IBM SPSS STATISTICS 19 Made Simple Colin D. Gray & Paul R. Kinnear School of Psychology, University of Aberdeen First published 2012 by Psychology Press 27 Church Road, Hove, East Sussex BN3 2FA Simultaneously published in the USA and Canada by Psychology Press, 711 Third Avenue, New York NY 10017 Psychology Press is an imprint of the Taylor & Francis Group, an informa business Copyright © 2012 Psychology Press Printed and bound in Great Britain by TJ International Ltd, Padstow, Cornwall, from pdf files supplied by the authors. Cover design by Hybert Design. All rights reserved. No part of this book may be reprinted or reproduced or utilised in any form or by any electronic, mechanical, or other means, now known or hereafter invented, including photocopying and recording, or in any information storage or retrieval system, without permission in writing from the publishers. This book is not sponsored or approved by SPSS, and any errors are in no way the responsibility of SPSS. SPSS is a registered trademark and the other product names are trademarks of SPSS Inc. SPSS Screen Images © SPSS Inc. SPSS UK Ltd, First Floor St Andrew’s House, West Street, Woking, Surrey, GU21 1EB, UK. Windows is a registered trademark of Microsoft Corporation. For further information, contact: Microsoft Corporation, One Microsoft Way, Redmond, WA 98052-6399, USA. This publication has been produced with paper manufactured to strict environmental standards and with pulp derived from sustainable forests. British Library Cataloguing in Publication Data A catalogue record for this book is available from the British Library. Library of Congress Cataloging-in-Publication Data Gray, Colin D. IBM SPSS statistics 19 made simple / Colin D. Gray and Paul R. Kinnear. p. cm. Includes bibliographical references and index. ISBN 978-1-84872-069-5 (soft cover) 1. SPSS (Computer file) 2. Social sciences--Statistical methods--Computer programs. I. Kinnear, Paul R. II. Title. HA32.G727 2012 005.5'5--dc23 2011025014 ISBN: 978-1-84872-069-5 (pbk only) Contents Preface xiii CHAPTER 1 INTRODUCTION 1 1.1 AIMS AND OBJECTIVES 1 1.2 MEASUREMENTS AND DATA 2 1.2.1 Variables: quantitative and qualitative 2 1.2.2 Levels of measurement: scale, ordinal and nominal data 2 1.2.3 Univariate, bivariate and multivariate data sets 3 1.3 DISTRIBUTIONS 4 1.3.1 The three most important properties of a distribution 5 1.3.2 Some common distribution shapes 6 1.4 EXPERIMENTAL VERSUS CORRELATIONAL RESEARCH 7 1.4.1 A simple experiment 8 1.4.2 A more complex experiment 9 1.4.3 A correlational study 10 1.4.4 Quasi-experiments 13 1.5 CHOOSING A STATISTICAL TEST 13 1.5.1 Considerations in choosing a statistical test 13 1.5.2 Five common research situations 14 1.6 A SIGNIFICANT DIFFERENCE? 15 1.6.1 Independent or related samples? 15 1.6.2 Flow chart 16 1.7 ARE TWO VARIABLES ASSOCIATED? 17 1.7.1 Flow chart 18 1.7.2 Measuring association in ordinal data 18 1.7.3 Measuring association in nominal data: contingency tables 19 1.7.4 Multi-way contingency tables 20 1.8 MAKING PREDICTIONS 20 1.8.1 Flow chart 20 1.8.2 Simple regression 21 1.8.3 Multiple regression 21 1.8.4 Predicting category membership 22 1.9 FROM A SINGLE SAMPLE TO THE POPULATION 22 1.9.1 Flow chart 22 1.9.2 Goodness-of-fit: nominal data 23 1.9.3 Inferences about the mean of a single population 23 1.10 THE SEARCH FOR LATENT VARIABLES 24 1.11 MULTIVARIATE STATISTICS 24 iii iv Contents 1.12 A FINAL WORD 25 CHAPTER 2 Getting started with IBM SPSS Statistics 19 26 2.1 INTRODUCTION 26 2.2 DESCRIBING THE DATA FROM THE CAFFEINE EXPERIMENT 29 2.2.1 Opening SPSS 29 2.2.2 The SPSS Statistics Data Editor 30 2.2.3 Working in Variable View 30 2.2.4 Working in Data View 35 2.2.5 Computing the group means 39 2.2.6 The SPSS Statistics Viewer 42 2.2.7 The output from the Means procedure 46 2.2.8 Histograms 47 2.2.9 Editing items in the Viewer 52 2.2.10 Ending the session 53 2.2.11 Resuming work on a saved data set 53 2.3 THE FOUR DRUGS EXPERIMENT 53 2.3.1 In Variable View 54 2.3.2 In Data View 56 2.3.3 Using the Means procedure 57 2.3.4 The histograms 59 2.4 PRINTING FROM THE STATISTICS VIEWER 60 2.5 USING SPSS SYNTAX 64 2.5.1 The Syntax Editor 65 2.5.2 Running the Means procedure with Syntax 66 2.5.3 Looping functions in Syntax 69 2.6 A FINAL WORD 69 CHAPTER 3 Editing data sets 70 3.1 MORE ON THE DATA EDITOR 70 3.1.1 A preliminary check on the default settings 70 3.1.2 Inserting new variables 71 3.1.3 Rearranging the order of variables in the Data Editor 72 3.1.4 Inserting case numbers 73 3.1.5 Using Syntax to insert case numbers into a data set 76 3.1.6 Inserting case numbers into an empty Data Editor 76 3.1.7 Using Syntax to insert case numbers into an empty Data Editor 77 3.1.8 Changing the default settings for Width and Decimals 77 3.1.9 String variables 78 3.1.10 Specifying missing values 80 3.1.11 Changing the Alignment settings 82 3.1.12 Opening an SPSS file 82 3.1.13 Entering data from other applications 83 3.1.14 Creating new variables while in DataView 84 3.1.15 Adding new cases while in Data View 85 3.2 VALIDATION OF THE DATA 85 Contents v 3.2.1 Validation of the data by defining rules 85 3.2.2 Handling multiple duplications in a large data set 91 3.3 EDITING ITEMS IN THE VIEWER 92 3.3.1 Changing the format of a table 92 3.3.2 Widening, narrowing and hiding columns in a Report table 94 3.3.3 Deleting rows and columns from a Report table 95 3.3.4 Editing the captions in a table 96 3.3.5 Changing the number of decimal places displayed in a table 97 3.4 SELECTING CASES 97 3.5 AGGREGATING DATA 99 3.6 SORTING DATA 101 3.7 FILE-MERGING: ADDING CASES 103 3.8 FILE-MERGING: ADDING VARIABLES 105 3.9 THE COMPUTE VARIABLE COMMAND 108 3.9.1 Calculating functions of several variables 108 3.9.2 Conditional transformations 109 3.10 THE RECODE COMMAND 112 3.11 A FINAL WORD 115 CHAPTER 4 Describing and exploring your data 117 4.1 INTRODUCTION 117 4.2 DESCRIBING NOMINAL DATA 119 4.2.1 Describing nominal data on one attribute 119 4.2.2 Two attributes: contingency tables 121 4.3 DESCRIBING CONTINUOUS OR SCALE DATA 130 4.3.1 Histograms of height and weight for males and females 131 4.3.2 Obtaining scatterplots of weight against height 132 4.3.3 Statistics of height in males and females 135 4.4 DESCRIBING SMALL DATA SETS 137 4.4.1 The stem-and-leaf display and the boxplot 138 4.4.2 Exploring a small data set 142 4.4.3 Some of the statistical output from Explore 143 4.5 DESCRIBING DATA FROM MULTIPLE RESPONSE QUESTIONNAIRES 145 4.5.1 Data for the Multiple Response procedure 147 4.5.2 Creating a multiple response set 148 4.5.3 Obtaining the crosstabulations 150 4.5.4 Finding the frequencies of specific transport profiles 155 4.6 A FINAL WORD 159 CHAPTER 5 More on graphs and charts 161 5.1 INTRODUCTION 161 5.2 BAR CHARTS 162 5.2.1 Simple bar charts 162 5.2.2 Clustered bar charts 163 5.2.3 Panelled bar charts 166 vi Contents 5.2.5 Editing a bar chart 167 5.3 ERROR BAR CHARTS 171 5.4 BOXPLOTS 173 5.5 PIE CHARTS 173 5.6 LINE GRAPHS 174 5.6.1 The Visual Binning procedure 175 5.6.2 Plotting line graphs 177 5.7 USING CHART TEMPLATES 179 5.8 A FINAL WORD 183 CHAPTER 6 Comparing averages: Two-sample and one- sample tests 184 6.1 OVERVIEW 184 6.2 THE INDEPENDENT-SAMPLES T TEST WITH SPSS 185 6.2.1 Running the t test 185 6.2.2 Interpreting the output 188 6.2.3 Two-tailed and one-tailed p-values 190 6.2.4 Measuring effect size 191 6.2.5 Reporting the results of a statistical test 193 6.3 THE RELATED-SAMPLES (OR PAIRED-SAMPLES) T TEST 194 6.3.1 Preparing the data file 195 6.3.2 Exploring the data 195 6.3.3 Running the t test 196 6.3.4 Interpreting the output 197 6.3.5 Measuring effect size 198 6.3.6 Reporting the results of the test 199 6.3.7 A one-sample test 199 6.4 NONPARAMETRIC TESTS 199 6.4.1 Nonparametric tests in SPSS 200 6.4.2 Independent samples: The Mann-Whitney U test 201 6.4.3 Output from the Mann-Whitney U test 203 6.4.4 Exact tests 206 6.4.5 Effect size 207 6.4.6 The report 208 6.5 THE WILCOXON MATCHED-PAIRS TEST 209 6.5.1 The Wilcoxon matched-pairs tests in SPSS 209 6.5.2 The output 210 6.5.3 Effect size 214 6.5.4 The report 215 6.6 THE SIGN AND BINOMIAL TESTS 215 6.6.1 The sign test in SPSS 216 6.6.2 Bernoulli trials: the binomial test 219 6.7 EFFECT SIZE, POWER AND NUMBER OF PARTICIPANTS 223 6.8 A FINAL WORD 225 Contents vii CHAPTER 7 The one-way ANOVA 227 7.1 INTRODUCTION 227 7.2 ENTERING THE DATA 233 7.3 RUNNING THE ONE-WAY ANOVA ON GLM 236 7.3.1 Finding the GLM menu 236 7.3.2 Descriptives and the ANOVA summary table 241 7.3.3 The profile plot 243 7.3.4 Measuring effect size 244 7.3.5 Report of the analysis 246 7.4 MAKING COMPARISONS AMONG THE TREATMENT MEANS 247 7.4.1 Planned and unplanned comparisons 247 7.4.2 Linear contrasts 251 7.4.3 Helmert contrasts 252 7.5 POWER AND EFFECT SIZE IN THE ONE-WAY ANOVA 258 7.6 ALTERNATIVES TO THE ONE-WAY ANOVA 260 7.7 A FINAL WORD 266 CHAPTER 8 Between subjects factorial experiments 268 8.1 INTRODUCTION 268 8.1.1 An experiment with two treatment factors 268 8.1.2 Main effects and interactions 270 8.1.3 Profile plots 271 8.2 HOW THE TWO-WAY ANOVA WORKS 272 8.2.1 The two-way ANOVA 272 8.2.2 Degrees of freedom 273 8.2.3 The two-way ANOVA summary table 274 8.3 THE TWO-WAY ANOVA WITH SPSS 275 8.3.1 Entering the data 276 8.3.2 Exploring the data: boxplots 277 8.3.3 Choosing a factorial ANOVA 277 8.3.4 Output for a factorial ANOVA 280 8.3.5 Measuring effect size in the two-way ANOVA 282 8.3.6 Reporting the results of the two-way ANOVA 285 8.4 FURTHER ANALYSIS 286 8.4.1 A problem with multiple comparisons 286 8.4.2 Unpacking significant main effects: post hoc tests 286 8.4.3 The analysis of interactions 287 8.5 TESTING FOR SIMPLE MAIN EFFECTS WITH SYNTAX 289 8.5.1 Using the MANOVA command to run the univariate ANOVA 289 8.5.2 Including simple effects in a MANOVA command 291 8.6 HOW MANY PARTICIPANTS SHALL I NEED? 297 8.7 MORE COMPLEX EXPERIMENTS 298 8.7.1 Three-way interactions 298 8.7.2 The three-way ANOVA 299 8.7.3 How the three-way ANOVA works 301 8.7.4 The three-way ANOVA with SPSS 302 viii Contents 8.7.5 Follow-up analysis following a significant three-way interaction 304 8.7.6 Testing for simple interactions and simple, simple main effects 305 8.7.7 Unplanned multiple comparisons 308 8.8 A FINAL WORD 309 CHAPTER 9 Within subjects experiments 311 9.1 INTRODUCTION 311 9.1.1 Rationale of a within subjects experiment 311 9.1.2 How the within subjects ANOVA works 312 9.1.3 A within subjects experiment on the effect of target shape on shooting accuracy 315 9.1.4 Order effects: counterbalancing 316 9.1.5 Assumptions underlying the within subjects ANOVA: homogeneity of covariance 317 9.2 A ONE-FACTOR WITHIN SUBJECTS ANOVA WITH SPSS 318 9.2.1 Entering the data 319 9.2.2 Running the one-factor within subjects ANOVA 319 9.2.3 Output for a one-factor within subjects ANOVA 322 9.2.4 Effect size in the within subjects ANOVA 327 9.3 HOW MANY PARTICIPANTS SHALL I NEED? 328 9.4 NONPARAMETRIC EQUIVALENTS OF THE WITHIN SUBJECTS ANOVA 329 9.4.1 The Friedman test for ordinal data 329 9.4.2 Cochran’s Q test for nominal data 333 9.5 THE TWO-FACTOR WITHIN SUBJECTS ANOVA 336 9.5.1 Preparing the data set 337 9.5.2 Running the two-factor within subjects ANOVA 338 9.5.3 Output for a two-factor within subjects ANOVA 341 9.5.4 Unpacking a significant interaction with multiple comparisons 344 9.6 A FINAL WORD 346 CHAPTER 10 Mixed factorial experiments 349 10.1 INTRODUCTION 349 10.1.1 A mixed factorial experiment 349 10.1.2 Classifying mixed factorial designs 351 10.1.3 Rationale of the mixed ANOVA 351 10.2 THE TWO-FACTOR MIXED FACTORIAL ANOVA WITH SPSS 354 10.2.1 Preparing the SPSS data set 354 10.2.2 Running the ANOVA 355 10.2.3 Output for the two-factor mixed ANOVA 357 10.2.4 Simple effects analysis with syntax 364 10.3 THE THREE-FACTOR MIXED ANOVA 369 10.3.1 The two three-factor mixed designs 369 10.3.2 Two within subjects factors 370 10.3.3 Using syntax to test for simple effects 375 10.3.4 One within subjects factor and two between subjects factors: the A×B×(C) mixed factorial design 380 Contents ix 10.4 THE MULTIVARIATE ANALYSIS OF VARIANCE (MANOVA) 387 10.4.1 What the MANOVA does 387 10.4.2 How the MANOVA works 389 10.4.3 Assumptions of the MANOVA 392 10.4.4 Application of the MANOVA to the shape recognition experiment 393 10.5 A FINAL WORD 397 CHAPTER 11 Measuring statistical association 399 11.1 INTRODUCTION 399 11.1.1 A correlational study 400 11.1.2 Linear relationships 401 11.2 THE PEARSON CORRELATION 402 11.2.1 Formula for the Pearson correlation 402 11.2.2 The range of values of the Pearson correlation 403 11.2.3 The sign of a correlation 404 11.2.4 Testing an obtained value of r for significance 405 11.2.5 A word of warning about the correlation coefficient 405 11.2.6 Effect size 406 11.3 CORRELATION WITH SPSS 407 11.3.1 Preparing the SPSS data set 407 11.3.2 Obtaining the scatterplot 408 11.3.3 Obtaining the Pearson correlation 408 11.3.4 Output for the Pearson correlation 408 11.4 OTHER MEASURES OF ASSOCIATION 409 11.4.1 Spearman’s rank correlation 409 11.4.2 Kendall’s tau statistics 410 11.4.3 Rank correlations with SPSS 411 11.5 NOMINAL DATA 412 11.5.1 The one-sample approximate chi-square goodness-of-fit test with three or more categories 413 11.5.2 Running a chi-square goodness-of-fit test on SPSS 414 11.5.3 Measuring effect size following a chi-square test of goodness-of-fit 418 11.5.4 Testing for association between two qualitative variables in a contingency table 420 11.5.5 Analysis of contingency tables with SPSS 425 11.5.6 Getting help with the output 431 11.5.7 Some cautions and caveats 432 11.5.8 Other problems with traditional chi-square analyses 440 11.6 DO DOCTORS AGREE? COHEN’S KAPPA 440 11.7 PARTIAL CORRELATION 443 11.9 A FINAL WORD 447 CHAPTER 12 Regression 448 12.1 INTRODUCTION 448 12.1.1 Simple, two-variable regression 449 12.1.2 Residuals 450