Tip: Use Score → Percentile to find rank, or Percentile → Score to find a cutoff value.
What this normal percentile calculator does
This calculator helps you work with values from a normal distribution. In practical terms, it tells you either:
- the percentile rank for a score (for example, “What percentile is a test score of 85?”), or
- the score at a chosen percentile (for example, “What score is the 90th percentile?”).
It is useful for exam analysis, quality control thresholds, employee performance metrics, and any process that is approximately bell-shaped.
How to use it
1) Score → Percentile
Enter your observed value x, the distribution mean μ, and the standard deviation σ. The calculator returns:
- the z-score,
- the percentile at or below that value,
- and the percentage above that value.
2) Percentile → Score
Enter a percentile between 0 and 100 (exclusive), along with mean and standard deviation. You’ll get:
- the corresponding z-score, and
- the actual score cutoff in original units.
Core formulas behind the calculator
First, the score is standardized using:
z = (x - μ) / σ
For Score → Percentile, the calculator uses the standard normal cumulative distribution function (CDF):
Percentile = Φ(z) × 100
For Percentile → Score, it applies the inverse normal CDF:
x = μ + σ × Φ-1(p), where p = percentile / 100.
Interpretation guide
| Percentile | Meaning |
|---|---|
| 50th | Exactly average (median in a normal distribution) |
| 75th | Higher than 75% of values, lower than 25% |
| 90th | Top 10% threshold |
| 95th | Top 5% threshold (often used for screening cutoffs) |
| 99th | Top 1% (very rare in many contexts) |
Example
Suppose scores are normally distributed with mean 100 and standard deviation 15. A score of 130 gives:
- z = (130 - 100) / 15 = 2.0
- Percentile ≈ 97.7th
So a score of 130 is better than roughly 97.7% of scores in that distribution.
Common mistakes to avoid
- Using a negative or zero standard deviation (not valid).
- Confusing “percent” with “percentile” (90% is not always 90th percentile in raw data).
- Applying normal assumptions to strongly skewed data without checking fit.
- Entering percentile 0 or 100 for inverse calculations (these imply infinite tails).
When this tool is most reliable
Results are best when the data are approximately normal (bell-shaped), especially near the center. For heavily skewed or bounded data, percentile estimates from empirical data or nonparametric methods may be more appropriate.