Intermediate Statistics: A Modern Approach, Third Edition

Book Preface

This book is written for behavioral and social science students at the advanced undergraduate or beginning graduate level. The text emphasizes conceptual understanding, the effective use of statistical software to run the analyses, and the correct interpretation of results. Two statistical software packages, SAS and SPSS, are an integral part of each chapter. An annotated printout is given from at least one of the programs for each analysis. The annotations highlight what the numbers mean and how to interpret the results. The explanation appears on the printout or on the same page to enhance learning efficiency. The assumptions underlying each analysis are given special attention, and the reader is shown how to test the critical assumption( s) using SAS and SPSS. Power analysis is an integral part of the book. There are no computational formulas in this text. I took the position that they were not needed many years ago, and it is even truer today.

The instructional mix of strategies that is employed to illustrate each statistical technique consists of two parts (a) First, I use definitional formulas on small data sets to convey conceptual insight into what is being measured, and (b) Then, I proceed directly to the packages to efficiently process data. I feel very strongly about using these strategies.

The most significant change in this edition is the addition of a chapter on hierarchical linear modeling using HLM6. This material is important because correlated observations occur frequently in social science research and just a SMALL amount of dependence causes the type I error rate to be several times greater than one wishes! Since HLM involves a series of regressions, this new chapter is placed after the material on regression. The distinction between fixed and random factors is important, and so it is emphasized. The chapter on HLM was written by Dr. Natasha Beretvas of the University of Texas at Austin. I thank her very much for her contribution.

The third edition features newer versions of SPSS (Release 12.0) and SAS (Release 8.0). Much of the material on importing data into SAS or SPSS that previously appeared in chapter 1 was deleted. Importing data into these two programs is now much easier so this material was no longer necessary.

The exercises involve a mixture of numerical, conceptual and computer related problems. I have de-emphasized purely numerical exercises, for I agree entirely with Cobb (l987, p. 323) that, “computing rules are just the skin of our subject; it is focus that reveals the skeleton of fundamental concepts and connections that hold the body of knowledge together.” Regarding exercises, it is important to note that there are 3 new exercises for each chapter. Answers are provided for half of the exercises and an Instructor’s Solutions CD is available to adopters. A computer example of real data integrates many of the concepts. A CD containing all of the book’s data sets is included in the back of the book.

The reader should have a background of a one quarter course in statistics that covered at least the t tests for independent and dependent samples.

I am very grateful to the reviewers of this text: Dale Berger of Claremont Graduate University, Michael Milburn of University of Massachusetts, Mary Lou Kerwin of Rowan University, Gordon Brooks of Ohio University, and Roderick Gillis of the University of Miami. I am also indebted to some individuals at my publisher. Larry Erlbaum continues to be very supportive. Debra Riegert was instrumental in motivating me to write this third edition.

Jim Stevens