Comment on Maxwell and Delaney

Abstract

Serlin (1991) praises the Maxwell and Delaney (1990) text Designing Experiments and Analyzing Data: A Model Comparison Perspective as an excellent reference work yet expresses the opinion that it is too advanced to be used as a first-year graduate textbook. I must disagree with this opinion. In his review, Serlin begins by highlighting one sentence from the preface of the book concerning the authors' presumption that the reader will have had high school algebra and an undergraduate statistics course. He then uses this criterion to evaluate the content of the book and its appropriateness as a textbook for novice graduate students. In explaining his concerns, Serlin lists several terms that apparently are not accompanied by an explicit definition the first time they appear in the book. He interprets this to mean that the authors must be supposing students will have prior knowledge of these topics and expresses doubt that they could have obtained the necessary background from a single undergraduate statistics course. Serlin does not mention, however, that some of these early references are simply informing readers of the location where these topics are later discussed. He also fails to mention that these later discussions, rather than presuming knowledge from prior coursework, begin by introducing the topics at a basic level. To cite one example, Serlin construes that prior study of within-subjects designs is a prerequisite for the book from one sentence that simply refers to a later treatment of the topic (However, we will see in Chapter 13 that the Bonferroni approach may be preferable . . . in within-subjects designs . . ., p. 184). In fact, 237 pages of the text are devoted to introducing within-subjects designs and analyses. Clearly Serlin's reasoning falls short of providing a meaningful assessment of the material students must know in advance to understand this book. What is most disconcerting about Serlin's review, however, is the conclusion he draws from his content analysis that the difficulty level of the book may limit its usefulness as a standard graduate statistics textbook. My own experience as a first-year graduate student belies this assessment. I had the opportunity to use this text (in manuscript form) during the second semester of a three-semester grad-