Z-Score Calculator

Three z-score calculation tools

Standardize a value

Find z from a score, mean, and spread

Also returns the normal probabilities tied to the solved z-score.

Z = (x − μ) / σ

Convert z and probability

Start from any one known area

Enter any one value; the other five values are calculated.

p ⇄ Z

Probability between two z-scores

Measure the area inside two bounds

The curve and text report the same interval.

Φ(z₂) − Φ(z₁)

Shaded: P(-1 < Z < 0)

Probability calculations use the standard normal model. In the converter, x is the standard-normal value and Z is its cutoff. The calculations do not test whether source data are normally distributed. Numerical definitions and check values follow the NIST standard normal table.

Standardized position

Express one observation as a signed distance from its reference mean

A z-score removes the original unit by dividing a deviation by a positive reference standard deviation. It enables within-scale comparison but does not automatically create a normal probability.

Step-by-step guide · Standardized position

Standardize an exam score against its declared cohort

Standardization: z = (x−μ)/σ
Center position: x=μ gives z=0
Scale requirement: Standard deviation > 0

Identify the reference group
Confirm that 70 and 8 describe the same exam, cohort, scale, and population or sample convention relevant to the score. Mixing references invalidates the quotient.
Find the signed deviation
Subtract the mean: 82−70=12 points. The positive sign places the score above the reference center.
Divide by spread
Twelve divided by eight gives z=1.5. The score lies one and a half declared standard deviations above the mean.
Report inputs with the result
Write z=1.5 using mean 70 and standard deviation 8 for the specified cohort. A bare standardized number cannot be audited or interpreted across a changed reference group.

Express one observation as a signed distance from its reference mean: A score of 82 in a cohort with mean 70 and standard deviation 8 has a straightforward standardized position.

Standardize an exam score against its declared cohort

A score of 82 in a cohort with mean 70 and standard deviation 8 has a straightforward standardized position.

Identify the reference group
Confirm that 70 and 8 describe the same exam, cohort, scale, and population or sample convention relevant to the score. Mixing references invalidates the quotient.
Find the signed deviation
Subtract the mean: 82−70=12 points. The positive sign places the score above the reference center.
Divide by spread
Twelve divided by eight gives z=1.5. The score lies one and a half declared standard deviations above the mean.
Report inputs with the result
Write z=1.5 using mean 70 and standard deviation 8 for the specified cohort. A bare standardized number cannot be audited or interpreted across a changed reference group.

Build the reference statistics before standardizing

If a supplied data list has an audited total and count, use the fraction calculator to preserve their quotient exactly before decimal reporting. Use the standard deviation calculator to decide whether the values represent the complete population or a sample estimate. The z-score page accepts the declared results; it does not infer their provenance or recompute them from raw records.

For a broader check of median, modes, extrema, range, and frequency, the descriptive statistics calculator helps reveal skewness or clusters that a mean and standard deviation compress. Standardization remains arithmetically possible for many shapes, but distribution shape governs what further claims are sensible.

Separate relative position from normal-tail area

Always encoded by z

The sign indicates which side of the mean contains x, and the magnitude counts declared standard-deviation units. This interpretation follows directly from the formula.

Requires a normal model

Converting z into a percentile or tail probability uses a distribution, usually the standard normal. That step requires a justified model and is not guaranteed by calculating z.

Requires inferential design

Calling a score unusual, significant, or evidence against a hypothesis can require sampling assumptions, multiple-comparison rules, and a decision threshold beyond this descriptive transformation.

Guard against a meaningless standardized quotient

Standard deviation must be strictly positive. When every reference value is identical, division by zero makes a z-score undefined rather than infinitely informative.
x, mean, and standard deviation must use the same unit and scale. A Celsius score cannot be standardized against a Fahrenheit mean and spread without coherent conversion.
Population parameters μ and σ and sample statistics x̄ and s are not interchangeable labels. Record which reference was used, especially when comparing reports from different studies.
A negative z-score is not inherently bad and a positive one is not inherently good. Direction depends on what the measured quantity means; lower time can be desirable while lower quality may not be.

Important limitations

Probability results assume a standard normal model; this page does not test whether source data are normal. In the converter, x denotes the standard-normal value and Z is its cutoff. Symmetric central, mean-to-Z, and two-tail probabilities determine |Z| but cannot determine its original sign. Probabilities are not accepted as percentages.

Review the stated scope, calculation method, assumptions, supported inputs, and effective dates, and obtain qualified advice before using a high-stakes result.

This formula passed its registered engineering tests and is published under an explicit site-owner policy authorization. No independent expert, QA, editorial, legal, or third-party review is claimed. Verify consequential results with a qualified professional.

Standardized position

Express one observation as a signed distance from its reference mean

A z-score removes the original unit by dividing a deviation by a positive reference standard deviation. It enables within-scale comparison but does not automatically create a normal probability.

Step-by-step guide · Standardized position

Standardize an exam score against its declared cohort

Standardization: z = (x−μ)/σ
Center position: x=μ gives z=0
Scale requirement: Standard deviation > 0

Identify the reference group
Confirm that 70 and 8 describe the same exam, cohort, scale, and population or sample convention relevant to the score. Mixing references invalidates the quotient.
Find the signed deviation
Subtract the mean: 82−70=12 points. The positive sign places the score above the reference center.
Divide by spread
Twelve divided by eight gives z=1.5. The score lies one and a half declared standard deviations above the mean.
Report inputs with the result
Write z=1.5 using mean 70 and standard deviation 8 for the specified cohort. A bare standardized number cannot be audited or interpreted across a changed reference group.

Express one observation as a signed distance from its reference mean: A score of 82 in a cohort with mean 70 and standard deviation 8 has a straightforward standardized position.

Standardize an exam score against its declared cohort

A score of 82 in a cohort with mean 70 and standard deviation 8 has a straightforward standardized position.

Identify the reference group
Confirm that 70 and 8 describe the same exam, cohort, scale, and population or sample convention relevant to the score. Mixing references invalidates the quotient.
Find the signed deviation
Subtract the mean: 82−70=12 points. The positive sign places the score above the reference center.
Divide by spread
Twelve divided by eight gives z=1.5. The score lies one and a half declared standard deviations above the mean.
Report inputs with the result
Write z=1.5 using mean 70 and standard deviation 8 for the specified cohort. A bare standardized number cannot be audited or interpreted across a changed reference group.

Build the reference statistics before standardizing

Separate relative position from normal-tail area

Always encoded by z

The sign indicates which side of the mean contains x, and the magnitude counts declared standard-deviation units. This interpretation follows directly from the formula.

Requires a normal model

Converting z into a percentile or tail probability uses a distribution, usually the standard normal. That step requires a justified model and is not guaranteed by calculating z.

Requires inferential design

Calling a score unusual, significant, or evidence against a hypothesis can require sampling assumptions, multiple-comparison rules, and a decision threshold beyond this descriptive transformation.

Guard against a meaningless standardized quotient

Standard deviation must be strictly positive. When every reference value is identical, division by zero makes a z-score undefined rather than infinitely informative.
x, mean, and standard deviation must use the same unit and scale. A Celsius score cannot be standardized against a Fahrenheit mean and spread without coherent conversion.
Population parameters μ and σ and sample statistics x̄ and s are not interchangeable labels. Record which reference was used, especially when comparing reports from different studies.
A negative z-score is not inherently bad and a positive one is not inherently good. Direction depends on what the measured quantity means; lower time can be desirable while lower quality may not be.

Three z-score calculation tools

Identify the reference group

Find the signed deviation

Divide by spread

Report inputs with the result

Standardize an exam score against its declared cohort

Build the reference statistics before standardizing

Separate relative position from normal-tail area

Always encoded by z

Requires a normal model

Requires inferential design

Guard against a meaningless standardized quotient

Important limitations

Three z-score calculation tools

Identify the reference group

Find the signed deviation

Divide by spread

Report inputs with the result

Standardize an exam score against its declared cohort

Build the reference statistics before standardizing

Separate relative position from normal-tail area

Always encoded by z

Requires a normal model

Requires inferential design

Guard against a meaningless standardized quotient