ABSTRACT
This chapter addresses the logic of validation for uses and interpretations of such derived scores. The derived score, which can be seen as an extension of the original score that measures an application construct, should require a second round of validation, looping through scoring, generalization, extrapolation, and use a second time. Validation would begin with the test developer’s investigation of achievement tests used to measure reading and mathematics achievement on continuous scales. Then, a second looping through Kane’s four stages would address the derived scores created when scale scores are mapped into “Below Basic,” “Basic,” “Proficient,” and “Advanced” according to cut scores defined by a judgmental standard setting process. Teacher effectiveness estimates from value-added models may be regarded as derived scores created by a complex measurement procedure to measure an application construct far removed from individual student scores obtained at a single point in time.
