Reverse: Target Percentile → Score
The Z-Score Formula
Standardizing a value strips away its original units and expresses position relative to the spread of the data:
z = (x − μ) ÷ σ
Reverse it to recover a raw score from a target percentile: x = μ + z × σ, where z is the standard normal cutoff for that percentile (about +1.28 for the 90th, +1.645 for the 95th, +1.96 for the 97.5th). Because the standard normal is symmetric, the 5th percentile is just the negative of the 95th.
Worked Example: An 85 on a Tough Exam
The class averaged 75 with a standard deviation of 10. Your 85 gives z = (85 − 75) ÷ 10 = +1.00, which lands at the 84.1th percentile — you beat about 84% of the class. Now suppose the department awards the top 10% a prize. The 90th percentile needs z ≈ 1.282, so the cutoff score is 75 + 1.282 × 10 = 87.82. Your 85 falls just short; you'd need roughly an 88 to make the cut. Same distribution, two completely different stories depending on which direction you read it.
When Z-Scores Are (and Aren't) the Right Tool
Z-scores assume the data are at least roughly symmetric and bell-shaped. For genuinely skewed distributions (income, house prices), the mean and standard deviation are themselves misleading, and a percentile rank from the raw data (or a log transform) is more honest. For test scores, heights, measurement errors, and most naturally-occurring biological and industrial metrics, the normal assumption behind this calculator is a very good approximation.
Real-World Case Study: SAT Score Percentile Ranking
A student scored 1350 on the SAT (national mean: ~1050, SD: ~200). Z-score = (1350 − 1050) ÷ 200 = +1.50. This corresponds to the 93.3rd percentile — beating 93% of test-takers. A peer with a 1200 scored z = 0.75 (77.3rd percentile). The gap of 150 points sounds large, but in percentile terms it's 16 percentage points — significant but not transformative for admissions. The calculator confirmed that the student needed at least 1450 (z = 2.00, 97.7th percentile) to reach the top 2.5% for Ivy League competitiveness.
Key takeaway: Z-scores turn raw scores into meaningful comparisons across different distributions. The same score can be exceptional in one context and average in another.
Z-Score Quick Reference: Common Benchmarks
The table below maps z-scores to percentiles and real-world equivalents for normally distributed metrics like test scores, heights, and IQ.
| Z-Score | Percentile | % Within ±z | Real-World Example |
|---|---|---|---|
| −3.00 | 0.1% | 99.7% | Extremely rare low value |
| −2.00 | 2.3% | 95.4% | SAT 650 (well below average) |
| −1.00 | 15.9% | 68.2% | Below average but common |
| 0.00 | 50.0% | — | Exactly at the mean |
| +1.00 | 84.1% | 68.2% | Above average (top third) |
| +2.00 | 97.7% | 95.4% | IQ 130, gifted range |
| +3.00 | 99.9% | 99.7% | Extremely rare high value |
Note: The 68-95-99.7 rule holds for perfectly normal distributions. Real data may deviate slightly, but this table is accurate to within 1% for most test scores, heights, and measurement errors.
Frequently Asked Questions
What is a z-score?
It's how many standard deviations a value is from the mean: z = (x − μ) ÷ σ. It puts values from different scales on a common footing.
How do you interpret a z-score?
Sign = direction from mean; magnitude = distance in SDs. ~68% of values sit within ±1, 95% within ±2, 99.7% within ±3.
What percentile does a z-score correspond to?
The area under the normal curve to the left of z. z = 1.00 → 84.1th percentile; z = −1.00 → 15.9th. We compute it via the normal CDF.
Can a z-score be negative?
Yes — any value below the mean is negative. A 60 with mean 75, SD 10 gives z = −1.5 (about the 6.7th percentile).
How do you find the value at a given percentile?
Use x = μ + z × σ with the inverse-normal z for that percentile. Enter a target % and the tool returns the cutoff score.