The Use of Estimated True Scores,Standard Errors of Measurement,and Confidence Intervals in Reporting Assessment Data

Abstract

Assessment data play an important role in educational decision making. All too often we act as if we have total confidence in the test scores we interpret, though we know that error of measurement is always present. Three procedures that help to determine the degree of confidence we can place in any raw score are discussed in this article (a) computing the standard error of measurement, (b) computing the estimated true score, and (c) constructiong conference intervales. These three procedures are easy to and require only mathematical skills individuals. The process of pupil assessment should these three simple procedures.