When can I use the empirical rule?

Only when data is approximately normally distributed (bell-shaped and symmetric). Check with a histogram, Q-Q plot, or normality test first. For non-normal data, use Chebyshev's theorem instead, which gives weaker but universally valid bounds.

What does it mean to be "within 2 standard deviations"?

It means the value is between μ − 2σ and μ + 2σ. For IQ (μ=100, σ=15), within 2 SDs means between 70 and 130. About 95% of the population falls in this range, so values outside it are considered unusual.

Is the empirical rule exact?

It's an approximation. The exact values for a normal distribution are 68.27%, 95.45%, and 99.73%. The rounded 68-95-99.7 is close enough for quick mental calculations and back-of-envelope estimates.

How do I identify an outlier using this rule?

Values beyond 3 standard deviations are often flagged as potential outliers since less than 0.3% of normal data falls there. For a stricter criterion, some use 2σ (outside 95%). However, for non-normal data, the thresholds should be adjusted.

What is the relationship between the empirical rule and z-scores?

The empirical rule uses z = 1, 2, 3 as boundaries. Any z-score tells you how many standard deviations a value is from the mean. The rule simplifies: |z| ≤ 1 is common (68%), |z| ≤ 2 is typical (95%), |z| > 3 is very rare (0.3%).

Why is the empirical rule better than Chebyshev's theorem for normal data?

Chebyshev guarantees at least 75% within 2σ for any distribution; the empirical rule gives 95% for normal data. Chebyshev has to work for heavily skewed and heavy-tailed distributions, so its bounds are conservative. When you know data is normal, the empirical rule is much more informative.

Empirical Rule Calculator (68-95-99.7)

Apply the 68-95-99.7 rule to normal distributions with bell curve visualization, section probabilities, z-score lookup, and Chebyshev comparison table.

Mean (μ)

Standard Deviation (σ)

Query Value (x)

68% Interval (±1σ)

85.00 to 115.00

μ ± 1σ

95% Interval (±2σ)

70.00 to 130.00

μ ± 2σ

99.7% Interval (±3σ)

55.00 to 145.00

μ ± 3σ

Z-score of 130.00

2.0000

2.0 standard deviations from mean

Percentile (below x)

97.72%

2.28% above

Within ±kσ

Within 2σ (95%)

|z| = 2.00

Bell Curve Visualization

−3σ−2σ−1σμ+1σ+2σ+3σ

■ 68%■ 95%■ 99.7%| x = 130.0

Section Probabilities

Section	Range	Probability
Below μ − 3σ	−∞ to 55.0	0.13%
μ − 3σ to μ − 2σ	55.0 to 70.0	2.14%
μ − 2σ to μ − 1σ	70.0 to 85.0	13.59%
μ − 1σ to μ	85.0 to 100.0	34.13%
μ to μ + 1σ	100.0 to 115.0	34.13%
μ + 1σ to μ + 2σ	115.0 to 130.0	13.59%
μ + 2σ to μ + 3σ	130.0 to 145.0	2.14%
Above μ + 3σ	145.0 to ∞	0.13%

Empirical Rule vs. Chebyshev\'s Theorem

k (SDs)	Empirical (Normal)	Chebyshev Min	Actual Normal
1	68%	0.0%	68.269%
1.5	—	55.6%	86.639%
2	95%	75.0%	95.450%
2.5	—	84.0%	98.758%
3	99.7%	88.9%	99.730%
4	—	93.8%	99.994%

Planning notes, formulas, and examples

About the Empirical Rule Calculator (68-95-99.7)

The empirical rule (also called the 68-95-99.7 rule) states that for normally distributed data, approximately 68% of values fall within one standard deviation of the mean, 95% within two standard deviations, and 99.7% within three. This simple rule lets you quickly estimate probabilities, identify unusual values, and set expectations without needing z-tables.

This calculator applies the empirical rule to any normal distribution. Enter a mean and standard deviation, and it computes the three key intervals, shows where any query value falls on the bell curve, and breaks the distribution into eight sections with exact probabilities. A color-coded bell curve visualization shows the 68/95/99.7 bands with your query value marked.

The comparison table between the empirical rule (which applies only to normal distributions) and Chebyshev's theorem (which applies to all distributions) shows why the empirical rule gives much tighter estimates — and when you can safely use it versus when you need the more conservative Chebyshev bounds.

When This Page Helps

The empirical rule is the most commonly taught concept in introductory statistics, yet students often struggle to apply it to specific contexts. This calculator bridges theory and application by computing exact intervals for any mean and standard deviation, visualizing the bell curve with color-coded bands, and showing where specific values fall.

The section probability breakdown goes beyond the basic 68/95/99.7 numbers to show exact percentages in each section of the curve — useful for computing more specific probabilities. The Chebyshev comparison table builds understanding of when the empirical rule applies and what to do when it doesn't.

How to Use the Inputs

Enter the mean (μ) and standard deviation (σ) of your normal distribution.
Use presets for common distributions: IQ scores, SAT, height, birth weight, blood pressure.
Enter a query value (x) to see its z-score and percentile on the bell curve.
Review the three empirical rule intervals showing 68%, 95%, and 99.7% ranges.
Examine the color-coded bell curve visualization with your query value marked.
Check the section probabilities table for exact percentages in each region.
Compare empirical rule results with Chebyshev's theorem in the reference table.

Formula used

68% Rule: P(μ − σ ≤ X ≤ μ + σ) ≈ 0.6827
95% Rule: P(μ − 2σ ≤ X ≤ μ + 2σ) ≈ 0.9545
99.7% Rule: P(μ − 3σ ≤ X ≤ μ + 3σ) ≈ 0.9973

Z-score: z = (x − μ) / σ
Percentile: P(X ≤ x) = Φ(z)

Section probability:
P(kσ to (k+1)σ) = Φ(k+1) − Φ(k)

Example Calculation

Result: 68%: 85-115, 95%: 70-130, 99.7%: 55-145, z(130) = 2.00, Percentile: 97.7%

For IQ scores (μ=100, σ=15), the middle 68% falls between 85 and 115. An IQ of 130 has z-score 2.00, placing it at the 97.7th percentile — right at the boundary of the 95% interval. Only about 2.3% of the population scores higher.

Tips & Best Practices

Remember: about 68% within 1 SD, 95% within 2 SDs, 99.7% within 3 SDs.
Each tail beyond 3σ contains only about 0.15% of the data — that's 3 in 2,000.
The sections between σ boundaries contain about 34%, 13.5%, 2.35%, and 0.15% each.
For quick percentile estimates: ±1σ excludes 16% per tail, ±2σ excludes 2.5% per tail.
Always verify normality before applying the empirical rule to real data.
Compare with Chebyshev: at 2σ, Chebyshev guarantees only 75% vs. the empirical 95%.

Origin and Significance

The empirical rule derives from the properties of the normal (Gaussian) distribution, discovered independently by Abraham de Moivre, Pierre-Simon Laplace, and Carl Friedrich Gauss. The rule' name reflects that it was verified empirically across many natural phenomena before the mathematical theory was fully developed. Heights, weights, test scores, measurement errors, and many biological variables follow approximate normal distributions.

The Sections of the Bell Curve

The normal curve can be divided into eight sections: the center (μ ± 0), two sections from μ to μ ± σ (about 34.13% each), two from μ ± σ to μ ± 2σ (about 13.59% each), two from μ ± 2σ to μ ± 3σ (about 2.14% each), and two tails beyond ± 3σ (about 0.13% each). These percentages sum to 100% and form the basis for all normal distribution calculations.

Applications in Quality Control

In manufacturing quality control, the empirical rule underlies Six Sigma methodology. A "six sigma" process has defect rates so low that the specification limits are 6 standard deviations from the process mean. Even 3σ control limits catch 99.7% of variation, making out-of-control signals extremely rare under normal operation. Walter Shewhart's control charts, the foundation of statistical process control, are direct applications of the empirical rule.

Sources & Methodology

Last updated: March 8, 2026

Frequently Asked Questions

Only when data is approximately normally distributed (bell-shaped and symmetric). Check with a histogram, Q-Q plot, or normality test first. For non-normal data, use Chebyshev's theorem instead, which gives weaker but universally valid bounds.