Z-Score Calculator | Standard Score & Percentile
Calculate z-score, percentile rank, and p-values from any data point, mean, and standard deviation. Includes reverse calculation and step-by-step solutions.
What Is the Z-Score Calculator | Standard Score & Percentile?
The Z-Score Calculator is a full-featured statistics tool that converts any raw data point into a standardized score, and goes far beyond a basic z-score computation. In a single click you get the z-score, the exact percentile rank, one-tailed and two-tailed probabilities, and a plain-English interpretation of the result.
It also works in reverse: switch to "Find X from Z-Score" mode and enter a known z-score with the distribution parameters to recover the corresponding raw data value. Every result includes a step-by-step arithmetic breakdown so you can follow and verify each calculation.
Probabilities are computed using the error function (erf) approximation, accurate to 6+ decimal places, far more precise than reading from a printed z-table.
Formula
z = (x − μ) / σx = μ + z × σP(Z < z) = Φ(z)P(Z > z) = 1 − Φ(z)P(|Z| > |z|) = 2 × min(Φ(z), 1−Φ(z))How to Use
Mode 1, Calculate Z-Score x → z
- 1Select mode: Make sure "Calculate Z-Score" is highlighted (it is the default).
- 2Data Point (x): Enter the individual observation you want to standardize, e.g. a student's exam score, a product's weight, or a stock's daily return.
- 3Population Mean (μ): Enter the average of the full dataset or distribution.
- 4Standard Deviation (σ): Enter the standard deviation, this must be greater than zero.
- 5Click Calculate: Instantly see the z-score, percentile rank, all three probability measures, a step-by-step solution, and an interpretation.
Mode 2, Find X from Z-Score z → x
- 1Switch mode: Click "Find X from Z-Score", the button turns orange to confirm the switch.
- 2Z-Score (z): Enter the known z-score. Negative values are valid (e.g. −1.645 for the 5th percentile).
- 3Mean (μ) and Std Dev (σ): Enter the distribution parameters as above.
- 4Click Calculate: The calculator returns the raw data value (x), its percentile, and all probability measures.
Example Calculation
Example 1, Student exam score
A student scored 92 on a test where the class mean is 78 and the standard deviation is 8.
The student is 1.75 standard deviations above the mean, placing them at the 95.99th percentile, scoring higher than ~96% of the class.
Example 2, Population height
A person is 183 cm tall. Population mean is 170 cm, σ = 10 cm.
This person is taller than approximately 90.32% of the population.
Example 3, Reverse: find the cutoff score for top 5%
Using the same test (μ = 78, σ = 8), what score marks the top 5%? The z-score for the 95th percentile is 1.645.
A student must score 91.16 or higher to be in the top 5% of the class.
Example 4, Manufacturing quality control
A factory targets bolt diameter of 10 mm with σ = 0.05 mm. A bolt measures 10.12 mm.
The bolt is 2.4σ outside target, beyond the standard ±2σ control limit. This signals a quality issue requiring investigation.
Understanding Z-Score | Standard Score & Percentile
What Is a Z-Score?
A z-score (also called a standard score or standardized score) is a statistical value that describes how far, and in which direction, a data point is from the mean of its distribution, measured in units of the standard deviation. Because z-scores are dimensionless, they let you compare values from completely different datasets on a single, universal scale.
- Positive z-score, the value is above the mean.
- Negative z-score, the value is below the mean.
- Z = 0, the value equals the mean exactly.
- |z| = 1, the value is exactly one standard deviation from the mean.
Breaking Down the Z-Score Formula
The formula z = (x − μ) / σ has three inputs:
| Variable | Name | What it represents |
|---|---|---|
| x | Observed value | The individual data point being standardized |
| μ (mu) | Population mean | The average of the entire distribution or dataset |
| σ (sigma) | Population std dev | How spread out the distribution is; must be > 0 |
| z | Z-score (output) | Standard deviations x is from μ, positive = above, negative = below |
For sample data (when the full population is unavailable), substitute the sample mean x̄ for μ and the sample standard deviation s for σ. The formula is otherwise identical.
The Standard Normal Distribution
When you compute a z-score you are transforming your data onto the standard normal distribution, a symmetric, bell-shaped curve with μ = 0 and σ = 1. This reference distribution is the backbone of inferential statistics because its probabilities are fully tabulated (z-table) and can be computed precisely with the cumulative distribution function Φ(z).
- Φ(z) gives the left-tail probability, the fraction of the distribution below z.
- The right-tail probability is 1 − Φ(z), the fraction above z.
- The two-tailed probability is 2 × min(Φ(z), 1 − Φ(z)), used in hypothesis testing.
The Empirical Rule (68 – 95 – 99.7)
For any normally distributed dataset, z-scores map to the following coverage ranges, a fact known as the empirical rule:
- z = ±1, covers 68.27% of all values (within one standard deviation of the mean).
- z = ±2, covers 95.45% of all values.
- z = ±3, covers 99.73% of all values.
- |z| > 3, only 0.27% of values; considered statistical outliers.
Quick rule: if |z| > 2, the value is in the outer 5% of the distribution, unusual. If |z| > 3, it is in the outer 0.3%, extremely rare.
Z-Score to Percentile Reference Table
The most commonly used z-score critical values and their exact percentiles:
| Z-Score | Percentile | Left tail P(Z < z) | Right tail P(Z > z) | Common use |
|---|---|---|---|---|
| −3.000 | 0.13th | 0.00135 | 0.99865 | Extreme low outlier |
| −2.326 | 1.00th | 0.01000 | 0.99000 | 1% significance (one-tail) |
| −1.960 | 2.50th | 0.02500 | 0.97500 | 5% significance (two-tail) |
| −1.645 | 5.00th | 0.05000 | 0.95000 | 5% significance (one-tail) |
| −1.282 | 10.00th | 0.10000 | 0.90000 | 10% significance (one-tail) |
| 0.000 | 50.00th | 0.50000 | 0.50000 | Mean of distribution |
| +1.282 | 90.00th | 0.90000 | 0.10000 | Top 10% |
| +1.645 | 95.00th | 0.95000 | 0.05000 | 5% significance (one-tail) |
| +1.960 | 97.50th | 0.97500 | 0.02500 | 5% significance (two-tail) |
| +2.326 | 99.00th | 0.99000 | 0.01000 | 1% significance (one-tail) |
| +3.000 | 99.87th | 0.99865 | 0.00135 | Extreme high outlier |
Real-World Applications of Z-Scores
Z-scores appear across virtually every data-driven discipline:
Education & Standardized Testing
- SAT, GRE, and IQ tests convert raw scores to a common scale using z-score transformations.
- Grading "on a curve", assigning letter grades based on percentile rank rather than raw score.
- Identifying students who need support (z < −2) or are gifted (z > 2).
Medicine & Clinical Research
- Pediatric growth charts express height and weight as z-scores relative to population norms.
- Clinical trials use z-statistics to test whether a treatment effect is statistically significant.
- Lab results are flagged when they fall beyond ±2σ of the reference range for a patient's demographic.
Finance & Risk Management
- The Altman Z-Score uses five financial ratios to predict corporate bankruptcy risk.
- Value at Risk (VaR) models use normal distribution z-scores to estimate portfolio loss probabilities.
- Six Sigma quality processes target z ≥ 6, meaning fewer than 3.4 defects per million opportunities.
Data Science & Machine Learning
- Z-score normalization (standardization) scales features to mean = 0, σ = 1 before ML training.
- Outlier detection: data points with |z| > 3 are flagged as anomalies for review.
- Feature scaling ensures variables measured in different units contribute equally to distance-based algorithms (k-NN, SVM, PCA).
Z-Score vs. T-Score: When to Use Which
Both z-scores and t-scores measure how far a sample statistic is from the null hypothesis value, but they differ in when they are appropriate:
| Situation | Z-Score | T-Score |
|---|---|---|
| Population σ is known | ✓ | |
| Large sample (n ≥ 30) | ✓ | (either works) |
| Small sample (n < 30) | ✓ | |
| σ must be estimated from sample | ✓ | |
| Normal population assumed | ✓ | ✓ |
As sample size grows, the t-distribution approaches the standard normal distribution, so z and t give practically identical results for n ≥ 30.
How This Calculator Computes Probabilities
Most z-tables are rounded to four decimal places, which limits precision for scientific or professional work. This calculator uses the Abramowitz & Stegun error function (erf) approximation with a maximum error of 1.5 × 10⁻⁷, providing results accurate to six or more decimal places without requiring a lookup table. The underlying math:
- The standard normal CDF is Φ(z) = 0.5 × (1 + erf(z / √2)).
- The erf function is approximated by a fifth-degree polynomial, fast and highly accurate.
- All three probability outputs (left tail, right tail, two-tailed) are derived from this single CDF value.
Frequently Asked Questions
What is a z-score?
A z-score (also called a standard score or standardized score) is a statistical measure that describes how many standard deviations a data point is from the mean of its distribution. A z-score of +1.5 means the value is 1.5 standard deviations above the mean; −2 means it is 2 standard deviations below. Z-scores allow fair comparison of values from completely different datasets by putting them on a common, unit-free scale.
What is the z-score formula?
The formula is z = (x − μ) / σ, where x is the raw data point, μ (mu) is the population mean, and σ (sigma) is the population standard deviation. For sample data, substitute x̄ (sample mean) and s (sample standard deviation). The result is the signed number of standard deviations the value lies above or below the mean.
What does a z-score of 0, +1, or −2 mean?
A z-score of 0 means the value equals the mean exactly. +1 means the value is one standard deviation above the mean, roughly the 84th percentile in a normal distribution. −2 means two standard deviations below, roughly the 2.3rd percentile. In a normal distribution, about 68% of data falls within ±1σ, 95% within ±2σ, and 99.7% within ±3σ.
How do I convert a z-score to a percentile?
The percentile is found using the cumulative distribution function (CDF) of the standard normal distribution, Φ(z). This calculator computes it automatically using the error function. Common reference points: z = 1.28 → 90th percentile; z = 1.645 → 95th percentile; z = 2.326 → 99th percentile. Negative z-scores correspond to percentiles below 50%.
What is the difference between one-tailed and two-tailed probability?
A one-tailed probability is the area under the normal curve on one side of z, either P(Z < z) (left tail) or P(Z > z) (right tail). A two-tailed probability measures the chance of a value being at least as extreme in either direction: P(|Z| > |z|) = 2 × min(Φ(z), 1 − Φ(z)). Use two-tailed tests when testing whether a value is unusually large OR unusually small; use one-tailed when you care about one direction only.
When should I use a z-score vs. a t-score?
Use a z-score when you know the population standard deviation (σ) and have a large sample (n ≥ 30). Use a t-score (from the t-distribution) when working with a small sample where you estimate σ from sample data. As sample size increases, the t-distribution converges to the standard normal, making z and t nearly identical for large n.
Can a z-score be negative?
Yes. A negative z-score simply means the raw value is below the mean. For example, a score of 60 when the mean is 75 and σ is 10 gives z = (60 − 75) / 10 = −1.5. The sign shows direction (above or below the mean) while the magnitude shows distance from it.
How is a z-score used in hypothesis testing?
In a z-test, you compute a z-statistic from your sample data and compare it to a critical value. Common critical values: ±1.645 (5% significance, two-tailed) and ±1.96 (5% significance, two-tailed for proportions). If your z-statistic exceeds the critical value, you reject the null hypothesis. The p-value is the probability of observing a z-score as extreme as yours if the null hypothesis were true, this calculator computes it directly.
What is considered an outlier based on z-score?
The most common rule is that any data point with |z| > 3 is a statistical outlier, it lies more than three standard deviations from the mean and represents fewer than 0.3% of values in a normal distribution. Some fields use |z| > 2 (the outer 5%) as their threshold, especially in quality control.
What is z-score normalization in machine learning?
Z-score normalization (also called standardization) rescales every feature in a dataset so it has mean = 0 and standard deviation = 1, using the same formula: z = (x − μ) / σ. This is a standard preprocessing step before algorithms like SVM, k-NN, logistic regression, and PCA, ensuring that no single feature dominates due to its original scale.