Skip to content
Omnikalc

Standard error of measurement calculator

Enter the standard deviation of the test, its reliability, an observed score, the mean of the group and the confidence level to get the standard error of measurement, a score band and the Kelley estimate of the true score.

Updated on
Your inputs

Value out of range

Value out of range

Value out of range

Value out of range

Results

Standard error of measurement

4.50

Lower limit of the band around the observed score

99.2

Upper limit of the band around the observed score

116.8

Estimated true score (Kelley)
107.3
Standard error of estimation
4.29
Lower limit around the estimated true score
98.9
Upper limit around the estimated true score
115.7
Standard error of the difference between two scores
6.36

The standard error of measurement calculator gives the SEM of a test from its standard deviation and reliability, a confidence band around an observed score, the Kelley estimate of the true score with its band, and the standard error of the difference between two scores.

How the calculation works

  • Enter the standard deviation of the test and its reliability.
  • Enter the observed score, the mean of the group and the confidence level.
  • Read the SEM, the bands and the estimated true score.

Example

A test with a standard deviation of 15 and a reliability of 0.91 has an SEM of 4.50. A score of 108 has a band from 99.2 to 116.8. With a group mean of 100, the estimated true score is 107.3, with a band from 98.9 to 115.7 and a standard error of estimation of 4.29. The standard error of the difference between two scores is 6.36.

Why it matters

The calculator helps psychologists, teachers, clinicians and test users who report a score with its uncertainty, decide whether a score is above a cut line, or judge whether two scores really differ. It converts a reliability coefficient into a margin in score units.

Good to know

Use the reliability estimate that fits the use: internal consistency, test-retest or an alternate form give different SEMs. The SEM assumes the same error for all score levels, which is only approximately true. Report the confidence level with the band. A band around the observed score and a band around the estimated true score answer different questions, and the first is the usual one in test manuals. Do not use this for a single item.

The Cronbach alpha calculator gives a reliability estimate, the Kuder-Richardson calculator does the same for right-wrong items, the reliable change index calculator tests a change and the Spearman-Brown calculator predicts the effect of test length.

Frequently asked questions

What is the standard error of measurement?

The standard error of measurement, SEM, is the standard deviation of the errors of measurement of a test. It tells how much a person’s observed score would vary around the true score on repeated testing. It is SEM = SD × √(1 − r), where SD is the standard deviation of the scores and r the reliability.

How is the confidence band used?

Add and subtract a multiple of the SEM to an observed score: ±1.96 SEM gives a 95 percent band. A score of 108 with an SEM of 4.5 has a band of about 99 to 117. The band shows that two close scores cannot be taken as different, and that a cut score has a zone of uncertainty around it.

What is the Kelley estimate of the true score?

Because of regression to the mean, the best estimate of the true score is pulled toward the group mean: T = r × X + (1 − r) × M. The more reliable the test, the closer it is to the observed score. Its standard error is SD × √(r × (1 − r)), which is smaller than the SEM, and the band is centered on the estimate.

How does it relate to reliability?

For a given SD, a more reliable test has a smaller SEM. A reliability of 0.91 on a scale with SD 15 gives an SEM of 4.5, while a reliability of 0.75 gives 7.5. The SEM is in the unit of the score, so it is easier to interpret than the reliability coefficient, which depends on the group.

Was this calculator helpful?

Thanks for your feedback!

Results are estimates provided for information purposes. They do not replace professional advice.