Equating error is usually small relative to the magnitude of measurement error, but it could be one of the major sources of error contributing to mean scores of large groups in educational measurement, such as the year-to-year state mean score fluctuations. Though testing programs may routinely calculate the standard error of equating (SEE), the…

In this study we examined variations of the nonequivalent groups equating design for tests containing both multiple-choice (MC) and constructed-response (CR) items to determine which design was most effective in producing equivalent scores across the two tests to be equated. Using data from a large-scale exam, this study investigated the use of…

In four articles adapted from the Educational Testing Service (ETS) Research Report Series, Issue 2 of ETS Research Spotlight provides a small taste of the range of assessment-related research capabilities of the ETS Research and Development Division. Those articles cover assessment-related research aimed at developing models of student learning,…

Several current international comparative large-scale assessments of educational achievement (ICLSA) make use of "Rasch models", to address functions essential for valid cross-cultural comparisons. From a historical perspective, ICLSA and Georg Rasch's "models for measurement" emerged at about the same time, half a century ago. However, the…

"Educational Measurement" has been the bible in its field since the first edition was published by ACE in 1951. The importance of this fourth edition of "Educational Measurement" is to extensively update and extend the topics treated in the previous three editions. As such, the fourth edition documents progress in the field and…