A Conceptual Guide to Statistics Using SPSS
Publication Year: 2012
DOI: http://dx.doi.org/10.4135/9781506335254
Subject: Quantitative Techniques for Business & Management Research, Research Methods for Sociology
 Chapters
 Front Matter
 Back Matter
 Subject Index

 Chapter 1: Introduction
 Goal of This Book: Conceptual Understanding
 Features That Will Help You in This Book
 A Note on Data Files
 Chapter 2: Descriptive Statistics
 Introduction to Descriptive Statistics
 Computing Descriptive Statistics in SPSS
 Interpreting the Output
 A Closer Look: Eyeballing a Hypothesis Test
 A Closer Look: Assessing for Normality
 Chapter 3: The ChiSquared Test for Contingency Tables
 Introduction to the ChiSquared Test
 Computing the ChiSquared Test in SPSS
 A Closer Look: Fisher's Exact Test
 The ChiSquared Test for Testing the Distribution of One Categorical Variable
 Chapter 4: Correlation
 Behind the Scenes: Conceptual Background of Correlation
 Covariance Versus Correlation
 Computing Correlation (and Covariance) in SPSS
 Interpreting the Correlation Output
 A Closer Look: Partial Correlations
 Visualizing Correlations
 Chapter 5: One and TwoSample tTests
 Conceptual Background of the tTest
 Behind the Scenes: The tRatio
 Computing the OneSample tTest Using SPSS
 Computing the PairedSamples tTest Using SPSS
 Computing the IndependentSamples tTest Using SPSS
 Connections: A Comparison of the Independent and PairedSamples tTests
 Visualizing the Results from the tTest
 A Closer Look: Testing the Assumptions Underlying the tTest
 Chapter 6: OneWay ANOVA
 Behind the Scenes: Conceptual Background of the Analysis of Variance (ANOVA)
 Computing the OneWay ANOVA Using SPSS
 Interpreting the ANOVA Output
 A Closer Look: Custom Contrasts in OneWay ANOVA
 Making the Most of Syntax: Custom Contrasts Using Syntax
 Connections: On the Equivalence of OneWay ANOVA and tTests
 Plotting the Results of the OneWay ANOVA
 A Closer Look: Testing Assumptions in OneWay ANOVA
 Chapter 7: Two and HigherWay ANOVA
 Conceptual Background of the HigherOrder ANOVA
 Behind the Scenes: Modeling Two and HigherWay ANOVA with the GLM
 Computing the TwoWay ANOVA Using SPSS
 Interpreting the ANOVA Output
 Making the Most of Syntax: Custom Contrasts in TwoWay ANOVA
 A Closer Look: MultipleLine Contrasts
 Connections: Equivalence Between Main Effects Tests and Custom Contrasts
 Plotting the Results of the TwoWay ANOVA
 Chapter 8: WithinSubject ANOVA
 Conceptual Background of the WithinSubjects ANOVA
 Behind the Scenes: Modeling the WithinSubjects ANOVA
 Computing the WithinSubjects ANOVA Using SPSS
 Interpreting the ANOVA Output
 Plotting the Results of WithinSubjects ANOVA
 Making the Most of Syntax: Custom Contrasts in WithinSubjects ANOVA
 Connections: Equivalence Between WithinSubjects ANOVA and PairedSamples tTests
 Chapter 9: MixedModel ANOVA
 Conceptual Background of the MixedModel ANOVA
 Computing the MixedModel ANOVA Using SPSS
 A Closer Look: Testing the Assumptions of the MixedModel ANOVA
 Interpreting the Output of the MixedModel ANOVA
 Plotting the Results of the MixedModel ANOVA in SPSS
 Making the Most of Syntax: Custom Contrast Tests in MixedModel ANOVA
 Chapter 10: Multivariate ANOVA
 Conceptual Background of the Multivariate ANOVA
 Computing MANOVA Using SPSS
 Interpreting the SPSS Output
 A Closer Look: Which Multivariate Test to Report?
 Making the Most of Syntax: Testing Custom Contrasts in MANOVA
 A Closer Look: When to Use MANOVA Versus WithinSubjects ANOVA
 Chapter 11: Linear Regression
 Behind the Scenes: Conceptual Background of Linear Regression
 Computing Linear Regression in SPSS
 Interpreting the Linear Regression Output
 Connections: Understanding the Meaning of Partial and Semipartial Correlations
 A Closer Look: Hierarchical Regression
 A Closer Look: Testing Model Assumptions
 Chapter 12: Analysis of Covariance
 Conceptual Background of Analysis of Covariance (ANCOVA)
 Computing ANCOVA in SPSS
 Understanding the Output of ANCOVA
 Visualizing Results of the ANCOVA
 Making the Most of Syntax: Custom Hypothesis Testing in ANCOVA
 A Closer Look: Evaluating the Assumptions of ANCOVA
 Chapter 13: Factor and Components Analysis
 Conceptual Background of the Factor Analysis
 Background Issues
 Computing the Factor Analysis in SPSS
 Interpreting the SPSS Output of Factor Analysis
 A Closer Look: Reporting the Results in a Research Paper
 Chapter 14: Psychometrics
 Conceptual Background of Psychometrics
 Preliminary Psychometrics in SPSS
 Computing Formal Psychometrics Analyses in SPSS
 Interrater Reliability in SPSS
 Chapter 15: Nonparametric Tests
 Conceptual Background of Nonparametric Tests
 The Sign Test (for OneSample Hypotheses)
 The Wilcoxon RankSum Test (for Independent Samples)
 The Wilcoxon SignedRank Test (for Paired Samples)
 Connections: A Comparison to the PairedSamples tTest
 The KruskalWallis Test (for BetweenSubjects Comparisons)
 Friedman's Rank Test (for WithinSubject Comparisons)
 Chapter 16: Matrix Algebra
 Conceptual Background of Matrix Algebra
 Overview of Matrix Algebra in SPSS
 Making the Most of Syntax: Solving the General Linear Equation in SPSS
 A Closer Look: Custom Hypothesis Testing Using Matrix Algebra
 Appendix: Commented Syntax for OneWay ANOVA Using Matrix Algebra in SPSS

Dedication
EB—I dedicate this book to my parents for encouraging my childhood nerdlihood, and to my wife for continuing to find it endearing.
Copyright
Copyright © 2012 by SAGE Publications, Inc.
All rights reserved. No part of this book may be reproduced or utilized in any form or by any means, electronic or mechanical, including photocopying, recording, or by any information storage and retrieval system, without permission in writing from the publisher.
FOR INFORMATION:
SAGE Publications, Inc.
2455 Teller Road
Thousand Oaks, California 91320
Email: order@sagepub.com
SAGE Publications Ltd.
1 Oliver's Yard
55 City Road
London EC1Y 1SP
United Kingdom
SAGE Publications India Pvt. Ltd.
B 1/I 1 Mohan Cooperative Industrial Area
Mathura Road, New Delhi 110 044
India
SAGE Publications AsiaPacific Pte. Ltd.
33 Pekin Street #0201
Far East Square
Singapore 048763
Printed in the United States of America
Library of Congress CataloginginPublication Data
Berkman, Elliot T.
A conceptual guide to statistics using SPSS / Elliot T. Berkman, Steven P. Reise.
p. cm.
ISBN 9781412974066 (pbk.)
1. SPSS for Windows. 2. Statistics—Computer programs. 3. Social sciences—Statistical methods—Computer programs. I. Reise, Steven Paul II. Title.
HA32.B47 2012
005.5′5—dc22
2011009895
This book is printed on acidfree paper.
11 12 13 14 15 10 9 8 7 6 5 4 3 2 1
Acquisitions Editor: Christine Cardone
Editorial Assistant: Sarita Sarak
Production Editor: Eric Garner
Copy Editor: Liann Lech
Typesetter: C&M Digitals, (P) Ltd.
Proofreader: Wendy Jo Dymond
Indexer: Brian Clark
Cover Designer: Gail Buschman
Marketing Manager: Liz Thornton
Permissions Editor: Karen Ehrmann
Preface
This book grew out of our experiences across many years of teaching introductory statistics to graduate students and advanced undergraduates in psychology. We noticed that our students faced a special set of challenges in learning statistics compared to other topics covered in the psychology curriculum. It was often the case that our students had little or no background in statistics and were consequently unfamiliar with thinking about the world in statistical or probabilistic terms. Even when they were familiar with statistics, our students often just didn't like it. And to make matters worse, in addition to their usual course load, they were also busy completing heavy research expectations (for graduate and honors students) or assisting faculty with their research. The dilemma for these students became how to simultaneously learn the challenging theoretical material taught in statistics class and to come away with the practical computational skills needed to advance their research?
The current text proposes to aid students by drawing clear connections between the theoretical and computational aspects of statistics, emphasizing the importance of understanding theoretical concepts during computation, and demonstrating how and where they fit in to SPSS, an IBM Company*. The text not only demonstrates how to use SPSS to advanced computation but also aids students' understanding of the theoretical concepts by teaching them in another, more practical context.
Our goal in this book is to clearly map the theories and techniques taught in a statistics class to the procedures in SPSS. The text teaches students how to perform standard and advanced statistical tests using both the pointandclick menus and syntax functions and how to integrate the SPSS functions with the statistical theory taught in class. The theoretical foundation underlying each topic are introduced before the computational steps in order to remind students of the logic of each statistical test. In this way, a conceptual link is created between the statistical test and the computational steps, and attention is drawn to testspecific issues. Presenting the material in this way also helps to give students a better understanding of the test output because they know which parameters were used “behind the scenes” in the computation. To better fit the material to the needs of a graduatelevel audience, advanced options and variations on each test are discussed, and the syntax commands are presented. This gives students more flexibility in tailoring their analyses to a wide variety of experimental paradigms.
It was impossible to cover all of the many statistical tests offered in the SPSS package. Instead, our goal was to provide coverage on any statistical test that might appear in a peerreviewed psychology journal article. The text features detailed chapters on the tests most commonly used by psychologists, as well as several newer tests that are increasing in popularity. Each chapter is structured similarly so students familiar with the text will be able to quickly flip open the book to learn a new topic.
The book is organized in parallel to many standard statistics textbooks covering correlation, ttests, ANOVA and MANOVA, multiple regression, and nonparametric tests. Each chapter begins with a brief conceptual introduction featuring test assumptions and a sketch of the mathematical operations behind a procedure. This is followed by an illustrated and annotated stepbystep guide to computation with references back to the introduction where possible and concludes with a discussion of the output.
Target AudienceThis book is intended for anyone who not only wants to know how to use SPSS to compute a variety of statistical tests, but who also wants to understand the reasons behind each step and the conceptual meaning of the output. This includes advanced undergraduates in the social sciences; master's students in psychology, education, economics, public health, biological sciences, and counseling psychology; PhD students in the social sciences; and faculty in all these fields seeking a deeper understanding of SPSS than that offered by the usual stepbystep procedural guides. Most of the examples are drawn from research in social and personality psychology, but the tests used are common across many fields that make use of empirical behavioral data.
This text is sufficiently detailed to serve as a standalone guide to SPSS, but also is intended to complement a statistics textbook for a variety of undergraduate and graduate statistics courses in the social sciences. Because we cover topics ranging from ttests and regression to factor analysis and matrix algebra, and because we describe both basic and advanced features of SPSS for each, we are confident that SPSS users at all levels of expertise will find something new and useful in this book.
Special FeaturesBehind the ScenesThe Behind the Scenes sections explain the conceptual machinery underlying the statistical tests. In contrast to merely presenting the equations for computing the statistic, these sections describe the idea behind each test in plain language. In writing these sections, we sought to answer, in conceptual terms, the questions, What does SPSS do with your data to transform it into the test statistic? Which parts of the data are important for this calculation? and How does the output relate to the meaning of the test? After that, and only where it is helpful to building a conceptual understanding, we give the equation for the test and explain each part in terms of the idea behind the test. Several Behind the Scenes sections also contain schematic diagrams that are intended to clarify how different patterns of data relate to key ideas in the test. These sections were written specifically for introductory students seeking to make a connection between the ideas taught in a statistics course or textbook and the SPSS procedure.
ConnectionsThe Connections sections use SPSS to demonstrate the equivalence among tests that are often treated as distinct. Particularly for introductory students, the syllabus of a statistics course can seem like a laundry list of unrelated tests. The layout of SPSS also supports this impression by segregating similar tests into different menus. The purpose of the Connections sections is to provide a “bigger picture” perspective by highlighting the conceptual similarities across tests. We do this by showing commonalities within a family of tests (e.g., those based on the general linear model) and by relating entirely different types of tests to each other (e.g., between nonparametric tests and ANOVAbased tests). We also use the Connections sections to point out similarities in the SPSS output across different but related statistical tests.
A Closer LookThe A Closer Look sections feature advanced topics that are beyond the scope of other introductory SPSS books. These sections teach the reader how to use SPSS to compute tests or display output that is can be important to report in a research paper but that SPSS does not compute or display by default. Though the topics are more advanced or specialized, the A Closer Look sections are nonetheless written so that introductory students can understand when and why they might want to use them and that more advanced students can quickly learn how to compute them. Topics covered in A Closer Look sections include custom hypothesis tests among group means in ANOVA, assumption checking in the General Linear Model, and saving predicted scores in multiple regression.
Making the Most of SyntaxIn the Making the Most of Syntax sections we describe statistical tests and output options that are exclusive to syntax. These include extensive treatment of custom hypothesis testing in ANOVA, MANOVA, ANCOVA, and regression, and an entire chapter on the advanced matrix algebra functions available only through syntax in SPSS. Our emphasis on the powerful capacity of the syntax functions is unique among introductory SPSS books. In order to help the reader learn how to use syntax in your own research, we provide the general form and also a specific example of each syntax function. As always, we emphasize conceptual understanding by linking the specifics of the syntax functions to the general idea behind the test.
This section also highlights the value of using syntax for all statistical tests even when other options are available. Syntax is the easiest way to rerun statistical tests with slight variations or with different variables. And by describing the syntax corresponding to every topic, this book teaches the reader to create a syntax log that provides a complete record of your data analysis process from data cleaning all the way through to figures for publication.
Data FilesEach of the statistical tests covered here is accompanied by an example data set, and the screenshots and output that are shown in each chapter are based on these data sets. Our intention is that the reader can follow along and practice analysis using these data sets, so we have made the data files available on the book webpage at http://www.sagepub.com/berkman. We hope it is clear from the content of the data sets that they are simulated and intended for illustrative purposes only.
AcknowledgmentsWe would like to acknowledge the insightful feedback from our brilliant colleagues in statistics education, Emily Falk and Hongjing Lu, as well as the willingness of many of our students to serve as proofreaders and guinea pigs for this book over the last few years. We also appreciate helpful comments from several expert reviewers in the field. They made our jobs easier and improved the book substantially.
*Note: SPSS was acquired by IBM in October 2009.
About the Authors
Elliot T. Berkman is Assistant Professor of Psychology and director of the Social and Affective Neuroscience Laboratory at the University of Oregon. He has been teaching statistics to graduate students using SPSS for the past six years. In that time, he has been awarded the UCLA Distinguished Teaching Award and the Arthur J. Woodward Peer Mentoring Award. He has published numerous papers on the social psychological and neural processes involved in goal pursuit. His research on smoking cessation was recognized with the Joseph A. Gengerelli Distinguished Dissertation Award. He received his PhD in 2010 from the University of California, Los Angeles.
Steven P. Reise is professor, chair of Quantitative Psychology, and codirector of the Advanced Quantitative Methods training program at University of California, Los Angeles. Dr. Reise is an internationally renowned teacher in quantitative methods; in particular, the application of item response theory models to personality, psychopathology, and patient reported outcomes. In recognition of his dedication to teaching, Dr. Reise was named “Professor of the Year” in 1995–96 by the graduate students in the psychology department at UC Riverside, and was awarded the 2008 Psychology Department Distinguished teaching award. Most recently, in recognition of his campuswide and global contributions, Dr. Reise was awarded the University of California campuswide distinguished teaching award. Dr. Reise has spent the majority of the last twenty years investigating the application of latent variable models in general and item response theory (IRT) models in particular to personality, psychopathology and health outcomes data. In 1998, Dr. Reise was recognized for his work and received the Raymond B. Cattell award for outstanding multivariate experimental psychologist. Dr. Reise has over 70 refereed publications, including, two Annual Review Chapters, two contributions to American Psychological Association Handbooks, several articles in leading journals such as Psychological Assessment and Psychological Methods, and, finally, along with Dr. Susan Embretson, Dr. Reise has the leading textbook on item response theory called Item Response Theory for Psychologists (2000 and forthcoming). He received his PhD from the Department of Psychology at the University of Minnesota in 1990.

Appendix: General Formulation of Contrasts Using LMATRIX
In Chapter 7, we walked through a single example of how to use the LMATRIX function to compute a few custom contrasts in syntax. But the “LMATRIX” function in SPSS is a powerful tool that can compute nearly any contrast among any combination of group means. In the most general terms, the steps for figuring out the right syntax are as follows:
 Write down the contrast coefficients for each cell in a table like the one shown in the figure. The first factor should be along the rows, and the second factor should be along the columns.
If you have a threeway ANOVA, make separate tables for each level of the first factor, with the second factor levels in the rows and the third factor levels in the columns. (For example, suppose we wanted to look at the gender of the guests in addition to their side and relationship to the couple. Then there would be two 2 × 4 tables like the one shown in the figure, one for males and one for females.)
 Compute the marginal sums across the rows and across the columns. If you have a threeway ANOVA, make a new table that is identical in form to the ones you made for each level of the first factor and contains a sum of all the other tables.
 After the “/LMATRIX =” tag, list out all of the factors and all of the interactions in the same order as they are listed in the GLM and DESIGN tags. For example,
/LMATRIX = IV1 IV2 IV1*IV2
for two factors, or
/LMATRIX = IV1 IV2 IV3 IV1*IV2 IV1*IV3 IV1*IV2*IV3
for three factors.
 Write down the cell values and marginal sums for each term based on the tables you generated in Step 2. With two factors (one with 2 levels and the other with 4 levels), the general form (based on the tables) is
 Remove any term (and its coefficients) if all the coefficients are equal to 0.
For example, suppose we wanted to compute the following contrast based on the data set “Wedding.sav” from Chapters 6 and 7, which tests whether the difference in dancing between coworkers and family on the bride's side is different between male and female guests.
The corresponding syntax is
/LMATRIX = gender*relation 0 −1 0 1 0 1 0 −1
gender*side*relation 0 −1 0 1 0 0 0 0 0 1 0 −1 0 0 0 0
Syntax Index
 /*, 9, 281–282
 ALPHA, 79, 242, 246
 ANALYSIS, 30, 56, 227
 ANOVA, 185–186, 196, 202
 BAR, 64, 66–67, 94, 121
 BINOMIAL, 252
 BLANK, 227, 232
 BOXPLOT, 70, 166
 BY, 24, 42, 66, 78–79, 89, 91, 95, 106, 117–118, 121, 166, 170, 209–210, 215, 219, 254, 262, 275, 280
 CAPWIDTH, 67
 CHANGE, 196, 202
 CI or CIN, 56, 94, 122, 246
 COEFF, 185–186, 196, 202
 COMPARE VARIABLES, 166
 COMPUTE, 140, 202, 269–270, 272–282
 CONTRAST, 105–106
 CORR, 185, 242
 CORRELATIONS, 38, 41, 188, 193, 227–228, 238
 CRITERIA, 56, 79, 227
 CROSS, 97
 CROSSTABS, 23
 DEPENDENT, 185–186, 193, 196, 202
 DESCRIPTIVES, 10, 185–186, 254, 257, 265–266
 DESIGN, 79, 89, 91, 106, 113, 117–118, 147–148, 170, 209–210, 217, 219, 273, 275, 280, 281, 287
 DIST, 199
 EIGEN, 227
 EMMEANS, 209, 213
 END MATRIX, 269, 271, 273, 282
 EQUAL, 30
 ERRORBAR, 67
 ETASQ, 106
 EXAMINE, 70, 166
 EXE, 140, 202
 EXPECTED, 30
 EXTRACTION, 227
 FACTOR, 227
 FACTORS, 227
 FCDF, 270, 277, 280, 282
 FORMAT, 9, 23, 227, 232
 FREQUENCIES, 7, 9, 239
 FRIEDMAN, 266
 GET, 270, 272–273
 GINV(X), 270, 274, 279, 281–282
 GLM, 78–79, 89, 91, 105–106, 113, 116, 118, 131–132, 137–138, 147, 170, 209–210, 215, 217, 219, 275–276, 278, 280, 287
 GRAPH, 44, 47, 64, 66, 94, 97, 121, 215
 GROUPED, 121
 GROUPS, 60, 91
 HISTOGRAM, 9, 97
 HOMOGENEITY, 95, 106, 147, 151, 170, 172, 209
 ICC, 246
 IGRAPH, 47, 67
 INITIAL, 227
 INTERVAL, 64, 66, 94, 122
 ITERATE, 227
 JITTER, 47
 KMATRIX, 89, 138, 176–177, 216
 KW, 262
 LISTWISE, 39, 41–42
 LMATRIX, 89, 112–113, 115–116, 118, 158–160, 169, 175–177, 208, 216–217, 280, 285–289
 MATRIX, 269, 271, 273, 281
 MEAN, 64, 66, 67, 94, 121, 140, 185, 242
 METHOD, 78–79, 185–186, 193, 196, 202, 227–228
 MISSING, 30, 38–39, 41–42, 56
 MMATRIX, 137–138, 158–160
 MODEL, 242, 246
 MSUM(X), 270, 276, 281
 MW, 254
 N, 185
 NCOL, 274, 276, 281
 NONE, 70
 NORMAL, 9, 97, 199
 NOSIG, 38
 NOTABLE, 9
 NOTOTAL, 70, 166
 NPAR TESTS, 30, 252, 254, 257, 262, 266
 NROW, 274, 279, 281–282
 NTILES, 9
 OUTS, 185, 196, 202
 PAF, 227
 PAIRED, 56, 139–140, 257, 259
 PAIRS, 56 139–140, 259
 PAIRWISE, 38–39, 41
 PARAMETER, 209, 211
 PARTIAL CORR, 41–42
 PLOT, 70, 79, 96, 106, 131, 147, 155, 166, 170, 227, 232
 POLYNOMIAL, 132
 PPLOT, 199
 PRED, 185, 187
 PRINT, 38, 79, 95, 106, 132, 134, 147, 149, 151, 170, 172, 209, 211, 227, 269–270, 272, 277, 280, 282
 PROFILE, 79, 106, 131, 147, 155, 170
 PROMAX, 227
 R, 185–186, 196, 202
 REG, 227
 REGRESSION, 185–186, 193, 196, 202
 RELIABILITY, 241, 246
 REPR, 227
 ROTATION, 227, 232
 ROWOP, 97
 ROWVAR, 97
 RSSCP, 134, 147, 149, 151, 170
 SAVE, 10, 185–187, 193, 198, 227
 SCALE, 241–242, 246
 SIG, 185
 SIGNIFICANCE, 42
 STATISTICS, 9, 11, 23, 38, 70, 166, 185–186, 190, 196, 202, 242, 254, 257, 262, 265–266
 STDDEV, 185
 SUMMARY, 242
 TABLES, 23, 209
 TESTVAL, 53–54, 246
 TOTAL, 242
 TTEST, 53–54, 56, 60, 91
 TTEST PAIRS, 139–140, 259
 TWOTAIL, 38, 42
 TYPE, 199, 246
 VARIABLES, 7, 9, 10, 38, 41–42, 53, 60, 70, 91, 166, 188, 193, 199, 227, 238–239, 241, 246
 VARIANCE, 242
 WILCOXON, 257
 WITH, 44, 56, 139–140, 209–210, 215, 219, 257, 259
 WSDESIGN, 132, 137, 147–148, 170
 WSFACTOR, 131–132, 137, 147, 170
 XPROD, 38
 ZPP, 185–186, 190

168976 Loading...
Also from SAGE Publishing
 CQ Library American political resources opens in new tab
 Data Planet A universe of data opens in new tab
 Lean Library Increase the visibility of your library opens in new tab
 SAGE Journals Worldclass research journals opens in new tab
 SAGE Research Methods The ultimate methods library opens in new tab
 SAGE Stats Data on demand opens in new tab