Models for uncertainty in educational testing

The aim of this monograph is to present a survey of modern statistical techniques applied to educational and psychological testing. A theme that runs through this book is that of making inferences about sources of variation or uncertainty. The author shows how information about these sources can be used for improved (shrinkage) estimation of certain elementary quantities. Among the topics covered are essay rating, summarizing item-level properties, test equating, small-area estimation, and longitudinal analysis with missing data. Throughout, examples are given using real data sets that exemplify these applications.