What is an Empirical Rule Calculator?
The empirical rule, also called the 68-95-99.7 rule, says that in a normal distribution about 68% of values lie within one standard deviation of the mean, about 95% within two and about 99.7% within three. Enter a mean and standard deviation and this calculator gives those three ranges in your units, the exact normal percentages, the percent between or outside any two values, and a bell curve graph.
What the Empirical Rule Calculator does
Enter the mean and standard deviation to get the intervals μ ± σ, μ ± 2σ and μ ± 3σ, the rule-of-thumb and exact normal coverage for each (68.27%, 95.45% and 99.73%), the share outside each range, and Chebyshev’s guaranteed minimum for comparison. A second table splits the curve into the 34%, 13.5%, 2.35% and 0.15% bands, so you can read off any percentage the rule gives.
Add an optional value x to see its z-score, its percentile and which band it falls in. Add a second value x₂ as well to get the percent of values between the two and the percent outside them. When both values sit on whole standard deviations, the calculator adds the empirical rule bands; otherwise it reports the exact normal share, the same as the Normal Distribution Calculator.
This tool is also known as a 68-95-99.7 rule calculator, a 68 95 99 rule calculator or an empirical rule solver. Each result comes with a step-by-step solution and an empirical rule graph that shades the three ranges on the normal curve.
How to use the Empirical Rule Calculator
- Enter the Mean, μ and the Standard deviation, σ of your data or population.
- Read the 68%, 95% and 99.7% ranges at the top of the results, and the band table for the 34-13.5-2.35-0.15 split.
- Optionally enter a Value to locate, x to see how many standard deviations it is from the mean and its percentile.
- To find the percent between two values, enter both x and x₂. The percent outside them is shown underneath.
- Check the first table for the exact normal percentages, and Chebyshev’s bound if your data may not be normal.
Empirical rule chart: the 34-13.5-2.35-0.15 breakdown
The normal curve is symmetric, so each of the three ranges splits evenly around the mean. Half of 68% gives 34% between the mean and one standard deviation on each side. The next band holds half of 95% − 68%, or 13.5%, on each side. The third holds half of 99.7% − 95%, or 2.35%, and the last 0.15% on each side lies beyond three standard deviations.
To find the percent between any two whole-σ cutoffs, add the bands between them. For example, from μ − σ to μ + 2σ is 34 + 34 + 13.5 = 81.5%, and from the mean to μ + 2σ is 34 + 13.5 = 47.5%. The percent above μ + σ is 13.5 + 2.35 + 0.15 = 16%, which puts μ + σ at about the 84th percentile.
| Band | Rule of thumb | Exact (normal) | Cumulative percentile at the upper edge |
|---|---|---|---|
| Below μ − 3σ | 0.15% | 0.13% | 0.15% |
| μ − 3σ to μ − 2σ | 2.35% | 2.14% | 2.5% |
| μ − 2σ to μ − σ | 13.5% | 13.59% | 16% |
| μ − σ to μ | 34% | 34.13% | 50% |
| μ to μ + σ | 34% | 34.13% | 84% |
| μ + σ to μ + 2σ | 13.5% | 13.59% | 97.5% |
| μ + 2σ to μ + 3σ | 2.35% | 2.14% | 99.85% |
| Above μ + 3σ | 0.15% | 0.13% | 100% |
Empirical rule formula
| Symbol | Meaning |
|---|---|
| μ | mean |
| σ | standard deviation |
| k | number of standard deviations from the mean |
| z | z-score: how many standard deviations a value is from the mean |
| Φ | standard normal cumulative distribution |
To calculate the empirical rule by hand, multiply σ by 1, 2 and 3, then subtract each result from the mean and add it to the mean. Those six numbers are the bounds of the 68%, 95% and 99.7% ranges.
The exact shares come from the standard normal distribution; see the NIST/SEMATECH e-Handbook of Statistical Methods for its definition and properties.
The rounded 95% is a slight overstatement of 95.45%; the exact 95% range is μ ± 1.96σ. That is why 1.96, not 2, appears in 95% confidence intervals.
Empirical rule worked example
Adult heights (μ = 170 cm, σ = 7 cm)
Problem. Adult heights in a population are roughly normal with μ = 170 cm and σ = 7 cm. Apply the empirical rule and locate a height of 160 cm.
- 01μ ± σ = 170 ± 7 = 163 to 177 cm (about 68%).
- 02μ ± 2σ = 170 ± 14 = 156 to 184 cm (about 95%).
- 03μ ± 3σ = 170 ± 21 = 149 to 191 cm (about 99.7%).
- 04For x = 160: z = (160 − 170) / 7 = −1.4286, so it is between 1 and 2 SDs below the mean.
Answer. About 68% of adults are 163–177 cm tall, 95% are 156–184 cm and nearly all are 149–191 cm. A height of 160 cm lies outside μ ± σ but inside μ ± 2σ, at about the 7.66th percentile, so it is short but not unusual.
More empirical rule examples
Percent between two values (IQ scores). IQ is scaled to μ = 100 and σ = 15. What share of people score between 85 and 130? 85 is z = −1 and 130 is z = 2, so add the bands: 34% + 34% + 13.5% = 81.5%. The exact normal figure is 81.86%.
Percent above a value (IQ above 130). 130 is μ + 2σ. Above it lie the two outer bands, 2.35% + 0.15% = 2.5%, or about 1 person in 40. The exact figure is 2.28%.
Percent outside a range (exam scores). Scores are normal with μ = 68 and σ = 9. The range 59 to 77 is μ ± σ, so about 68% fall inside and 32% fall outside it, 16% on each side. In a class of 200, expect about 64 students outside that range (exact 31.73%, about 63).
Expected counts (fill volumes). A machine fills bottles with μ = 500 ml and σ = 2.5 ml, and the specification is 495 to 505 ml, which is μ ± 2σ. About 95% are in spec and 5% are out, so a run of 10,000 bottles gives roughly 500 rejects. The exact normal share outside is 4.55%, about 455 bottles.
From a percentage back to a value. Which exam score marks the top 16%? Since 16% lies above μ + σ, the cutoff is 68 + 9 = 77. The top 2.5% starts at μ + 2σ = 68 + 18 = 86. For other percentages, use the Z-Score to Percentile Calculator or the Percentile to Z-Score Calculator.
How to interpret the 68-95-99.7 rule
Think of the three ranges as typical, unusual and extreme. A value inside μ ± σ is ordinary; about a third of all values fall outside that range, so being just beyond one standard deviation is common. A value beyond μ ± 2σ happens about 1 time in 20, so it is unusual. A value beyond μ ± 3σ happens about 3 times in 1,000, so it is rare enough to check for a data error or a special cause.
Each range is about individual values, not about the mean. “95% of values lie within μ ± 2σ” describes the spread of the data. It is not a 95% confidence interval, which describes uncertainty in an estimated mean and uses the standard error instead of σ.
To place one value on the scale, find its z-score: z = (x − μ) / σ. The sign tells you the side of the mean and the size tells you which band it falls in. The z table gives the exact percentile for any z.
When the data are not normal: Chebyshev’s inequality
The 68-95-99.7 figures only hold for bell-shaped data. Chebyshev’s inequality gives a floor that holds for every distribution with a finite variance: at least 1 − 1/k² of values lie within k standard deviations of the mean. It says nothing for k = 1, guarantees 75% for k = 2 and 88.9% for k = 3.
In practice, real skewed data usually sit between the two: more spread out in one tail than the normal predicts, but well inside Chebyshev’s worst case.
| k | Empirical rule (normal) | Exact normal | Chebyshev minimum (any shape) |
|---|---|---|---|
| 1 | ≈ 68% | 68.27% | 0% (no guarantee) |
| 2 | ≈ 95% | 95.45% | 75% |
| 3 | ≈ 99.7% | 99.73% | 88.89% |
Empirical rule in Excel, Google Sheets, TI-84 and R
There is no built-in empirical rule function, but each tool can compute the ranges and the exact shares. With the mean in cell A1 and the SD in B1:
- Excel or Google Sheets, ranges:
=A1-2*B1and=A1+2*B1give the 95% range; change 2 to 1 or 3 for the others. - Excel or Google Sheets, exact share:
=NORM.DIST(A1+B1,A1,B1,TRUE)-NORM.DIST(A1-B1,A1,B1,TRUE)returns 0.6827. UseNORM.S.DIST(1,TRUE)-NORM.S.DIST(-1,TRUE)for the standard normal. - TI-83/84:
normalcdf(μ−σ, μ+σ, μ, σ)returns 0.6827; use 2σ or 3σ for 0.9545 and 0.9973. - R:
pnorm(1) - pnorm(-1)returns 0.6827, andmu + c(-1, 1) * 2 * sigmagives the 95% range.
Common use cases
- Quick checks on whether a value is typical, unusual (beyond 2σ) or extreme (beyond 3σ).
- Estimating ranges for normally distributed measurements, such as heights, test scores, IQ or fill volumes.
- Estimating how many items in a batch fall outside a specification without a z-table.
- Setting rough control limits at μ ± 3σ in process monitoring.
- Teaching how standard deviation describes spread for bell-shaped data.
Assumptions and limitations
- Only valid for approximately normal (symmetric, bell-shaped) data. For skewed data, use Chebyshev’s bound or look at the actual percentiles.
- The percentages are approximations; use the exact values (68.27%, 95.45%, 99.73%) or the Normal Distribution Calculator when precision matters.
- The band shares can only be added for cutoffs on whole standard deviations. For other cutoffs, the calculator switches to the exact normal share.
- With a sample, x̄ and s are estimates, so the ranges describe the population only approximately, especially for small samples. The Standard Deviation Calculator finds s from raw data.
- For individual-value ranges this rule describes the data spread; it is not a confidence interval for the mean. For that, see the Mean Confidence Interval Calculator.
Frequently asked questions
What is the 68-95-99.7 rule?
How do you calculate the empirical rule?
What percentage of data falls within 2 standard deviations of the mean?
What do 34%, 13.5%, 2.35% and 0.15% mean in the empirical rule?
How do I find a z-score for the empirical rule?
Does the empirical rule work for skewed data?
Why is a value beyond 3 standard deviations considered an outlier?
Cite this page
Statculator. (2026, November 14). Empirical Rule Calculator. Retrieved November 14, 2026, from https://statculator.com/normal-distribution/empirical-rule-calculator/