A textbook of statistics proper, giving statistical arguments and derivations, would be useful for many empirical workers and as a reference book for instructors in design-analysis. Innumerable statistics texts have been written by statisticians, but mainly for statistic students and statisticians. Those with an empirical perspective are mainly oriented toward engineering and agriculture (e.g., Box, Hunter, & Hunter, 1978; Draper & Smith, 1981; Montgomery, 1997; Snedecor & Cochran, 1980). In psychology, considerable portions of Maxwell and Delaney (1990) and Myers and Well (1991) seem to me to have the right spirit and level for a handbook of statistics for psychologists.
For several reasons, unfortunately, the statistical orientation perpetuates and amplifies itself. Among these reasons is the common statistical reaction that the cure for poor empirical judgment in design and analysis is more statistics. The more statistics the students know, the argument goes, the better they will be able to deal with the innumerable empirical complications that arise. In actuality, I believe, the opposite is true; the real need is for research judgment. In writing the core chapters for this book, I realized, slowly and tortuously, how much standard material has marginal usefulness and how much standard exposition is marred by inappropriate goals.
What is needed is not more statistics, but less; less statistics and more empirics; less statistics and more extrastatistics. This book is a step in this empirical direction.
With exercises requiring thought and judgment, however, including answers would short-circuit the learning process. What matters is the mental activity involved in seeking an answer, not the answer itself. Indeed, a wrong answer may be quite instructive. Answers are included in the companion Instructor's Manual.