Reviews the book Statistical approaches to measurement invariance by Roger E. Millsap. The reviewer found this book to be well written and feels it deserves to become the classic reference in the field, but also that the reader should be prepared to invest time to read it. Millsap’s book is not the easiest to read because it contains 432 equations. These equations are given in nine well-structured chapters and offer a near-comprehensive account of the current state of knowledge of measurement invariance. Millsap provides extensive background, is technically rigorous, and illustrates the approaches with interesting psychological examples. So why does the study of measurement invariance get so complex that Millsap devotes 432 equations to it? The reason is that "whatever it is a test purports to measure" cannot be observed directly; it is a latent variable. Because we measure the latent variable by means of a test that may be a biased measure of the latent variable, this is somewhat of a puzzle. To solve it, we need several measures of the latent variable, like items or subtests, and a statistical model that relates observed test scores with the latent variable. Studying measurement invariance entails checking whether this relation is equal across groups. Millsap’s book provides a broad and very thorough exposition of the most important topics associated with the statistical study of measurement invariance. Statistical Approaches to Measurement Invariance is not read easily because measurement invariance is not an easy topic. However, in the reviewer's opinion the book is a must read for psychometricians and for researchers who compare test scores across groups. Readers who invest their time in solving the 432 pieces of Millsap’s interesting puzzle will be equipped with invaluable knowledge of the nature of group differences in psychological measurement.
|Number of pages||1|
|Publication status||Published - 2011|