Quick Answer
To calculate a z-score from standard deviation, subtract the mean from the raw value, then divide by the standard deviation: z = (x - mean) / SD. The result tells you how many standard deviations the value sits above or below the mean.
TL;DR
Background: a student, analyst, or quality reviewer often has three numbers already in hand: one observed value, the mean, and the standard deviation. The practical question is not just "what is the z-score?" It is "is this value ordinary, worth explaining, or unusual enough to act on?" This guide answers that calculation question from the role of a senior statistician and data educator.
A z-score is a standardized distance from the mean. A standard deviation is the distance unit used for that standardization. A mean is the arithmetic center used as the reference point. If you still need to calculate the mean or SD first, use the mean and standard deviation calculator, sample standard deviation calculator, or population standard deviation calculator.
Formula
Z-score from standard deviation
| Symbol | Meaning | Example |
|---|---|---|
| x | The raw value you want to standardize | 90 quiz points |
| mean | The average of the reference group | 79.17 quiz points |
| SD | The standard deviation of the same reference group | 6.21 quiz points |
| z | The raw value expressed in SD units | 1.75 |
Use the same unit
Subtract the mean
Divide by the standard deviation
Interpret the sign and size
Worked Example
For a first-hand calculator check while drafting this article, we entered this 12-score quiz dataset: 68, 72, 74, 76, 77, 79, 80, 81, 83, 84, 86, 90. The sample mean is 79.17 and the sample standard deviation is 6.21. Now suppose the instructor wants to standardize the score 90.
Substitute the numbers
The score of 90 is about 1.75 sample standard deviations above the class mean. That is a strong result in this class, but it is not automatically an outlier. If the class scores are roughly bell-shaped, the empirical rule suggests that values within about two standard deviations are still common enough to interpret as high performance rather than a data problem.
What changes if the raw value is lower?
Use the z-score calculator when the mean and SD are already known. Use the descriptive statistics calculator when you want mean, variance, standard deviation, and z-score context from a full list of observations.
Sample or Population SD
Use the standard deviation that matches your reference group. If the listed values are a sample from a larger process, use sample standard deviation s. If the listed values are the complete fixed population you want to describe, use population standard deviation sigma. This choice changes the denominator in the SD calculation before the z-score is computed.
| Situation | Use this SD | Why it matters |
|---|---|---|
| One class section used to infer future sections | Sample SD | The class is a sample from a wider teaching process. |
| Every employee in one small team | Population SD | The list is the complete group being described. |
| Recent production parts used to judge future runs | Sample SD | The parts estimate ongoing process variation. |
| A fixed roster of final contest scores | Population SD | No larger group is being estimated. |
Decision criterion
Interpretation Bands
NIST describes the normal distribution as a central model for many measurement processes, and OpenStax introduces z-scores as the standard way to locate values on that model. The bands below are useful when the reference data are reasonably symmetric and the mean and standard deviation are meaningful summaries.
| Z-score range | Plain-language meaning | Typical action |
|---|---|---|
| -1 to +1 | Close to average | Usually treat as routine variation. |
| -2 to -1 or +1 to +2 | Noticeably low or high | Explain with context, especially for grades, labs, or process checks. |
| -3 to -2 or +2 to +3 | Unusual under a bell-shaped model | Review data quality, subgroup, timing, and practical consequences. |
| Less than -3 or greater than +3 | Very unusual under a normal model | Investigate before treating the value as routine. |
These are decision aids, not deletion rules. A z-score of 3.2 may be a sensor error, a special-cause event, or the most important observation in the dataset. If the decision is about unusual values, pair this method with the outlier detection guide and modified z-score article.
Decision Checklist
- The raw value, mean, and SD come from the same population, process, or comparison group.
- The standard deviation is not zero. If SD = 0, the z-score is undefined.
- The distribution is not extremely skewed or dominated by one outlier.
- The sample size is large enough that the mean and SD are stable for your purpose.
- You chose sample SD or population SD deliberately, not because a spreadsheet default picked it.
- You have a practical action rule before labeling a value as unusual.
For normal-probability work, continue with Standard Deviation and Normal Distribution. For broader interpretation of high and low spread, read How to Interpret Standard Deviation.
Common Mistakes
- Using mismatched groups:Do not compare a student score with the mean and SD from a different test unless that is the intended reference group.
- Treating z as a percent:A z-score of 1.75 is not 1.75%. It means 1.75 standard deviations above the mean.
- Ignoring direction:Positive and negative z-scores can have the same distance from average but opposite meanings.
- Assuming normality:Z-scores can standardize any numeric value, but normal-table probabilities require a suitable distribution shape.
- Deleting extreme values automatically:Investigate unusual z-scores before removing them. The value may be valid evidence rather than an error.
Weakest-section revision note
Further Reading
Sources
References and further authoritative reading used in preparing this article.