A Review of Testlet Response Theory and Its Applications

Abstract

During the past few years, there has been a surge in utilizing Bayesian inference. An ERIC search of the keywords Bayesian statistics reveals that 48 journal articles using Bayesian models were published between 1995 and 2000, and this number more than doubles when articles from 2001 to 2006 are consid ered. Such interest is well deserved, as recent advances in computer technology and sampling methodology have allowed researchers to demonstrate the poten tial of Bayesian models for solving practical problems in many areas of study. In psychometrics, the potential of Bayesian models is clearly shown in Testlet Response Theory and Its Applications, the new book by Wainer, Bradlow, and Wang. It is well established in the psychometric literature that lack of conditional independence among subsets of test items, known as testlets, may result in underestimated standard errors and thus overestimated measurement precision when conditional independence is assumed (e.g., Sireci, Thissen, & Wainer, 1991; Wainer, Bradlow, & Du, 2000). Therefore, it is not surprising that a num ber of strategies have been proposed for modeling these dependencies. One such approach is testlet response theory (TRT), which models the dependency because of testlets by adding a parameter to the familiar item response theory (IRT) models (Bradlow, Wainer, & Wang, 1999). It is on this model that led to the development of Testlet Response Theory and Its Applications, but the table of contents makes it clear that this book offers considerably more than the output of this single program of research. The six chapters comprising part I, for instance, could serve as a fine introductory book on test theory. These chap ters, written in accessible language, introduce true score theory, IRT, and TRT; connect the ideas behind these theories; and present comprehensive and thought ful examples. The authors assert that their intended audience includes research ers and professionals in statistics, psychometrics, and educational psychology; clearly, students could be added to this list. There are many insightful discus sions on various topics throughout the book that both researchers and students would find stimulating, but perhaps the most delightful occurs in chapter 3, which contains a three-page obiter dictum on competing interpretations of the