 |
GetTextbooks.co.uk
Compare Prices & Save up to 90% |
|
|
|
| Login | Sign up | My Wish List |
 | | Test Equating: Methods and Practices (Springer Series in Statistics) by Michael J. Kolen, Robert L. Brennan ISBN-10: 9780387944869 ISBN-10: 0-387-94486-9 ISBN-13: 9780387944869 ISBN-13: 978-0-387-94486-9 Hardcover 1999-08-13 Springer Find Lowest Price |
Editorials | ||
Product Description In recent years, many researchers in the psychology and statistical communities have paid increasing attention to test equating as issues of using multiple test forms have arisen and in response to criticisms of traditional testing techniques. This book provides a practically oriented introduction to test equating which both discusses the most frequently used equating methodologies and covers many of the practical issues involved. The main themes are: - the purpose of equating - distinguishing between equating and related methodologies - the importance of test equating to test development and quality control - the differences between equating properties, equating designs, and equating methods - equating error, and the underlying statistical assumptions for equating. The authors are acknowledged experts in the field, and the book is based on numerous courses and seminars they have presented. As a result, educators, psychometricians, professionals in measurement, statisticians, and students coming to the subject for the first time as part of their graduate study will find this an invaluable text and reference. | ||
Reviews | ||
Important for educational testing The Educational Testing Service and other institutions have for over 50 years done important research in testing methods,designs and analyses for a variety of purposes including (1) determining eligibility for university enrollment (2) competency for specific jobs and (3) trend information for public policy. Many of the advances have been described in the psychometric literature and the techniques described in this book are used to compare scores on different tests such as SAT tests taken at different times. On an elementary level most university professors have done a form of test equating with their own exams since the appropriate difficulty is sometimes hard to gauge. So-called curving a test is a form of adjusting scores by transforming to the standard normal curve. Unfortunately there are many better techniques that have been ignored and many psychometricians are not properly trained. These authors are among teh leaders in the field and have taught test equating. They found a need for a book of this type and they filled it. Many prominent statisticians have contributed to this research including Paul Holland, Don Rubin and Henry Braun. Much of this work was done by statisticians during their tenure at ETS and research continues at the ETS and at academic institutions. The authors provide a thorough treatment of the methods and the literature. This is good for a graduate level course, for a reference guide for practitioners and researchers in test equating. One important aspect of the problem is estimating the standard error of the test equating estimates. The delta method and parametric and nonparametric bootstrap methods are all described. An entire chapter is devote to item response theory and its application to test equating. The final chapter provides a very practical discussion of issues in the implementation of test equating methods and includes discussion of the choice of statistical procedures, determining the sample size, data collection, test specification and design, quality control and calibration. | ||
nice introduction to the details of test equating The Educational Testing Service and other institutions have for over 50 years done important research in testing methods,designs and analyses for a variety of purposes including (1) determining eligibility for university enrollment (2) competency for specific jobs and (3) trend information for public policy. Many of the advances have been described in the psychometric literature and the techniques described in this book are used to compare scores on different tests such as SAT tests taken at different times. On an elementary level most university professors have done a form of test equating with their own exams since the appropriate difficulty is sometimes hard to gauge. So-called curving a test is a form of adjusting scores by transforming to the standard normal curve. Unfortunately there are many better techniques that have been ignored and many psychometricians are not properly trained. These authors are among teh leaders in the field and have taught test equating. They found a need for a book of this type and they filled it. Many prominent statisticians have contributed to this research including Paul Holland, Don Rubin and Henry Braun. Much of this work was done by statisticians during their tenure at ETS and research continues at the ETS and at academic institutions. The authors provide a thorough treatment of the methods and the literature. This is good for a graduate level course, for a reference guide for practitioners and researchers in test equating. One important aspect of the problem is estimating the standard error of the test equating estimates. The delta method and parametric and nonparametric bootstrap methods are all described. An entire chapter is devote to item response theory and its application to test equating. The final chapter provides a very practical discussion of issues in the implementation of test equating methods and includes discussion of the choice of statistical procedures, determining the sample size, data collection, test specification and design, quality control and calibration. | ||