What does a z-score tell you?

A z-score tells you how many standard deviations a data point is from the mean. A z-score of 0 means the value equals the mean. Positive z = above mean; negative z = below mean. The magnitude tells you how unusual: |z| 2 is unusual, |z| > 3 is very unusual for normal data.

What is a good z-score?

Depends on context. Z-scores between -2 and +2 are typical (covering about 95% of data in a normal distribution). For "good" outcomes (test scores, athletic performance): higher z = better. For "bad" outcomes (stock losses, medical complications): lower z (more negative) means worse. Beyond +/-3 indicates unusual values.

How do I calculate z-score from sample data?

For population: z = (x - μ) / σ where μ = population mean and σ = population SD. For sample data, use z = (x - x̄) / s where x̄ = sample mean and s = sample SD. For hypothesis testing with sample data, use t-distribution if sample size is small (< 30).

What's the z-score of the mean?

The z-score of the mean is 0. By definition: z = (mean - mean) / SD = 0. This represents the center of the distribution. The mean is the 50th percentile (median equals mean in symmetric distributions).

How do I convert z-score to percentile?

Use standard normal distribution lookup table or software. Roughly: z = -2 → 2.5%, z = -1 → 16%, z = 0 → 50%, z = +1 → 84%, z = +2 → 97.5%. In Excel: =NORM.S.DIST(z, TRUE). Note: this assumes normal distribution; for non-normal data, use empirical percentiles.

Can z-scores be negative?

Yes. A negative z-score means the value is below the mean. Z = -1.5 means the value is 1.5 SD below the mean. Z-scores can range from negative infinity to positive infinity (though practical bounds are usually ±5 or so). Absolute value of z represents distance from mean.

What's the difference between z-score and standardized score?

Z-score is a standardized score with mean 0 and SD 1. Other standardized scores transform z-scores to different scales: IQ has mean 100, SD 15 (so IQ = z × 15 + 100). SAT has mean 500 per section, SD 100. T-scores have mean 50, SD 10. All convey relative position but on different scales.

Z-Score Calculator

Q: What's the z-score of the mean?

The z-score of the mean is 0. By definition: z = (mean - mean) / SD = 0. This represents the center of the distribution. The mean is the 50th percentile (median equals mean in symmetric distributions).

Q: How do I convert z-score to percentile?

Use standard normal distribution lookup table or software. Roughly: z = -2 → 2.5%, z = -1 → 16%, z = 0 → 50%, z = +1 → 84%, z = +2 → 97.5%. In Excel: =NORM.S.DIST(z, TRUE). Note: this assumes normal distribution; for non-normal data, use empirical percentiles.

Q: Can z-scores be negative?

Yes. A negative z-score means the value is below the mean. Z = -1.5 means the value is 1.5 SD below the mean. Z-scores can range from negative infinity to positive infinity (though practical bounds are usually ±5 or so). Absolute value of z represents distance from mean.

Q: What's the difference between z-score and standardized score?

Z-score is a standardized score with mean 0 and SD 1. Other standardized scores transform z-scores to different scales: IQ has mean 100, SD 15 (so IQ = z × 15 + 100). SAT has mean 500 per section, SD 100. T-scores have mean 50, SD 10. All convey relative position but on different scales.

Enter a data value, the population mean, and standard deviation to compute the z-score. The z-score tells you how many standard deviations a value is from the mean.

A z-score tells you how unusual a value is compared to a distribution. Specifically, it measures how many standard deviations a data point sits from the mean. A z-score of 0 means the value is exactly at the mean; +1 means one standard deviation above; -2 means two standard deviations below. Z-scores let you compare values across different distributions and quickly assess how typical or extreme a value is.

This calculator returns the z-score from a single value, population mean, and standard deviation. The formula z = (x - μ) / σ is straightforward but the interpretation is what matters in practice. For roughly normally distributed data (which describes many real-world measurements), 68% of values have |z| < 1, 95% have |z| < 2, and 99.7% have |z| < 3 (the "68-95-99.7 rule" or "empirical rule").

Z-scores are foundational to confidence intervals, hypothesis testing, and standardized scoring systems (SAT, IQ, height percentiles for children). They're how statisticians and scientists communicate "how unusual is this?" in a standardized way that works across different measurement scales.

Inputs

Data Value (x)

Population Mean (μ)

Standard Deviation (σ)

Results

Z-Score

2.0000

Percentile

97.72%

Above This Value

2.28%

Interpretation

Within 3 standard deviations (uncommon)

Last updated: May 29, 2026

Formula

**Z-score formula:** z = (x - μ) / σ Where: - **x**: the data value of interest - **μ**: population mean - **σ**: population standard deviation **Interpretation:** - **z = 0**: value equals mean - **z = +1**: value is 1 SD above mean - **z = -1.5**: value is 1.5 SD below mean - **|z| > 2**: unusual (occurs in <5% of normal distribution) - **|z| > 3**: very unusual (occurs in <0.3%) **Worked example: test score** You scored 85. Class mean: 75. Class SD: 5. z = (85 - 75) / 5 = 10 / 5 = 2 You scored 2 standard deviations above the mean. In a normally distributed class, only ~2.5% of students score higher than you. **The 68-95-99.7 rule (empirical rule):** For data approximately normally distributed: | Range | Percentage of data | |---|---| | |z| ≤ 1 (within ±1 SD) | 68.27% | | |z| ≤ 2 (within ±2 SD) | 95.45% | | |z| ≤ 3 (within ±3 SD) | 99.73% | | |z| ≤ 4 | 99.99% | **Z-score and percentiles:** Z-score corresponds to a specific percentile in the standard normal distribution: | Z | Percentile | Interpretation | |---|---|---| | -3 | 0.13% | Bottom 0.13% | | -2 | 2.28% | Bottom 2.28% | | -1 | 15.87% | Bottom 15.87% | | 0 | 50% | Median | | +1 | 84.13% | Top 15.87% | | +2 | 97.72% | Top 2.28% | | +3 | 99.87% | Top 0.13% | **Standardized scores (built from z-scores):** - **IQ**: mean 100, SD 15. Score of 130 → z = 2; top 2.28% of population. - **SAT**: mean 1000, SD 200 (each section). Score of 1400 → z = 2. - **Body Mass Index (BMI) percentile**: pediatric reports use z-scores. - **Height/weight percentile (CDC)**: z-scores converted to percentile rank. **One-tailed vs two-tailed:** - **One-tailed z**: how unusual on one side only (e.g., how good or how bad). - **Two-tailed z**: how unusual either way (absolute value). - **Critical z values**: 1.645 for 5% one-tail, 1.96 for 5% two-tail. **Standardizing data:** z-scores are the "common currency" of statistics. By standardizing data, you can: - Compare values from different distributions. - Combine measurements with different units. - Identify outliers (|z| > 3). - Build predictive models with normalized features. **Using z-score for outliers:** A common rule: |z| > 3 indicates outlier. But this depends on: - Sample size (n). - Distribution shape. - Domain context. For small samples, modified z-score using median may be better. **Standard normal distribution:** z-scores follow the standard normal distribution N(0, 1) — bell-shaped curve with mean 0 and SD 1. This is the basis for all parametric tests and confidence intervals. **Confidence intervals from z:** For 95% confidence: z = ±1.96 For 99% confidence: z = ±2.576 For 90% confidence: z = ±1.645 CI = mean ± (z × SD / √n) **Z-test in hypothesis testing:** Test statistic: z = (sample mean - hypothesized mean) / (population SD / √n) Compare to critical z (e.g., ±1.96 for α = 0.05 two-tail). **Z-score in different contexts:** | Context | Use | |---|---| | Test grades | Compare student to class | | Stock returns | Volatility measure | | Quality control | Process capability | | Anthropometric | Growth percentiles | | Polling | Margin of error | | Risk management | Value at Risk | | Psychometrics | Standardized test scoring |

How to use this calculator

Enter the data value (x) you want to assess.
Enter the population mean (μ) of the distribution.
Enter the population standard deviation (σ).
Calculator returns the z-score.
Interpret: |z| < 1 = typical, |z| > 2 = unusual, |z| > 3 = very unusual.
For sample data (not population), use sample mean and sample SD with t-distribution.

Worked examples

Test score evaluation

**Scenario:** Student score: 92. Class mean: 75. Class SD: 8. **Calculation:** z = (92 - 75) / 8 = 17/8 = 2.125. Student scored 2.125 SD above mean. Corresponds to ~98.3rd percentile of a normal distribution. **Result:** Student is top ~1.7% of class. Strong performance. In a normal distribution, only ~2% of students would score this high or higher.

Stock price daily move

**Scenario:** Stock dropped 5% today. Mean daily move: 0.1%. SD: 1.5%. **Calculation:** z = (-5 - 0.1) / 1.5 = -3.4. The daily move is 3.4 SD below the mean. **Result:** |z| > 3 indicates very unusual day — beyond the 99.7% expected range. In ~250 trading days/year, you'd expect ~1 day with |z| > 3 in a normally-distributed market. This signals unusual market conditions.

Pediatric height percentile

**Scenario:** Child age 5 is 110 cm tall. Average 5-year-old height: 108 cm. SD: 4 cm. **Calculation:** z = (110 - 108) / 4 = 0.5. Child is 0.5 SD above mean. Corresponds to ~69th percentile. **Result:** Child is at about 69th percentile for height — taller than 69% of 5-year-olds. Normal range (z between -2 and +2): 100-116 cm. Outside this range may prompt growth chart review.

When to use this calculator

**Calculate z-scores for:**

- **Standardized testing**: SAT, IQ, GRE comparison across years. - **Quality control**: how unusual a measurement is. - **Hypothesis testing**: z-test statistics. - **Confidence intervals**: standard error calculations. - **Outlier detection**: |z| > 3 thresholds. - **Pediatric assessment**: growth chart standardized scores. - **Sports analytics**: player performance vs league averages. - **Risk management**: portfolio risk in standard deviations.

**Interpreting z-scores:**

| Z | Meaning | |---|---| | 0 | Equal to mean | | -1 to 1 | Within 1 SD of mean | | -2 to 2 | Within 2 SD (95% of normal data) | | -3 to 3 | Within 3 SD (99.7%) | | ±2 to ±3 | Unusual but plausible | | ±3 to ±4 | Very unusual | | Beyond ±4 | Extremely unusual or data error |

**Standardized scoring systems (built from z):**

- **IQ**: mean 100, SD 15. z × 15 + 100 = IQ. - **SAT**: mean 1000 (combined). z × 200 + section mean = SAT. - **ACT**: mean 21 (composite). Similar normalization. - **NIH/CDC growth charts**: z-scores converted to percentiles.

**Z vs t distribution:**

- **Z (standard normal)**: known population SD. - **t (Student's)**: estimated SD from sample. Use when sample size < 30. - **As sample size grows**: t approaches z.

**Common confusions:**

- **Z-score vs percentile**: z is in SD units; percentile is in 0-100 range. - **Population vs sample**: z-score uses population mean and SD. - **Standardized vs raw**: z is dimensionless, raw is in original units.

**Z-score advantages:**

- **Comparable across distributions**: same scale (SD units). - **Direct probability interpretation**: maps to standard normal. - **Dimensionless**: no units; pure ratio. - **Standard statistical input**: required by many tests.

**Z-score limitations:**

- **Assumes normality**: for non-normal data, interpretation differs. - **Requires population parameters**: rarely known exactly. - **Sensitive to outliers**: when calculated from sample SD. - **Equal weighting**: doesn't account for measurement uncertainty.

**Modified z-score (for outliers):**

Modified z = 0.6745 × (x - median) / MAD

Where MAD = median absolute deviation. More robust to outliers than standard z.

**Common applications:**

| Field | Use | |---|---| | Education | Test scoring, percentile reports | | Medicine | Growth charts, lab result interpretation | | Manufacturing | Process control, defect rates | | Finance | Risk assessment, VaR calculations | | Sports | Player ranking, season comparisons | | Marketing | Customer segmentation | | Research | Standardizing variables for analysis |

**Reporting z-scores:**

- Always include mean and SD reference. - Specify whether one-tailed or two-tailed. - Note distribution assumption (normal?). - For non-normal data, use percentile instead.

**Cautions:**

- For non-normal data, z-scores don't have the typical probability interpretation. - Very large or small datasets affect the practical meaning of "unusual." - Domain context matters — z = 3 in one field may be unremarkable in another. - Multiple comparisons: when checking many z-scores, some will appear extreme by chance.

Common mistakes to avoid

Using sample SD when population SD is needed. Z-score requires population parameters.
Interpreting z-score as a percentage. It's in SD units; convert to percentile if needed.
Applying 68-95-99.7 rule to non-normal data. The rule only holds for normal distributions.
Treating |z| > 2 as "significant" without considering sample size and multiple tests.
Confusing z-score with z-test. Z-score is for individual values; z-test is for hypothesis testing.
Forgetting absolute value when discussing magnitude. Both +2 and -2 are "2 SD from mean."
Using z when t-distribution is needed (small samples). t-test more appropriate for n < 30.

Frequently Asked Questions

Sources & further reading

SponsoredShop Top Deals on AmazonSupport CalcMountain — browse top-rated products at no extra cost to you.