Reviewer: James C. Torner, PhD, MS (University of Iowa College of Public Health)
Description: Increasingly, researchers attempt to assess quality of life, subjective states, and attitudes. This book provides detailed guidance on scale development, reliability, validity, bias, administration, ethics, and generalization theory. Prior editions of this book were published in 2003, 1995, and 1989.
Purpose: The purpose is to provide a resource to researchers in the health sciences (e.g., epidemiology, medicine, etc.) on the subject of measurement. The book succeeds (as did prior editions) in covering aspects of psychosocial measurement with relevant health-related examples.
Audience: Although the target audience is health researchers, the book is appropriate not just for researchers, but also for graduate students. The authors are well regarded and their book is a staple on the bookshelves of many researchers interested in measurement of subjective states, attitudes, etc.
Features: Starting with the importance of conducting a literature review and an overview of basic comments, including validity and measurement error, the book moves on to item development. It clearly describes how best to scale responses and when particular response approaches are most appropriate. The book delves into dealing with missing items, weighting items, standardization of scores, and methods of categorization and establishing cutpoints based on distribution of the data, judgment, and approaches to determining sensitivity and specificity. It has a new chapter on generalizability theory (G-theory) and a new and more straightforward approach to defining and assessing validity. It describes how to measure change in different types of studies, measures of association, and growth curves. The chapter on item response theory covers complex models, classical test theory, and differential item functioning. Another new chapter covers methods of administration, including computer-assisted administration. The book ends with chapters on ethics and reporting test results based on the Standards for Reporting Diagnostic Accuracy (STARD) initiative.
Assessment: The authors have added new material and expanded other material in important ways since the first edition. While all chapters are useful, the updated discussion of validity and the new section on G-theory will stand out as important references for readers. I have always kept a copy of the first edition of this book on my shelves. It is regularly loaned out (and never returned), so I have replaced copy after copy over the years. I look forward to replacing my old edition with this new, more fleshed out, more complete, and very up-to-date book. I will continue to use it in my classes.