Equating Subscores Using Total Scaled Scores as an Anchor. Research Report. ETS RR-11-07

Because the demand for subscores is ever increasing, this study examined two different approaches for equating subscores: (a) equating a subscore on the new form to the same subscore in the old form using internal common items as the anchor to conduct the equating, and (b) equating a subscore on the new form to the same subscore in the old form using equated total scores as the anchor to conduct the equating. Equated total scores can be used as an anchor to equate the subscores because the total equated scores are comparable across both the new and the old forms. Data from 2 tests (Tests X and Y) were used to conduct the study, and results showed that when the number of internal common items was large (approximately 50% of the total subscore), then using common items to equate the subscores was preferable. However, when the number of common items was small (approximately 25% of the total subscore, which is common practice), then using total scaled scores (TSS) to equate the subscores was preferable. Using raw subscores (not equating) resulted in a considerable amount of bias for both tests. Indices to measure equating variability and bias are appended. (Contains 4 tables, 5 figures and 4 notes.)