Book Review: Ordinal Measurement in the Behavioral Sciences

Abstract

Reviews the book, "Ordinal Measurement in the Behavioral Sciences," by Norman Cliff and John A. Keats (see record 2002-06630-000). Both authors, particularly Norman Cliff, have argued for years against the treatment of most common forms of psychological measurement as interval or ratio data. Instead, the authors have suggested that ordinal test theory is a more appropriate method for scoring and analyzing measures such as attitude scales, supervisory performance ratings, and even mental ability tests. There are probably three types of readers who may pick up the book (and for whom the book might be targeted). One is the strong quantitatively oriented researcher, who has likely encountered Cliff and Keats before and either accepts or rejects their calls for more ordinal measurement. For this reader, the book might provide a concise compilation of previously read perspectives and methods. A second type of reader might be a student or a researcher who resembles the development psychologist in the joke above. He or she might have been referred to the book by a more knowledgeable colleague or might suspect that they in fact have ordinal data and might be looking for methods of scoring or analyzing the data. This type of reader will come away without the answers being sought. The third type is the applied researcher who frequently develops or uses scales or measures as either independent variables or dependent variables (the applied researcher). Because he or she relies on measured variables more than experimentally manipulated variables, it is important that these measures be appropriate, reliable, and valid. This third type of reader should be the intended audience of the book, and whether these researchers adopt ordinal test theory is a test of the impact of the book. (PsycINFO Database Record (c) 2012 APA, all rights reserved)