The standard error of measure indicates the amount of uncertainty that a sample (such as a normative sample) is truly representative of the general population. In the case of administering standardized tests, it conveys the level of uncertainty that a single test performance observed by the evaluator represents how the child would do if it were administered multiple times.
The standard error of measure is used to determine the confidence interval. It is half the indicated confidence interval. A high standard error of measure implies a lower reliability in test scores. It represents how confident you can be that the scores on a test are accurate. It is always important to remember, however, that a test score may be reliable and even accurate but if it is not valid then the score is useless.
Additionally, the standard error of measure often demonstrates how little information is gained from scoring a standardized test, assuming it is valid, reliable and free of linguistic, cultural or SES bias. A child’s standard score may qualify him or her for special education services. However, with the confidence interval included, that decision often can not be determined. For example, you are evaluating a 3 yr 2 month old girl with the PLS-5 English. She receives a raw score of 30, which converts to a standard score of 76. With a mean of 100 and a standard deviation of 15, this score is between 1.5 and 2 SD below the mean. To be 90% sure that her true score is reflected in your results, her confidence interval would range from 71 to 85 SS, or moderately delayed to within normal limits.
Example from Sept 25-27, 2011. Survey of Republican Primary: Margin of Error +/- 3 points
Mitt Romney- 23%; Rick Perry- 19; Herman Cain-17; Newt Gingrich- 11; Ron Paul- 6; Jon Huntsman- 4; Michele Bachmann- 3; Rick Santorum-3
The way the candidates are listed, Mitt Romney appears to be in the lead. However, if the margin of error is taken into account, the top three candidates are essentially tied. Mitt Romney could be as low as 20%, Rick Perry could be as high as 22%, and Herman Cain could be as high as 20%. Due to inherent errors in measurement, we can not conclusively say that Mitt Romney is in the lead. This is why it is always important to include the confidence interval in any score presented.
Resources
NYCDOE Initial Guidance Document for Speech and Language Evaluators
