Introduction to Measurement Theory:

In general, this text is a really great balance between precision and conciseness. I had McDonald's Test Theory but that's just too much for many students in their first trip thru psychometric theory. (It's required later in our program.) The problems are good, with answers in the back of the book so these can be assigned as pencil and paper exercises for understanding. I especially like the chapter where the conclusions of Classical Test Theory are laid out and proven, which demonstrates the value of taking a few axioms as far as you can go. The one downside of this book is that it is showing its age now. There aren't glaring issues with it and it doesn't have a lot of "if you are doing this on a pocket calculator..." kinds of statements that mar other older books, but better mention of topics such as confirmatory factor analysis would be helpful.

I pair this text with one on IRT in my graduate psychometrics course. This classic text provides an excellent overview of classical test theory in concise, no-nonsense terms while still being easy to read for an audience of non-statisticians. The chapters are well-organized with helpful study aids. The book begins with a brief introduction to psychological/educational measurement and a bit of history. The second chapter is a review of basic mathematical concepts. The fun really starts in chapter 3 with the basic propositions of classical test theory and a list of results--this chapter is a marvel of brevity and clarity. The next chapters on reliability (ch. 4), validity (ch. 5), test construction (ch. 6), and equating (ch. 7) build sound practice upon this sound theoretical foundation. Scaling issues are covered superficially in chapter 8 and then a number of important issues are raised in chapter 9 ("Special Considerations..."). The final two chapters ("Controversies and current developments..." and a chapter on IRT) are most dated and least helpful (although, like the rest of the text, the material is covered well and these chapters would serve as a good foundation for understanding recent developments). In comparison, other psychometric texts tend to be more comprehensive at the expense of being quite shallow. These other texts, which also tend to be verbose, may be better-suited to undergraduate courses designed to gently introduce unsophisticated students to the bare basics of measurement. Allen & Yen is better suited for serious undergrads and graduate-level students who wish to understand classical test theory and to apply it well. Special note for specialists: while this book includes many issues critical to designing selection tests, industrial psychologists and HR professionals will need additional training on selection (e.g., job analysis) and on legal issues; other specialists will probably also find the book to be quite focused on measurement. Special note on versions: I own both the 1979 burgundy hardcover and 2001 blue-black paperback editions and, as far as I can tell, they are identical.

This looks a small book, but the content is comprehensive. The authors clearly get to the point which other this type of books do not.