Z-Score Calculator
Calculate how far an observed value lies above or below a reference mean in standard-deviation units. Explore normal-model percentiles and tail probabilities, see the calculation steps, and understand when these estimates are appropriate.
Z-Score Calculator
Enter an observed value, the reference mean and a positive standard deviation. Use the same measurement units for all three values. Results remain blank until you calculate.
Calculation Results
Step-by-Step Calculation
What Is a Z-Score?
A z-score, also called a standard score, describes the position of a value relative to a reference mean. It expresses the difference in units of standard deviation rather than the original measurement units.
For example, a test score of 85 has a different meaning in a group whose mean is 70 than in a group whose mean is 90. The standard deviation also matters: a difference of 10 points may be large in a tightly clustered group and relatively small in a group with wide variation.
A z-score helps put this relationship on a common scale. A positive score is above the mean, a negative score is below it, and zero indicates that the observed value equals the mean.
Z-Score Formula
z = (x − μ) / σ
Where:
x = observed value
μ = reference mean
σ = positive standard deviationFirst subtract the mean from the observed value. This gives the signed difference from the center of the reference distribution. Then divide that difference by the standard deviation to express it in standard-deviation units.
If the mean and standard deviation are estimated from a sample, they are commonly represented by x̄ and s. The same arithmetic produces a standardized score, but inference about the population requires additional assumptions and an appropriate sampling method.
How to Use This Calculator
1. Enter the observed value. This is the score or measurement you want to examine.
2. Enter the reference mean. Choose a mean that belongs to the relevant group and measurement scale.
3. Enter standard deviation. It must be positive. A standard deviation of zero makes the formula undefined.
4. Calculate and interpret. Read the z-score first. Use the percentile and tail probabilities only when a normal model is a reasonable description of the data.
Worked Example: Standardizing a Test Score
Suppose a student receives a score of 85. The reference group has a mean of 70 and a standard deviation of 10.
Step 1 — Identify the values: x = 85, μ = 70, σ = 10.
Step 2 — Subtract the mean: 85 − 70 = 15.
Step 3 — Divide by standard deviation: z = 15 / 10 = 1.5.
Interpretation: The score is 1.5 standard deviations above the reference mean. If the scores follow a normal distribution with this mean and standard deviation, the estimated cumulative percentile is approximately 93.32%.
The percentile is not necessarily the student’s actual class rank. The exact empirical percentile depends on the observed scores in the group. The normal-model percentile is an estimate based on the assumed distribution.
Standard Normal Reference Table
The following values show approximate cumulative probability for selected z-scores under a standard normal distribution. Cumulative probability is the area to the left of the z-score.
| Z-score | Approximate percentile | Area above | Interpretation |
|---|---|---|---|
| -2.00 | 2.28% | 97.72% | Two standard deviations below the mean |
| -1.00 | 15.87% | 84.13% | One standard deviation below the mean |
| -0.50 | 30.85% | 69.15% | Half a standard deviation below the mean |
| 0.00 | 50.00% | 50.00% | At the mean |
| 0.50 | 69.15% | 30.85% | Half a standard deviation above the mean |
| 1.00 | 84.13% | 15.87% | One standard deviation above the mean |
| 1.50 | 93.32% | 6.68% | One and a half standard deviations above the mean |
| 2.00 | 97.72% | 2.28% | Two standard deviations above the mean |
| 3.00 | 99.87% | 0.13% | Three standard deviations above the mean |
These figures are rounded normal-model probabilities. They should not be interpreted as guaranteed percentages in every real-world dataset.
Understanding Positive, Negative and Zero Z-Scores
Positive z-score
A positive z-score means the observation is above the reference mean. A score of +1.5 indicates that the observation is one and a half standard deviations above the mean. Whether that is desirable depends on the context: a high examination score may be favorable, while a high defect count may not be.
Negative z-score
A negative z-score means the observation is below the mean. A score of −2 means the observation is two standard deviations below the reference mean. It does not automatically mean that the observation is poor or abnormal.
Z-score of zero
A score of zero means the observed value equals the mean. In a continuous normal distribution, the mean divides the theoretical distribution into equal areas of 50% below and 50% above.
Z-Score, Percentile and Probability: What Is the Difference?
These measures are related, but they answer different questions.
- Z-score: How many standard deviations is the value above or below the mean?
- Percentile: What percentage of the reference distribution lies at or below this value?
- Upper-tail probability: What proportion of the modeled distribution lies above the value?
Under a continuous normal model, the cumulative percentile and upper-tail probability add to 100%. The calculator estimates these areas from the normal distribution. The actual percentage of observations in a dataset can differ, particularly when the data are skewed or contain multiple clusters.
When Is the Normal Distribution Assumption Appropriate?
The z-score formula can be calculated for any value with a meaningful reference mean and positive standard deviation. Normality is not required for the arithmetic itself.
However, translating a z-score into a normal percentile or probability requires a suitable normal-distribution model. This model may be reasonable for some measurement processes, but it should not be assumed for every dataset.
Strong skewness, heavy tails, multiple peaks, truncation, dependence between observations or unusual sampling can make normal-model probabilities misleading. When raw observations are available, compare the theoretical result with a histogram, a normal probability plot or an empirical percentile calculation.
Population Standard Deviation vs. Sample Standard Deviation
The choice of standard deviation depends on the question and the data available. Population standard deviation describes the spread of an entire defined population. Sample standard deviation estimates spread from a sample, commonly using a denominator of n − 1 rather than n.
Use a standard deviation that matches the reference mean and the analysis you are performing. Mixing a mean from one group with a standard deviation from another can produce a numerical result that has little practical meaning.
For inferential tasks, such as estimating a population mean from a sample, a z-score alone may not be the appropriate statistic. Depending on the design and assumptions, a t-statistic or another method may be required.
Using Z-Scores to Compare Different Measurements
Standardization can help compare relative positions across measurements that use different units. For example, a student’s position in mathematics and science can be compared using scores standardized against the appropriate subject-specific groups.
This comparison is only meaningful when the reference populations, score definitions and measurement conditions are suitable. Equal z-scores do not mean that two raw measurements are equal, and they do not establish that two measurements have the same practical importance.
A standardized score also does not correct for biased sampling, unreliable measurement, inappropriate reference groups or confounding factors.
Outliers and Unusual Observations
A large absolute z-score can flag an observation for further investigation. For example, a value several standard deviations from the mean may be unusual under a normal model. But it is not automatically a data-entry error or a confirmed outlier.
Outlier assessment depends on the distribution, sample size, study design and purpose of the analysis. A rule based on a fixed z-score threshold may be unsuitable for skewed distributions or small datasets.
Before removing or changing an observation, verify the original measurement, investigate possible causes and document the decision. Do not delete data solely because a standardized score is large.
Common Z-Score Calculation Mistakes
- Mixing measurement units: Convert the value, mean and standard deviation to compatible units before calculating.
- Using an unrelated reference group: The mean and standard deviation must describe the group against which the value is being compared.
- Using zero standard deviation: The formula is undefined when the denominator is zero.
- Ignoring the sign: Negative scores indicate values below the mean; the sign is meaningful.
- Confusing score and percentile: A z-score is a standardized distance, not a percentage rank.
- Assuming normality without checking: The score can be computed without normality, but normal-model probability estimates may be unsuitable.
- Overinterpreting a large score: An unusual value requires context and verification rather than automatic removal.
- Relying only on rounded output: Use unrounded values for subsequent calculations when precision matters.
Calculation Methodology and Limitations
The calculator uses the entered value, mean and standard deviation to compute the z-score directly. It then uses an approximation to the standard normal cumulative distribution function to estimate the percentile and tail probabilities. Display rounding does not alter the underlying calculation.
The calculator does not estimate the mean or standard deviation from raw data, test distributional assumptions, calculate an empirical percentile or perform a hypothesis test. The user is responsible for selecting appropriate reference statistics.
The probability calculation is a numerical approximation suitable for general educational use. For high-stakes research or extreme tail probabilities, verify results with a validated statistical package and document the numerical method.
Frequently Asked Questions
1. What is a z-score?
A z-score expresses how many standard deviations an observed value lies above or below a reference mean.
2. What is the z-score formula?
The formula is z = (x − μ) / σ. Subtract the mean from the observed value and divide by the positive standard deviation.
3. Can a z-score be negative?
Yes. A negative score means the observation is below the reference mean.
4. What does a z-score of zero mean?
It means the observed value equals the reference mean.
5. Does calculating a z-score require normal data?
No. The formula itself does not require normality. The normal-model percentile and probability estimates do require an appropriate normal-distribution assumption.
6. What does the percentile result mean?
Under the normal model, it estimates the percentage of values at or below the entered observation.
7. Why must standard deviation be positive?
It is the denominator in the formula. Zero makes the formula undefined, and a standard deviation cannot be negative.
8. What does a z-score of 2 mean?
The value is two standard deviations above the mean. Under a standard normal model, the cumulative percentile is approximately 97.72%.
9. Is a z-score the same as a percentile?
No. A z-score measures standardized distance, while a percentile describes relative position in a distribution.
10. Does this calculator determine statistical significance?
No. Statistical significance requires an appropriate statistical test, hypotheses, sampling design and assumptions. A z-score alone is not sufficient.
Statistical References
These references provide further reading on statistical distributions, standardization, probability and data analysis.
NIST/SEMATECH e-Handbook of Statistical Methods
Reference material for statistical methods, probability distributions and data analysis.
OpenStax — Introductory Statistics
An educational textbook covering descriptive statistics, probability distributions and inferential methods.
NIST — Normal Probability Plot
Guidance on examining whether observed data are reasonably consistent with a normal distribution.
Penn State — Statistics Learning Resources
Statistics learning materials for probability, distributions and statistical inference.
Related Statistics Calculators
Continue with a related statistical task using an existing FreeCalz tool.
Disclaimer
This calculator is provided for educational and general informational purposes. Normal-distribution percentiles and probabilities are model-based estimates, not guarantees about observed data. Verify important analyses with appropriate statistical methods, validated software and qualified guidance where necessary.
