How Well Do the Angoff Design V Linear Equating Methods Compare With the Tucker and Levine Methods?

Abstract

Comparisons were made of three Angoff Design V equating methods and the Tucker and Levine Equally Reliable methods with respect to common item linear equating with non-equivalent populations. Forms of a professional certification test were equated with these five methods using (1) single-link equating of selected pairs of forms and (2) cyclical equating of selected forms to themselves, by means of equating chains. In the single-link equatings, raw score equivalents given by the Design V methods tended to fall between those obtained by use of the Tucker and Levine methods. The chain equatings produced similar estimated bias and estimated root mean squared error of score equiv alents for the five different methods.