Because some amount of error is expected with any measurement, statisticians developed the term Standard Error of Measurement (SEM) to account for small amounts of error in every result. (fastbridge.org)
The test has a standard error of measurement of about 2.5 points, so Cherry's true IQ could have been below 70. (newscientist.com)
Furthermore, they say, a test's standard error of measurement may be large enough to throw into question the use of the results. (pbs.org)