Three Sigma Rule in Statistics: Formula, Example, and Limits

Quick Answer

The three sigma rule says that, for an approximately normal distribution, almost all routine values should fall within three standard deviations of the mean. In formula form, the usual band is mean - 3s to mean + 3s for sample data, or mu - 3sigma to mu + 3sigma for a known population.

Under a normal model, about 99.73% of observations fall inside that band and about 0.27% fall outside it. That does not make every outside value an error. It makes the value unusual enough to investigate.

Use the Standard Deviation Calculator, Sample Standard Deviation Calculator, or Descriptive Statistics Calculator to compute the mean and standard deviation first. Use the Z-Score Calculator when you already know the mean, standard deviation, and value being checked.

Scenario: Applying the Rule

A student in a statistics course or an analyst on a quality team often has the same problem: the standard deviation has been calculated, but the next question is whether a new observation is routine variation or a signal.

As a statistics educator, I frame the three sigma rule as a screening rule, not a verdict. The objective is to create a defensible review band for a specific process: measurements inside the band are treated as expected variation, while measurements outside the band are checked for data entry mistakes, special causes, instrument drift, or a real process shift.

Formula

Here, x is the observation being checked, xbar is the sample mean, and s is the sample standard deviation. A z-score of 3.0 means the observation is three sample standard deviations above the mean. A z-score of -3.0 means it is three sample standard deviations below the mean.

Worked Example: Fill Weights

A packaging analyst checks whether new bottle fill weights are consistent with a stable line. The recent baseline sample, in grams, is: `499.4, 499.6, 499.8, 499.9, 500.0, 500.1, 500.2, 500.3, 500.4, 500.5, 500.7, 500.9`.

For those 12 baseline fills, the sample mean is 500.150 g and the sample standard deviation is 0.442 g. The three sigma limits are `500.150 - 3(0.442) = 498.824 g` and `500.150 + 3(0.442) = 501.476 g`.

The 501.7 g bottle is above the upper limit by 0.224 g. That is a statistical signal, but the next action is operational: check whether the filler nozzle was adjusted, whether the scale was calibrated, whether the product temperature changed, and whether nearby bottles show the same pattern.

Probability Table

The three sigma rule is a normal-distribution rule. The NIST normal-distribution reference gives the standard normal model behind these percentages. The exact values below are useful when deciding whether you need a quick screen or a formal probability calculation.

For a one-sided question, split the outside probability across the two tails. Under the normal model, values above mean + 3s account for about 0.135% of observations, and values below mean - 3s account for about 0.135%.

Decision Criteria

• Use it:The baseline process is stable, the distribution is roughly bell-shaped, and the cost of reviewing a small number of signals is acceptable.
• Use control charts:The data arrive over time and you need to distinguish routine variation from special-cause variation. Start with [Control Charts and Process Control](/learn/control-charts).
• Use a z-score:You need to report how far a single value is from the mean in standard deviation units. See [Z-Score Explained](/learn/z-score-explained).
• Use a robust method:The data are skewed, heavy-tailed, or already include extreme values. Compare [Robust Statistics](/learn/robust-statistics) and [Modified Z-Score Outlier Detection](/learn/modified-z-score-outlier-detection).
• Use a tolerance limit:A customer, regulator, or engineering specification already defines the acceptable range. A three sigma signal can support investigation, but it does not replace the requirement.

Checklist

• Is the baseline data from the same process, instrument, population, and time period as the value being checked?
• Was the suspicious value kept out of the baseline mean and standard deviation calculation?
• Does a histogram or normal probability plot look roughly normal?
• Are there enough observations for the standard deviation to be stable?
• Is the decision to investigate, quarantine, delete, or report the value defined before seeing the result?
• Would a domain limit, control chart, or robust statistic answer the practical question better?

Weakest Section Rewrite

Weak version: "The three sigma rule can identify unusual values."

Concrete substitution: "Use the three sigma rule when a stable fill-weight process has a baseline mean of 500.150 g and a sample standard deviation of 0.442 g. The review band is 498.824 g to 501.476 g, so a new 501.7 g bottle is not automatically bad product, but it is far enough from the baseline to check the scale, filler setting, and adjacent bottles."