How to Calculate Z Scores, Normal Probabilities and Percentiles

 

A Z score expresses how far a value lies from a mean in standard-deviation units. It can help compare standardized values, but it does not automatically turn a dataset into a normal distribution. To interpret a normal tail probability or percentile, you must also decide whether a normal model is appropriate for the question.

The Z-Score Calculator: Normal Probability & Percentile accepts a value, mean and standard deviation, or a Z score entered directly. It calculates areas to the left, to the right or between two bounds under a standard normal model.

Calculate the standardized value

Use Z = (x - mean) / standard deviation. All three quantities must use compatible units. Subtract the mean first, then divide the difference by the standard deviation. The resulting Z score has no measurement unit.

For a value of 115, a mean of 100 and a standard deviation of 15, Z = (115 - 100) / 15 = 1. The value lies one standard deviation above the mean. A value of 85 gives Z = -1, meaning one standard deviation below the same mean.

A zero standard deviation is not valid for this calculation because it would require division by zero. A negative standard deviation is also invalid. Check which standard deviation your task specifies rather than switching between sample and population formulas without explanation.

Find a left-tail probability

The left-tail area is the probability that a standard normal variable is less than or equal to a selected Z score. At Z = 1, that area is approximately 0.8413447, or 84.1345%. At Z = 0, it is 50%, because the standard normal curve is symmetric around zero.

The left-tail percentage also gives the percentile rank under this normal model. Z = 1 corresponds to approximately the 84.13th percentile. That is not necessarily the empirical percentile in your observed dataset, where distribution shape, ties and the chosen ranking convention may matter.

Find a right-tail probability

The right-tail area represents the probability of a standard normal value at or above the selected Z score. For Z = 1, the result is approximately 15.8655%. The left and right areas add to 100%, apart from display rounding.

Because the normal distribution is continuous, including or excluding the exact endpoint makes no difference to these probabilities. The probability of one exact value is zero, while the probability of an interval can be positive.

Calculate probability between two bounds

For an interval, enter the lower bound first and the upper bound second. The probability is the area under the normal curve between them. From Z = -1 to Z = 1, the result is approximately 68.2689%.

In raw-value mode, both bounds use the same mean and standard deviation. With a mean of 100 and standard deviation of 15, values from 85 through 115 map to the same Z interval. Equal bounds produce zero interval probability, and reversed bounds are rejected rather than silently swapped.

Separate a probability from a statistical test

A normal-model tail area is not automatically a hypothesis-test p-value. A test requires a null model, a suitable statistic and an alternative hypothesis. A two-sided test may involve both tails, while a t statistic requires a different distribution and degrees of freedom.

Entering a sample standard deviation does not make this tool account for parameter uncertainty or switch to Student's t. Likewise, it does not test normality, construct a confidence interval or calculate a sample size. Use the model required by your analysis.

Read the curve and rounding carefully

The illustration shows Z from -4 to 4, while calculations support -12 to 12. A tiny tail outside the visible curve can still have a nonzero numerical probability. Results are displayed to four decimal places as percentages, with very small nonzero values labeled as less than 0.0001%.

For definitions, consult the NIST/SEMATECH e-Handbook normal-distribution reference. The calculator runs locally without saving or transmitting entries; the hosting website may separately use analytics under its privacy policy.