98,99 €
Mehr erfahren.
- Herausgeber: John Wiley & Sons
- Kategorie: Wissenschaft und neue Technologien
- Sprache: Englisch
Features easy-to-follow insight and clear guidelines to perform data analysis using IBM SPSS®
Performing Data Analysis Using IBM SPSS® uniquely addresses the presented statistical procedures with an example problem, detailed analysis, and the related data sets. Data entry procedures, variable naming, and step-by-step instructions for all analyses are provided in addition to IBM SPSS point-and-click methods, including details on how to view and manipulate output.
Designed as a user’s guide for students and other interested readers to perform statistical data analysis with IBM SPSS, this book addresses the needs, level of sophistication, and interest in introductory statistical methodology on the part of readers in social and behavioral science, business, health-related, and education programs. Each chapter of Performing Data Analysis Using IBM SPSS covers a particular statistical procedure and offers the following: an example problem or analysis goal, together with a data set; IBM SPSS analysis with step-by-step analysis setup and accompanying screen shots; and IBM SPSS output with screen shots and narrative on how to read or interpret the results of the analysis.
The book provides in-depth chapter coverage of:
- IBM SPSS statistical output
- Descriptive statistics procedures
- Score distribution assumption evaluations
- Bivariate correlation
- Regressing (predicting) quantitative and categorical variables
- Survival analysis
- t Test
- ANOVA and ANCOVA
- Multivariate group differences
- Multidimensional scaling
- Cluster analysis
- Nonparametric procedures for frequency data
Performing Data Analysis Using IBM SPSS is an excellent text for upper-undergraduate and graduate-level students in courses on social, behavioral, and health sciences as well as secondary education, research design, and statistics. Also an excellent reference, the book is ideal for professionals and researchers in the social, behavioral, and health sciences; applied statisticians; and practitioners working in industry.
Sie lesen das E-Book in den Legimi-Apps auf:
Seitenzahl: 804
Veröffentlichungsjahr: 2013
Ähnliche
Table of Contents
Title Page
Copyright
Preface
Part 1: Getting Started With IBM SPSS®
Chapter 1: Introduction to IBM SPSS®
1.1 What is IBM SPSS?
1.2 Brief History
1.3 Types of IBM SPSS Files and File Name Extensions
Chapter 2: Entering Data in IBM SPSS®
2.1 The Starting Point
2.2 The Two Types of Displays
2.3 A Sample Data Set
2.4 The Variable View Display
2.5 Entering Specifications in the Variable View Display
2.6 Saving the Data File
2.7 Entering Specifications in the Data View Display
Chapter 3: Importing Data from Excel to IBM SPSS®
3.1 The Starting Point
3.2 The Importing Process
Part 2: Obtaining, Editing, and Saving Statistical Output
Chapter 4: Performing Statistical Procedures in IBM SPSS®
4.1 Overview
4.2 Using Dialog Windows to Setup the Analysis
4.3 The Output
Chapter 5: Editing Output
5.1 Overview
5.2 Changing the Wording of a Column Heading
5.3 Changing the Width of a Column
5.4 Viewing More Decimal Values
5.5 Editing Text in IBM SPSS Output Files
Chapter 6: Saving and Copying Output
6.1 Overview
6.2 Saving an Output File as an IBM SPSS Output File
6.3 Saving an Output File in Other Formats
6.4 Using Operating System Utilities to Copy an IBM SPSS Table to a Word Processing Document
6.5 Using the Copy and Paste Functions to Copy an IBM SPSS Output Table to a Word Processing Document
Part 3: Manipulating Data
Chapter 7: Sorting and Selecting Cases
7.1 Overview
7.2 Sorting Cases
7.3 Selecting Cases
Chapter 8: Splitting Data Files
8.1 Overview
8.2 The General Splitting Process
8.3 The Procedure to Split the Data File
8.4 The Data File after the Split
8.5 Statistical Analyses Under Split File
8.6 Resetting the Data File
Chapter 9: Merging Data from Separate Files
9.1 Overview
9.2 Adding Cases
9.3 Adding Variables
Part 4: Descriptive Statistics Procedures
Chapter 10: Frequencies
10.1 Overview
10.2 Numerical Example
10.3 Analysis Setup: Categorical Variables
10.4 Analysis Output: Categorical Variables
10.5 Analysis Setup: Quantitative Variables
10.6 Analysis Output: Quantitative Variables
Chapter 11: Descriptives
11.1 Overview
11.2 Numerical Example
11.3 Analysis Setup
11.4 Analysis Output
Chapter 12: Explore
12.1 Overview
12.2 Numerical Example
12.3 Analysis Setup
12.4 Analysis Output
Part 5: Simple Data Transformations
Chapter 13: Standardizing Variables to z Scores
13.1 Overview
13.2 Numerical Example
13.3 Analysis Setup
13.4 Analysis Output
13.5 Descriptive Statistics on Zneoneuro
13.6 Other Standard Scores
Chapter 14: Recoding Variables
14.1 Overview
14.2 Numerical Example
14.3 Analysis Strategy
14.4 Frequencies Analysis
14.5 Recoding an Original Variable Using Ranges
14.6 The Results of the Recoding
14.7 Recoding an Original Variable Using Individual Values
Chapter 15: Visual Binning
15.1 Overview
15.2 Numerical Example
15.3 Analysis Setup
Chapter 16: Computing New Variables
16.1 Overview
16.2 Computing an Algebraic Expression
16.3 The Outcome of Computing the Linear T Scores
16.4 Computing the Mean of a Set of Variables
16.5 Numerical Example of Computing the Mean of a Set of Variables
16.6 The Computation Process
16.7 The Outcome of Computing the Affiliation Subscale
Chapter 17: Transforming Dates to Age
17.1 Overview
17.2 The IBM SPSS® System Clock
17.3 Date Formats
17.4 The Date and Time Wizard
17.5 Using a Different Time Referent
17.6 Using Age in a Statistical Analysis
Part 6: Evaluating Score Distribution Assumptions
Chapter 18: Detecting Univariate Outliers
18.1 Overview
18.2 Numerical Example
18.3 Analysis Setup: Sample as a Whole
18.4 Analysis Output: Sample as a Whole
18.5 Analysis Setup: Considering the Categorical Variable of Sex
18.6 Analysis Output: Considering the Categorical Variable of Sex
Chapter 19: Detecting Multivariate Outliers
19.1 Overview
19.2 The Mahalanobis Distance
19.3 Numerical Example
19.4 Analysis Setup: Linear Regression
19.5 Analysis Output: Linear Regression
19.6 Strategies to Examine the Results
19.7 Examining the Data
Chapter 20: Assessing Distribution Shape: Normality, Skewness, and Kurtosis
20.1 Overview
20.2 Numerical Example
20.3 Analysis Strategy
20.4 Analysis Setup: Frequencies
20.5 Analysis Output: Frequencies
20.6 Analysis Set Up: Explore
20.7 Analysis Output: Explore
Chapter 21: Transforming Data to Remedy Statistical Assumption Violations
21.1 Overview
21.2 Numerical Example
21.3 Analysis Strategy
21.4 Analysis Setup: Frequencies of the Original doc_visits Variable
21.5 Analysis Output: Frequencies of the Original doc_visits Variable
21.6 Analysis Setup: Square Root Transformation
21.7 Analysis Setup: Log Base 10 Transformation
21.8 Analysis Setup: Reflected Inverse Transformation
21.9 Analysis Setup: Frequencies of All of the Transformed Variables
21.10 Analysis Output
Part 7: Bivariate Correlation
Chapter 22: Pearson Correlation
22.1 Overview
22.2 Numerical Example
22.3 Analysis Setup: Checking for Linearity
22.4 Analysis Output: Checking for Linearity
22.5 Analysis Setup: Correlating a Single Pair of Variables
22.6 Analysis Output: Correlating a Single Pair of Variables
22.7 Correlating Several Pairs of Variables
Chapter 23: Spearman Rho and Kendall Tau-b Rank-Order Correlations
23.1 Overview
23.2 The Spearman Rho Correlation
23.3 The Kendall Tau-b Correlation
23.4 Numerical Example Without Ties
23.5 Analysis Setup
23.6 Analysis Output
23.7 Numerical Example With Ties
23.8 Analysis Setup and Output
Part 8: Regressing (Predicting) Quantitative Variables
Chapter 24: Simple Linear Regression
24.1 Overview
24.2 Numerical Example
24.3 Analysis Setup
24.4 Analysis Output
24.5 The Y Intercept Issue
Chapter 25: Centering the Predictor Variable in Simple Linear Regression
25.1 Overview
25.2 Numerical Example
25.3 Analysis Strategy
25.4 Obtaining Descriptive Statistics on the Predictor Variable
25.5 Computing the Centered Predictor Variable
25.6 Analysis Setup: Simple Linear Regression Using BMI as the Predictor
25.7 Analysis Setup: Simple Linear Regression Using BMIcentered as the Predictor
25.8 Analysis Output From Both Regression Analyses
Chapter 26: Multiple Linear Regression
26.1 Overview
26.2 Numerical Example
26.3 Analysis Strategy
26.4 Analysis Setup: Standard Method
26.5 Analysis Output: Standard Method
26.6 Analysis Setup: Stepwise Method
26.7 Analysis Output: Stepwise Method
26.8 Analysis Setup: Automatic Linear Modeling
26.9 Analysis Output: Automatic Linear Modeling
Chapter 27: Hierarchical Linear Regression
27.1 Overview
27.2 Numerical Example and Analysis Strategy
27.3 Analysis Setup
27.4 Analysis Output
Chapter 28: Polynomial Regression
28.1 Overview
28.2 Numerical Example
28.3 Analysis Strategy
28.4 Obtaining the Scatterplot
28.5 Computing the Polynomial Variables
28.6 Analysis Setup: Linear Regression
28.7 Analysis Output: Linear Regression
Chapter 29: Multilevel Modeling
29.1 Overview
29.2 Numerical Example
29.3 Analysis Strategy
29.4 Aggregating the Optimism Variable
29.5 Centering the Level 2 optimismgroupmean Variable
29.6 Analysis Setup: Unconditional Model
29.7 Analysis Output: Unconditional Model
29.8 Analysis Setup: Mixed Level 1 Model
29.9 Analysis Output: Mixed Level 1 Model
29.10 Analysis Setup: Mixed Level 2 Model
29.11 Analysis Output: Mixed Level 2 Model
29.12 Analysis Setup: Hierarchical Model
29.13 Analysis Output: Hierarchical Model
29.14 Analysis Setup: Interaction Model
29.15 Analysis Output: Interaction Model
Part 9: Regressing (Predicting) Categorical Variables
Chapter 30: Binary Logistic Regression
30.1 Overview
30.2 Numerical Example
30.3 Analysis Setup
30.4 Analysis Output
Chapter 31: ROC Analysis
31.1 Overview
31.2 Numerical Example
31.3 Analysis Strategy
31.4 Binary Logistic Regression Analysis: Default Classification Cutoff
31.5 ROC Analysis: Setup
31.6 ROC Analysis: Output
31.7 Binary Logistic Regression Analysis: Revised Classification Cutoff
Chapter 32: Multinominal Logistic Regression
32.1 Overview
32.2 Numerical Example
32.3 Analysis Setup
32.4 Analysis Output
Part 10: Survival Analysis
Chapter 33: Survival Analysis: Life Tables
33.1 Overview
33.2 Numerical Example
33.3 Analysis Setup
33.4 Analysis Output
Chapter 34: The Kaplan–Meier Survival Analysis
34.1 Overview
34.2 Numerical Example
34.3 Analysis Strategy
34.4 Analysis Setup: Comparing Males and Females
34.5 Analysis Output: Comparing Males and Females
34.6 Analysis Setup: Comparing Males and Females with Stratification
34.7 Analysis Output: Comparing Males and Females with Stratification
Chapter 35: Cox Regression
35.1 Overview
35.2 Numerical Example
35.3 Analysis Setup
35.4 Analysis Output
Part 11: Reliability as a Gauge of Measurement Quality
Chapter 36: Reliability Analysis: Internal Consistency
36.1 Overview
36.2 Numerical Example
36.3 Analysis Setup
36.4 Analysis Output
Chapter 37: Reliability Analysis: Assessing Rater Consistency
37.1 Overview
37.2 Numerical Example: ICC
37.3 Analysis Setup: ICC
37.4 Analysis Output: ICC
37.5 Numerical Example: Kappa
37.6 Analysis Setup: Kappa
37.7 Analysis Output: Kappa
Part 12: Analysis of Structure
Chapter 38: Principal Components and Factor Analysis
38.1 Overview of Principal Components and Factor Analysis
38.2 Numerical Example
38.3 A Starting Place
38.4 Analysis Setup: Preliminary Analysis
38.5 Analysis Output: Preliminary Analysis
38.6 Our Analysis Strategy for the Main Analyses
38.7 Analysis Setup for the Four-Factor Structure
38.8 Analysis Output for the Four-Component/Factor Structure
38.9 The Three-Component/Factor Structure
38.10 Determining Which Solution to Accept
Chapter 39: Confirmatory Factor Analysis
39.1 Overview
39.2 Numerical Example
39.3 Drawing the Model
39.4 Analysis Setup
39.5 Analysis Output
39.6 Analysis Setup: Modified Model
39.7 Analysis Output: Modified Model
Part 13: Evaluating Causal (Predictive) Models
Chapter 40: Simple Mediation
40.1 Overview
40.2 Numerical Example
40.3 Analysis Strategy
40.4 The Independent Variable Predicting the Mediator Variable
40.5 The Independent Variable and the Mediator Predicting the Outcome Variable
40.6 The Unmediated Model with the Independent Variable Predicting the Dependent Variable
40.7 Consolidating the Results of the Mediation Model
40.8 Testing the Statistical Significance of the Indirect Effect
40.9 Testing the Statistical Significance of the Difference between the Direct Paths in the Unmediated and the Mediated Models
40.10 Determining the Relative Strength of the Mediated Effect
Chapter 41: Path Analysis Using Multiple Regression
41.1 Overview
41.2 Numerical Example
41.3 Analysis Strategy
41.4 The “Flat” Multiple Regression Analysis: Setup
41.5 The “Flat” Multiple Regression Analysis: Output
41.6 Path Analysis using Multiple Regression: Analysis 1
41.7 Path Analysis using Multiple Regression: Analysis 2
41.8 Path Analysis using Multiple Regression: Synthesis
41.9 Assessing the Statistical Significance of the Indirect Effects
41.10 Assessing the Strength of Each Indirect Effect
41.11 Evaluating the Possibility of Mediation
41.12 Testing the Statistical Significance of the Difference between the Direct Paths in the Unmediated and the Mediated Models
Chapter 42: Path Analysis Using Structural Equation Modeling
22.1 Overview
22.2 Path Analysis Based on SEM: Drawing the Model
22.3 Path Analysis based on SEM: Analysis Setup
22.4 Path Analysis based on SEM: Analysis Output
22.5 Path Analysis Based on SEM: Modified Model Output
Chapter 43: Structural Equation Modeling
43.1 Overview
43.2 Numerical Example
43.3 Analysis Strategy
43.4 Evaluating the Measurement Model: Drawing the Model
43.5 Evaluating the Measurement Model: Analysis Setup
43.6 Evaluating the Measurement Model: Analysis Output
43.7 Evaluating the Structural Model: Drawing the Model
43.8 Evaluating the Structural Model: Analysis Setup
43.9 Evaluating the Structural Model: Analysis Output
43.10 Evaluating the Structural Model: Synthesis
43.11 The Strategy to Configure and Analyze a Trimmed Model
43.12 Examining the Direct Effect of Efficacy on Statistics in Isolation
43.13 Examining the Mediated Effect of Efficacy on Statistics through Science
43.14 Synthesis of the Trimmed (Mediated) Model Results
43.15 Statistical Significance of the Indirect Effect: The Aroian Test
43.16 Comparing the Direct Effects of Efficacy on Statistics in the Simple Model and the Mediated Model: The Freedman-Schatzkin Test
43.17 The Relative Strength of the Mediated Effect
Part 14: Test
Chapter 44: One-Sample t Test
44.1 Overview
44.2 Numerical Example
44.3 Analysis Setup
44.4 Analysis Output
Chapter 45: Independent-Samples t Test
45.1 Overview
45.2 Numerical Example: Meeting the Homogeneity of Variance Assumption
45.3 Analysis Setup: Meeting the Homogeneity of Variance Assumption
45.4 Analysis Output: Meeting the Homogeneity of Variance Assumption
45.5 Magnitude of the Mean Difference
45.6 Numerical Example: Violating the Homogeneity of Variance Assumption
45.7 Analysis Setup: Violating the Homogeneity of Variance Assumption
45.8 Analysis Output: Violating the Homogeneity of Variance Assumption
Chapter 46: Paired-Samples t Test
46.1 Overview
46.2 Numerical Example
46.3 Analysis Setup
46.4 Analysis Output
46.5 Magnitude of the Mean Difference
Part 15: Univariate Group Differences: ANOVA and ANCOVA
Chapter 47: One-Way Between-Subjects ANOVA
47.1 Overview
47.2 Numerical Example
47.3 Analysis Strategy
47.4 Analysis Setup
47.5 Analysis Output
Chapter 48: Polynomial Trend Analysis
48.1 Overview
48.2 Numerical Example
48.3 Analysis Strategy
48.4 Analysis Setup
48.5 Analysis Output
Chapter 49: One-Way Between-Subjects ANCOVA
49.1 Overview
49.2 Numerical Example
49.3 Analysis Strategy
49.4 Analysis Setup: ANOVA
49.5 Analysis Output: ANOVA
49.6 Evaluating the ANCOVA Assumptions
49.7 Analysis Setup: ANCOVA
49.8 Analysis Output: ANCOVA
Chapter 50: Two-Way Between-Subjects ANOVA
50.1 Overview
50.2 Numerical Example
50.3 Analysis Setup
50.4 Analysis Output: Omnibus Analysis
50.5 Analysis Output: Simple Effects Tests
Chapter 51: One-Way Within-Subjects ANOVA
51.1 Overview
51.2 Numerical Example
51.3 Analysis Setup
51.4 Analysis Output
Chapter 52: Repeated Measures Using Linear Mixed Models
52.1 Overview
52.2 Numerical Example
52.3 Analysis Strategy
52.4 Restructuring the Data File
52.5 Analysis Setup: Autoregressive Covariance Structure
52.6 Analysis Output: Autoregressive Covariance Structure
52.7 Analysis: Compound Symmetry
52.8 Analysis: Unstructured Covariance
Chapter 53: Two-Way Mixed ANOVA
53.1 Overview
53.2 Numerical Example
53.3 Analysis Setup
53.4 Analysis Output
Part 16: Multivariate Group Differences: MANOVA and Discriminant Function Analysis
Chapter 54: One-Way Between-Subjects MANOVA
54.1 Overview
54.2 Numerical Example
54.3 Correlation Analysis
54.4 Analysis Setup: Manova
54.5 Analysis Output: MANOVA
Chapter 55: Discriminant Function Analysis
55.1 Overview
55.2 Numerical Example
55.3 Analysis Setup
55.4 Analysis Output
Chapter 56: Two-Way Between-Subjects MANOVA
56.1 Overview
56.2 Numerical Example
56.3 Analysis Setup
56.4 Analysis Output
Part 17: Multidimensional Scaling
Chapter 57: Multidimensional Scaling: Classical Metric
57.1 Overview
57.2 Numerical Example
57.3 Analysis Setup
57.4 Analysis Output
Chapter 58: Multidimensional Scaling: Metric Weighted
58.1 Overview
58.2 Numerical Example
58.3 Analysis Setup
58.4 Analysis Output
Part 18: Cluster Analysis
Chapter 59: Hierarchical Cluster Analysis
59.1 Overview
59.2 Numerical Example
59.3 Analysis Setup
59.4 Analysis Output
Chapter 60: k-Means Cluster Analysis
60.1 Overview
60.2 Numerical Example: -Means Clustering
60.3 Analysis Strategy
60.4 Transforming Cluster Variables to z Scores
60.5 Analysis Setup: -Means Clustering
60.6 Analysis Output: -Means Clustering
60.7 Follow-Up One-Way Between-Subjects Analysis
60.8 Numerical Example: One-Way ANOVA
60.9 Analysis Setup: One-Way ANOVA
60.10 Analysis Output: One-Way ANOVA
Part 19: Nonparametric Procedures for Analyzing Frequency Data
Chapter 61: Single-Sample Binomial and Chi-Square Tests: Binary Categories
61.1 Overview
61.2 Numerical Example
61.3 Analysis Strategy
61.4 Frequencies Analysis
61.5 Analysis Setup
61.6 Analysis Output
Chapter 62: Single-Sample (One-Way) Multinominal Chi-Square Tests
62.1 Overview
62.2 Numerical Example
62.3 Analysis Strategy
62.4 Frequencies Analysis
62.5 Analysis Setup: Omnibus Analysis
62.6 Analysis Output: Omnibus Analysis
62.7 Analysis Setup: Comparison of Categories 1 and 2
62.8 Analysis Output: Comparison of Categories 1 and 2
62.9 Analysis Setup: Comparison of Categories 1 and 3
62.10 Analysis Output: Comparison of Categories 1 and 3
62.11 Analysis Setup: Comparison of Categories 2 and 3
62.12 Analysis Output: Comparison of Categories 2 and 3
Chapter 63: Two-Way Chi-Square Test of Independence
63.1 Overview
63.2 Analysis Strategy
63.3 Numerical Example: 2 × 2 Chi-Square
63.4 Analysis Setup: 2 × 2 Chi-Square
63.5 Analysis Output: 2 × 2 Chi-Square
63.6 Numerical Example: 4 × 2 Chi-Square
63.7 Analysis Setup: 4 × 2 Chi-Square
63.8 Analysis Output: 4 × 2 Chi-Square
Chapter 64: Risk Analysis
64.1 Overview
64.2 Numerical Example
64.3 Analysis Setup
64.4 Analysis Output
Chapter 65: Chi-Square Layers
65.1 Overview
65.2 Numerical Example
65.3 Analysis Setup
65.4 Analysis Output
Chapter 66: Hierarchical Loglinear Analysis
66.1 Overview
66.2 Numerical Example
66.3 Analysis Setup
66.4 Analysis Output
66.5 The Next Steps
Appendix: Statistics Tables
References
Author Index
Subject Index
Copyright © 2013 by John Wiley & Sons, Inc. All rights reserved.
Published by John Wiley & Sons, Inc., Hoboken, New Jersey.
Published simultaneously in Canada.
No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, photocopying, recording, scanning, or otherwise, except as permitted under Section 107 or 108 of the 1976 United States Copyright Act, without either the prior written permission of the Publisher, or authorization through payment of the appropriate per-copy fee to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, (978) 750–8400, fax (978) 750–4470, or on the web at www.copyright.com. Requests to the Publisher for permission should be addressed to the Permissions Department, John Wiley & Sons, Inc., 111 River Street, Hoboken, NJ 07030, (201) 748–6011, fax (201) 748–6008, or online at http://www.wiley.com/go/permission.
Limit of Liability/Disclaimer of Warranty: While the publisher and author have used their best efforts in preparing this book, they make no representations or warranties with respect to the accuracy or completeness of the contents of this book and specifically disclaim any implied warranties of merchantability or fitness for a particular purpose. No warranty may be created or extended by sales representatives or written sales materials. The advice and strategies contained herein may not be suitable for your situation. You should consult with a professional where appropriate. Neither the publisher nor author shall be liable for any loss of profit or any other commercial damages, including but not limited to special, incidental, consequential, or other damages.
For general information on our other products and services or for technical support, please contact our Customer Care Department within the United States at (800) 762–2974, outside the United States at (317) 572–3993 or fax (317) 572–4002.
Wiley also publishes its books in a variety of electronic formats. Some content that appears in print may not be available in electronic formats. For more information about Wiley products, visit our web site at www.wiley.com.
Library of Congress Cataloging-in-Publication Data:
Meyers, Lawrence S.
Performing data analysis using IBM SPSS® / Lawrence S. Meyers,
Department of Psychology, California State University, Sacramento,
Sacramento, CA, Glenn C. Gamst, Department of Psychology, University of La
Verne, La Verne, CA, A. J. Guarino, Department of Biostatistics, MGH
Institute of Health Professions, Boston, MA.
pages cm
Includes bibliographical references and index.
ISBN 978-1-118-35701-9 (pbk.) – ISBN 978-1-118-51494-8 – ISBN
978-1-118-51492-4 (ePDF) – ISBN 978-1-118-51493-1 (ePub) – ISBN
978-1-118-51490-0 1. Social sciences–Statistical methods–Computer
programs. 2. SPSS (Computer file) I. Title.
HA32.M4994 2013
005.5'5–dc23
2013002844
Preface
The IBM SPSS® software package is one of the most widely used statistical applications in academia, business, and government. This book, Performing Data Analysis Using IBM SPSS, provides readers with both a gentle introduction to basic statistical computation with the IBM SPSS software package and a portal to the more comprehensive and statistically robust multivariate procedures. This book was written to be a stand-alone resource as well as a supplementary text for both undergraduate introductory and more advanced graduate-level statistics courses.
For most of the chapters, we provide a consistent structure that includes the following:
Overview:
This is a brief conceptual introduction that furnishes a set of relevant details for each statistical procedure being covered, including a few useful references that supply additional background information.
Numerical Example:
This includes a description of the research problem or question, the name of the data file, a description of the variables and how they are coded, and (often) a screenshot of the IBM SPSS Data View.
Analysis Strategy:
When the analysis is performed in stages, or when alternative data processing strategies are available, we include a description of how we have structured our data analysis and explain the rationale for why we have performed the analyses in the way presented in the chapter.
Analysis Setup:
This includes how to configure each dialog window with screenshots and is accompanied (within reason) with explanations for why we chose the particular options we utilized.
Analysis Output:
This elucidates the major aspects of the statistical output with pertinent screenshots and discussion.
Because of the multiple audience we are attempting to reach with this book, the complexity of the procedures covered varies substantially across the chapters. For example, chapters that cover IBM SPSS basics of data entry and file manipulation, descriptive statistical procedures, correlation, simple linear regression, multiple regression, one-way chi-square, t tests, and one and two-way analysis of variance designs are all appropriate topics for first- or second-level statistics and data analysis courses. The remaining chapters, data transformations, assumption violation assessment, reliability analysis, logistic regression, multivariate analysis of variance, survival analysis, multidimensional scaling, cluster analysis, multilevel modeling, exploratory and confirmatory factor analysis, and structural equation modeling, are all important topics that may be suitable for more advanced statistics courses.
There are 66 chapters in this book. They are organized into 19 sections or “Parts.” Different authors might organize the chapters in somewhat different ways and present them in a somewhat different order, as there is no fully agreed upon organizational structure for this material. However, except for the chapters presented in the early parts that show readers how to work with IBM SPSS data files, most of the data analysis chapters can be used as a resource on their own, allowing users to work with whatever analysis procedures meet their needs; the order in which users would choose to work with the chapters is really a function of the foundations on which the material is based (e.g., users should undertake structural equation modeling only after acquiring some familiarity with regression techniques and factor analysis).
Part 1, “Getting Started With IBM SPSS®,” consists of three chapters that provide the basics of IBM SPSS. Chapter 1 provides an introduction to IBM SPSS, Chapter 2 describes how to enter data, and Chapter 3 demonstrates how to import data from Excel to IBM SPSS.
Part 2, “Obtaining, Editing, and Saving Statistical Output,” consists of three chapters that describe ways of manipulating the IBM SPSS statistical output. Chapter 4 conveys how to perform a statistical procedure. Chapter 5 demonstrates how to edit statistical output. Chapter 6 provides information on saving and copying output.
Part 3, “Manipulating Data,” contains three chapters that focus on how to organize existing data. Chapter 7 examines the sorting and selecting of cases. Chapter 8 demonstrates how to split a data file. Chapter 9 discusses how to merge cases and variables.
Part 4, “Descriptive Statistics Procedures,” consists of three chapters that provide descriptive statistical summary capabilities. Chapter 10 focuses on the analysis of frequency counts for categorical variables. Chapter 11 describes how to compute measures of central tendency and variability. Chapter 12 provides additional options to examine variables in the data file.
Part 5, “Simple Data Transformations,” consists of five chapters that demonstrate how to manipulate variables. Chapter 13 describes how to standardize a variable through the creation of z scores. Chapter 14 demonstrates how to recode the values of a variable. Chapter 15 provides a discussion of visual binning used in categorizing data. Chapter 16 demonstrates how to compute a new variable from existing data, and Chapter 17 shows how to transform data into time variables.
Part 6, “Evaluating Score Distribution Assumptions,” consists of four chapters that examine the assumptions underlying most of the statistical procedures covered in the book. Chapter 18 focuses on the detection of univariate outliers, and Chapter 19 examines their multivariate counterpart. Chapter 20 focuses on the assessment of normality, and Chapter 21 demonstrates how to remedy assumption violations through data transformation.
Part 7, “Bivariate Correlation,” consists of two chapters dealing with correlation. Chapter 22 demonstrates how to perform a Pearson product moment correlation (r), and Chapter 23 depicts how to compute a Spearman rho and Kendall tau-b correlation.
Part 8, “Regressing (Predicting) Quantitative Variables,” consists of six chapters dealing with simple and multiple regression and multilevel modeling. Chapter 24 covers simple linear regression. Chapter 25 demonstrates how to center a predictor variable. Chapter 26 covers multiple linear regression. Chapter 27 covers hierarchical linear regression. Chapter 28 describes polynomial (curve estimation) regression. Chapter 29 provides an introduction to multilevel modeling.
Part 9, “Regressing (Predicting) Categorical Variables,” consists of three chapters that deal with logistic regression. Chapter 30 covers binary logistic regression. Chapter 31 demonstrates ROC (receiver operator curve) analysis. Chapter 32 examines multinomial logistic regression.
Part 10, “Survival Analysis,” consists of three chapters that depict various types of survival analysis. Chapter 33 demonstrates life table analysis. Chapter 34 covers the Kaplan–Meier procedure. Chapter 35 demonstrates the Cox regression procedure.
Part 11, “Reliability as a Gauge of Measurement Quality,” is covered in two chapters. Chapter 36 covers reliability analyses related to issues of internal consistency. Chapter 37 covers reliability analyses that focus on inter-rater reliability.
Part 12, “Analysis of Structure,” is covered in two chapters and deals with various types of factor analysis. Chapter 38 covers principal components analysis and factor analysis, and Chapter 39 covers confirmatory factor analysis.
Part 13, “Evaluating Causal (Predictive) Models,” contains four chapters that deal with model building, as it pertains to mediation analysis and structural equation modeling. Chapter 40 covers simple mediation analysis. Chapters 41 and 42 cover path analysis using multiple regression and Amos, respectively. Chapter 43 provides an introduction to structural equation modeling.
Part 14, “t Test,” consists of three chapters that cover various types of t tests. Chapter 44 demonstrates how to conduct a single sample t test. Chapter 45 covers the independent groups t test, and Chapter 46 covers the correlated samples t test.
Part 15, “Univariate Group Differences: ANOVA and ANCOVA,” consists of seven chapters that cover various one- and two-way analyses of variance procedures. Chapter 47 demonstrates the one-way between-subjects ANOVA using the IBM SPSS GLM (general linear model) procedure. Chapter 48 demonstrates a trend analysis using polynomial contrasts. Chapter 49 covers one-way between-subjects ANCOVA (analysis of covariance). Chapter 50 examines two-way between-subjects ANOVA. Chapter 52 covers one-way repeated linear mixed models, and Chapter 53 examines the two-way simple mixed design.
Part 16, “Multivariate Group Differences: MANOVA and Discriminant Function Analysis,” covers multivariate analysis of variance (MANOVA) and discriminant function analysis. Chapter 54 examines how to conduct a one-way between-subjects MANOVA. Chapter 55 covers discriminant function analysis, and Chapter 56 describes how to compute a two-way between-subjects MANOVA.
Part 17, “Multidimensional Scaling,” consists of two multidimensional scaling chapters. Chapter 57 describes multidimensional scaling using the classic metric approach, while Chapter 58 describes multidimensional scaling using the individual differences scaling approach.
Part 18, “Cluster Analysis,” consists of two cluster analysis chapters. Chapter 59 demonstrates hierarchical cluster analysis, while Chapter 60 depicts the k-means approach.
Part 19, “Nonparametric Procedures for Analyzing Frequency Data,” completes the book and consists of six chapters dealing with nonparametric statistical procedures. Chapter 61 covers the binomial test, and Chapter 62 covers the one-way chi-square test. Chapter 63 demonstrates the two-way chi-square test with observed versus expected frequencies. Chapter 64 demonstrates how to do a risk analysis, and Chapter 65 covers the chi-square layers procedure. Lastly, Chapter 66 demonstrates hierarchical log-linear analysis.
Entering ISBN 9781118357019 at booksupport.wiley.com allows users to access the IBM SPSS data sets that were used in each of the chapters. These files can be downloaded, and users can shadow our analyses of the data on their own computers, assuming that they have the IBM SPSS software on such systems.
Because this book has multiple intended audiences, we recommend several different reading strategies. For the beginning IBM SPSS user, we suggest a very careful reading of Chapters 1 through 9, before moving into Chapters 10, 11, 12, 22, 24, 45, 47, 50, and 62. For the advanced undergraduate student, graduate student, or researcher, the remaining chapters should be pursued as the need arises.
Part 1
Getting Started with IBM SPSS®
Chapter 1: Introduction to IBM SPSS®
1.1 What is IBM SPSS?
IBM SPSS is a computer statistical software package. This software can perform many types of data-oriented tasks such as recoding a variable (e.g., “flipping” the values of a reverse-worded survey item). It will perform these tasks for each case in the data set, even if there are tens of thousands of cases (a daunting job to perform by hand). IBM SPSS can also perform a huge range of statistical procedures, ranging from computing simple descriptive statistics such as the mean, standard deviation, and standard error of the mean, through some fundamental procedures such as correlation and linear regression, to a variety of multivariate procedures such as factor analysis, discriminant function analysis, and multidimensional scaling.
SPSS at one time was an acronym for Statistical Package for the Social Sciences but it is now treated as just a familiar array of letters. This is just as well, as researchers from a wide array of disciplines, not just those in the social sciences, use this software. Relatively recently, IBM purchased SPSS and beginning with version 19 has officially renamed the software as IBM SPSS.
1.2 Brief History
As described by Gamst, Meyers, and Guarino (2008), in the long-ago days, users did not have the luxury of pointing and clicking but instead actually typed syntax (SPSS computer code) as well as their data onto rectangular computer cards that were then physically read into a very large mainframe computer. Eventually, the cards gave way to computer terminals where users would type their data together with the syntax to structure their analysis via a keyboard and CRT (cathode ray tube) screen. The software finally reached the relatively early personal computers (PCs) in the middle 1980s, and it has gained considerable sophistication over the years.
As the program developed, one aspect has remained consistent: the statistical procedures are still driven by syntax. As we interact with the dialog windows, IBM SPSS is actually converting our actions and selections into its own code (syntax).
1.3 Types of IBM SPSS Files and File Name Extensions
There are three kinds of files with which we ordinarily work when using IBM SPSS: data files, output files, and syntax files. We constantly deal with data and output files; more seasoned users also use syntax files extensively. Each file type has its own file name extension and distinctive icon, as shown in Figure 1.1. We discuss data files in Chapter 2 and output files in Chapter 4 and leave any discussion of syntax files for more specialized applications in some of the later topics covered in the book (e.g., performing simple effects in analysis of variance). The story on each file type in simplified form is as follows:
Data File.
This is a spreadsheet containing the data that were collected from the participating entities or
cases
(e.g., students in a university class, patients in a clinic, retail stores in a national chain). In the data file, the variables are represented as columns; cases, as rows. This file type uses the extension .
sav
and its icon shows a grid.
Output File.
This file is produced when IBM SPSS has performed the requested statistical analysis (or other operations such as saving the data file). It contains the results of the procedure. This file type uses the extension .
spv
and its icon shows a window with a banner.
Syntax File.
This file contains the IBM SPSS computer code (syntax) that drives the analysis. This file type uses the extension .
sps
and its icon shows a window with horizontal lines.
Figure 1.1 The three types of IBM SPSS files with which we ordinarily work: data files (.sav), output files (.spv), and syntax files (.sps).
If the extensions do not show on your screen, here is what can be done to show the file extensions. If you are using Windows 7
select
Control Panel → Folder Options → View Tab
;
uncheck the checkbox for
Hide extensions for known file types
;
click
OK
.
Here is what that can be done to show the file extensions in Mac OS X:
Select
Finder → Preferences → Advanced Tab
.
Check the checkbox for
Show all filename extensions
.
Close the window.
Chapter 2: Entering Data in IBM SPSS®
2.1 The Starting Point
When opening the IBM SPSS software program, we are presented with the view shown in Figure 2.1. We can navigate to an existent file, run the tutorial, type in data, and so on. By selecting the choice Type in data or by selecting Cancel, we can reach the IBM SPSS spreadsheet (the Data View display). We will select Cancel.
Figure 2.1 The screen presented on opening IBM SPSS.
2.2 The Two Types of Displays
The spreadsheet that is initially displayed is shown in Figure 2.2. This view, which is the default display, is called by IBM SPSS the Data View because it is, quite literally, where we enter and view our data. But as shown in Figure 2.2, it is also possible to display the Variable View. Whether we are entering our own data or importing an already constructed data set (as described in Chapter 3), we will need to work in both the Data View and the Variable View screens. Although we can deal with these screens in any order, we strongly encourage those new to IBM SPSS to begin with the Variable View screen when entering a new data set.
Figure 2.2 The initial spreadsheet is presented in the Data View display.
2.3 A Sample Data Set
Figure 2.3 shows a very simple set of fictional results of a research study just to illustrate how to go through the steps of entering data. The variables and their meaning are as follows:
ID.
This is an arbitrary identification code associated with each research participant (case). The ID de-identifies participants, thus protecting their anonymity and guaranteeing confidentiality. The ID also allows us to review the original data (which should also contain the identification codes) if data entry questions or errors occurred.
Gender.
This indicates the gender of the participant; in
Figure 2.3
, M stands for male and F stands for female.
Extraversion.
A personality characteristic indicating, roughly speaking, the degree to which the person is outgoing. In this study, it is represented by a 10-point scale with 1 indicating
very low
and 10 indicating
very high
.
Sales in Thousands of Dollars.
This indicates sales figures for each participant (salesperson) during a given month in a given department of a large retail chain.
Figure 2.3 Raw data that are to be entered in IBM SPSS.
2.4 The Variable View Display
Selecting Variable View at the bottom of the new (blank) spreadsheet gives rise to the screen shown in Figure 2.4. This display is editable and allows us to specify the various properties of the variables in the data file. The columns address the following variable specifications:
Name.
A reasonably short but descriptive name of the variable. There is a 64-character maximum for English, but we ordinarily want very much shorter names. No spaces or special characters are allowed but underscores can be used. We suggest using letters and numbers only.
Type.
There are several types of data (e.g., scientific notation, date, string) that IBM SPSS can read, but we will restrict ourselves in this example to
Numeric
(i.e., regular numbers), which is the default.
Width.
This is the number of spaces the data occupy. The default is
8
.
Decimals.
This is the number of decimal places shown in the
Data View
for that variable. The default is
2
. Note that IBM SPSS computes values to 16 decimals no matter what; if we ask to see fewer, as most of us do, the displayed values will be rounded to the number of decimals specified here.
Label.
A phrase to describe the variable. This is often omitted by researchers if the variable name is sufficiently descriptive.
Values.
When entering information on a categorical variable (such as gender), it is appropriate for most of the analyses we cover in the book to use arbitrary numeric codes for the categories. This is the place that we can (and absolutely should) specify labels for each category code.
Missing.
We can designate a missing value by either leaving the cell empty in the process of data entry or using an arbitrary numeric code. Arbitrary numeric codes are useful when there are different reasons for defining a value as missing (e.g., the original source is not legible, there is a double answer) permitting us to differentiate why a value may be missing.
Columns.
This is the number of spaces the data are allowed to occupy. The default is
8
.
Align.
This specifies right, left, or center alignment in the
Data View
display. The default is
Right
.
Measure.
This specifies the scale of measurement of the variable. The options are
Scale
(representing approximately interval-level data at a minimum),
Ordinal
(containing only less than, equal to, and greater than information), and
Nominal
(representing categorical data), with the default shown initially as
Unknown
.
Role.
There are a variety of roles that can be assigned to variables, but we will restrict ourselves to
Input
, which is the default (and thus permits variables to be placed in all of our analyses).
Figure 2.4 The Variable View display.
2.5 Entering Specifications in the Variable View Display
In the Variable View window, double-click the cell under Name in the first row and type in the name of the first variable (ID). Then click in the cell directly below to allow IBM SPSS to fill in its defaults in the first row. This is shown in Figure 2.5. Modify the default specifications as follows:
Click the
Decimals
cell and change the specification to 0 by clicking on the down toggle.
Click the
Measure
cell and select
Scale
.
The finished first row for the ID variable is shown in Figure 2.6. We have repeated this process for the other three variables, and the completed specifications thus far are shown in Figure 2.7. Note the following (also shown in Figure 2.7):
We used a short
Name
for the dollar amount of sales (
sales
) and so supplied a
Label
to provide a more complete description of the variable.
The variable
gender
is a categorical (nominal) variable and we have indicated that in the
Measure
column;
extraversion
and
sales
are quantitative variables called by IBM SPSS as
Scale
measures.
The only job remaining is to specify the value labels for the gender variable. It is a categorical (nominal) variable and we will use the values (numerical codes) 1 for female and 2 for male. Such codes are arbitrary and idiosyncratic to each researcher and thus need to be specified so that (a) other researchers can understand the data and (b) these value labels can appear in some of our output.
Figure 2.5 The Variable View display with the first variable typed in and the defaults showing.
Figure 2.6 The Variable View display with all of the variables typed in, the defaults modified, and the first step in specifying the gender codes showing.
Figure 2.7 The variables now have most of their specifications.
To accomplish this, we click the cell under Values on the row for gender. This produces the Value Labels dialog window. Follow these steps as shown in Figure 2.8 to provide the specifications:
Type
1
in the
Value
panel.
Type
female
in the
Label
panel.
Click
Add
.
Type
2
in the
Value
panel.
Type
male
in the
Label
panel.
Click
Add
.
To register the labels with IBM SPSS, we click OK. The labels are now contained in the Values cell for gender (see Figure 2.9).
Figure 2.8 The Value Labels window with the gender codes entered.
Figure 2.9 The Variable View display with the specifications for all of the variables now complete.
2.6 Saving the Data File
We now do something that all users should do on a very frequent basis: we will save the data file (and will do so every time we make any modification to it). Select File Save As (or select the Save File icon shown in Figure 2.9). This opens a standard file-saving dialog screen in the operating system (e.g., Windows 7, Mac OS X). Navigate to the desired location, name the file in some way that makes sense, and save it. (IBM SPSS will display an acknowledgment of the saving operation—we close that acknowledgment window without saving it.) The name of the file will appear in the banner of the IBM SPSS window; double-clicking the file icon in the directory will directly open the file next time.
2.7 Entering Specifications in the Data View Display
With the specifications now in place and the file saved, the variable names are visible after selecting Data View and the file name is now shown in the banner (see Figure 2.10). In the Data View spreadsheet, we simply enter our numbers, using the arrow keys or tab key to move from cell to cell. Once the data have been entered as shown in Figure 2.10, we again save the data file (clicking the Save File icon will overwrite the older file with the modified file). A “dimmed” (“grayed”) icon indicates that the current version of the file as shown on the screen is saved. As soon as we make any changes to it, the icon becomes “normal” and signifies that what we see on the screen is not currently saved. In Figure 2.10, the icon is dim indicating that we have not modified the data file since the last time it was saved.
Figure 2.10 The Data View with the data entered and the file saved.
Chapter 3: Importing Data from Excel to IBM SPSS®
3.1 The Starting Point
In addition to entering data directly into IBM SPSS, it is possible to bring in (import) data that have been entered in another software spreadsheet. We illustrate this using Excel. The data set used in Chapter 2 was entered into an Excel worksheet and is shown in Figure 3.1. We will import this data set into IBM SPSS.
Figure 3.1 Excel spreadsheet containing the data set we wish to import.
3.2 The Importing Process
As described in Section 2.1, open IBM SPSS and reach the Data View display of a blank (new) data file. Select File → Open → Data to reach the Open Data dialog window shown in Figure 3.2. From the Files of type drop-down menu, select Excel; these files have a base .xls extension, some with different extra letters depending on the version of Excel (e.g., .xlsx, .xlsm). Then navigate through the storage drives to locate the Excel file containing the data (sales extraversion.xls). Selecting (clicking) its name in the panel will cause the file name to appear in the File name panel. This is also shown in Figure 3.2.
Figure 3.2 The Open Data dialog window.
With the name and type of file now identified, select Open. This produces the Opening Excel Data Source window shown in Figure 3.3. The checkbox for Read variable names from the first row of data is already checked as a default. We keep it that way because our Excel spreadsheet contains the variable names in the appropriate SPSS format. If the Excel file variable name is not in the acceptable SPSS format, SPSS will assign VAR0001, VAR0002, and so on to the variables. Clicking OK initiates the importing process, the result of which is shown in Figure 3.4.
Figure 3.3 The Opening Excel Data Source dialog window.
Figure 3.4 The imported data file needs to be named and saved.
The importing process has produced an IBM SPSS data file. It is currently untitled, as it has just been created. This file should be saved. Once saved, we would select the Variable View, modify the default specifications to match our data (overruling the IBM SPSS defaults), and again save the modified version of the data file. When finished, it would be indistinguishable from the one we created in Chapter 2.
IBM SPSS will always provide an acknowledgment of its actions; acknowledgment of the importing process is shown in Figure 3.5. This is the syntax that mediated the
Figure 3.5 The acknowledgment of importing need not be saved.
importing of the Excel file; a good deal of specific information is shown, but the general message is to deal with an .xls format to be found via the given path and capture all of the data fields with their names. It is sufficient to close that window without saving (there will be a prompt requesting a decision about saving the file when we attempt to close it).
Part 2
Obtaining, Editing, and Saving Statistical Output
Chapter 4: Performing Statistical Procedures in IBM SPSS®
4.1 Overview
The Analyze menu located in the IBM SPSS main menu bar (see Figure 4.1) gives us access to a range of statistical procedures (e.g., analysis of variance, linear regression), and we will use it extensively, but we will also be working to a certain extent with some of the other menus in the main menu (e.g., Data, Transform). All of this work will involve interacting with dialog windows and, in most circumstances, obtaining output.
Figure 4.1 The main menu of IBM SPSS.
Performing statistical analyses is carried out in two stages: analysis setup and viewing/interpreting the output. Here, we treat these processes in a generic manner, without being concerned about the details of the setup or the interpretation of the output; rather, our purpose is to present the general process of what will be done for most analyses.
4.2 Using Dialog Windows to Setup the Analysis
In order to set up an analysis, it is necessary to have a data file open, as the analysis will be performed on the active data file. We will use the data file named sales extraversion that was created in Chapter 2 to illustrate our analysis setup, and we will call upon the Bivariate Correlations procedure simply as a matter of convenience (we could have chosen any other procedure as they are all structured similarly) to illustrate how to perform a statistical analysis.
With the data file as the active window, select from the IBM SPSS menu Analyze → Correlate → Bivariate to reach the main Bivariate Correlations dialog window shown in Figure 4.2. The main window of a statistical procedure is the one where we identify the variables to be included in the analysis; it is almost always the window that opens when we invoke most procedures.
Figure 4.2 The main Bivariate Correlations dialog window.
Although there are many elements contained in these main dialog windows, working with them is pretty straightforward. Following are the major elements of a main dialog window using Bivariate Correlations as our medium:
The banner of the window provides the name of the procedure.
The unnamed panel in the upper left quadrant of the window, which we will call the
Variable List
panel, contains the names of the variables in the data file. They are listed in the order that they appear in the data file but it is possible to have them listed alphabetically (e.g., select
Edit → Options → General
tab and click
Alphabetical
under
Variable Lists
as shown in
Figure 4.3
).
The
Variables
panel identifies the variables that are to be included in the analysis. Variables may be moved into the
Variables
panel either by highlighting them in the
Variable Lists
panel and clicking the arrow button or by double-clicking the name of the variable.
The pushbuttons on the far right of the main window ordinarily open subordinate dialog screens in order to customize some aspect of the statistical analysis or to instruct IBM SPSS to print certain information in the output.
The pushbuttons on the bottom of the window enable some action to take place. They must be active to be eligible for selection. For example, the
OK
pushbutton (this enables the analysis to be performed) is not currently available (it is not active) because no variables have yet been moved into the
Variables
panel (there are no variables identified on which an analysis can be performed).
Figure 4.4 shows the two variables of extraversion and sales after they have been moved into the Variables panel. Note that the OK pushbutton is now active, but we will open one of the subordinate windows to demonstrate this aspect of the setup before selecting OK.
Figure 4.3 The General tab in editing the IBM SPSS system Options.
Figure 4.4 The main Bivariate Correlations dialog window with extraversion and sales entered into the Variables panel.
From the main dialog window, select the Options pushbutton. This opens the Options dialog screen shown in Figure 4.5. By way of illustrating how to interact with such screens, we have checked Means and standard deviations and retained the default selection of Exclude cases pairwise (we will explain these and other options in the context of the various analyses). Select Continue to return to the main dialog window and select OK to perform the analysis.
Figure 4.5 The Options screen of Bivariate Correlations.
4.3 The Output
Figure 4.6 displays the results (usually called output) of the analysis. It is typical in IBM SPSS output that much of the results of the analysis is contained in tables. Each table is headed by a title above it. In the Bivariate Correlation procedure, the first table presents the means and standard deviations of the variables; these are provided in the output because we requested in the Options dialog screen that these descriptive statistics be displayed.
Figure 4.6 Sample output from the Bivariate Correlations procedure.
The second table shows the Pearson correlation between extraversion and sales (a rather exaggerated .936). In the Bivariate Correlation procedure, IBM SPSS footnotes with asterisks show different probability levels. Here, there is only one correlation. Also shown in the printout is Sig. (2-tailed); this is the exact probability of obtaining a Pearson correlation of .936 based on an N of 6 cases assuming that the null hypothesis (maintaining that the population correlation value is zero) is true.
Chapter 5: Editing Output
5.1 Overview
As noted in Section 4.3, much of the output in IBM SPSS® is presented in the form of tables. The tables produced by IBM SPSS in its output are “generic” and may be less well formatted than we would prefer, especially if the table is to be included in a report. However, tables in the output can be edited to a certain extent before copying them to a word processing document or saving them to a PDF (Portable Document Format) file. In this chapter, we illustrate some simple editing tasks that can be done.
To generate an output table so that we can illustrate these editing tasks, we have performed a Bivariate Correlations analysis with four variables and have obtained the table of Descriptive Statistics shown in Figure 5.1. Assume that we wish to show this table in a presentation and, in that context we determine that (a) the Std. Deviation column is disproportionally wide and (b) we would prefer to have the word “Standard” written out rather than being abbreviated.
Figure 5.1 A table of Descriptive Statistics obtained from the Bivariate Correlations procedure.
To edit this table, we double-click on it. The result is shown in Figure 5.2. We can determine that it is editable because of three visual cues:
It becomes outlined in a dashed line.
It takes on a red arrow to its left.
The table title format changes to white font on a black background.
Figure 5.2 The table is now editable.
5.2 Changing the Wording of a Column Heading
Double-clicking the column heading Std. Deviation permits us to edit it (see Figure 5.3). We then type out the full term (with a hard return after Standard) to replace the abbreviation as shown in Figure 5.4.
Figure 5.3 Double-clicking the column heading Std. Deviation permits us to edit it.
Figure 5.4 The full wording of Standard Deviation is now the column heading.
5.3 Changing the Width of a Column
We can also adjust the width of the columns. Double-clicking in the table and placing the cursor on one of the vertical border lines of the Standard Deviation column gives
us a double horizontal arrow as shown in Figure 5.5. Clicking and dragging this border to the left allows us to narrow the column as shown in Figure 5.6. Clicking anyplace outside of the table takes it out of edit mode. Having modified the file, we now save it.
Figure 5.5 In editing mode, it is possible to move the vertical column border in or out.
Figure 5.6 The Standard Deviation column is now narrower.
5.4 Viewing More Decimal Values
IBM SPSS carries out its computations to 16 decimal places, but the full range of these decimal values is almost never displayed in the output tables, a strategy that makes a good deal of sense, given the level of measurement precision in our research instrumentation. Nevertheless, the fact that they are not displayed belies the fact that they are there—they are present but hidden from view, as it is a rare occurrence when we wish to see all of that information.
To view the full set of decimal values, we double-click the table to place it in edit mode. Then we double-click the entry whose full decimal values we wish to see. This is shown in Figure 5.7 where we have selected the entry for the standard deviation of social whose value is displayed in the table as 1.36015. By double-clicking that entry, we can see the full decimal value 1.3601470508735443 that was heretofore (and gratefully) hidden. Note that the ordinarily displayed tabled entry is a properly rounded representation of the full decimal value as computed by IBM SPSS.
Figure 5.7 Double-clicking a numerical entry in a table yields the decimal value to 16 places.
It is common that SPSS will print a sig. value (the probability of the statistic occurring by chance alone if the null hypothesis is true) to allow us to test the statistical significance of an obtained statistic (e.g., a Pearson r value) against the alpha level we have established. In many of the statistical procedures, the sig. value is given to three decimal places.
It is not uncommon for these probability values to be sufficiently low that the number of zero digits well exceeds the three-decimal printing limitations in the IBM SPSS tables. For example, the computed probability might be .000316. This conundrum is resolved by IBM SPSS, sometimes to the dismay of students, by presenting in the output table a sig. value of .000.
In presenting the results of any analysis in a report, probability values should never be reported as .000; rather, they should be reported as p < .001 (American Psychological Association, (2009)). This is because the probability is never zero, but just a very low value. Double-clicking the displayed value of .000 will yield the longer decimal.
Some probability values may be sufficiently low that IBM SPSS will present them in what is known as exponential notation. For example, the value of .000316 would be displayed as 3.16E−4. This notation is interpreted as follows:
3.16 is the base nonzero numeral in the expression.
E indicates that the value is written in exponential notation.
The dash is a minus sign directing us to move the decimal to the left, adding zeros as needed.
4 is the number of decimal places involved in the move.
Putting all this together, the exponential notation in this instance directs us to move the decimal in 3.16 four places to the left. In order to comply, it is necessary to add three zeros to the left of the 3; thus, the end result is .000316.
5.5 Editing Text in IBM SPSS Output Files
Although much of the information in output files is contained in tables, text is also produced. As an example, text in an output file will document the analysis setup by displaying the underlying syntax. This will be the case even when IBM SPSS provides an acknowledgment that it has carried out an instruction.
The output text can be edited. While we would not wish to change it, we may wish to copy such text to a word processing or other document. Double-clicking on the text gives us editing access to it, and we can, for example, copy and paste it where we wish.
Chapter 6: Saving and Copying Output
6.1 Overview
The statistical analysis of our data is contained in IBM SPSS® output files. These results need to be saved and often need to be copied into reports or other documents. In this chapter, we show how to accomplish these operations.
6.2 Saving an Output File as an IBM SPSS Output File
In Section 1.3, we indicated that IBM SPSS output files are associated with an .spv extension. The standard file-saving routine within IBM SPSS will save the output file in that format. To save a newly generated output file, select File → Save As or select the Save File icon (shown in Figure 2.9). This opens a standard file-saving dialog screen in the operating system (e.g., Windows 7, Mac OS X). Navigate to the desired location (e.g., a personal flash drive), name the file, and save it.
It is possible to directly open the output file later. Double-clicking the file icon in the directory will open the file provided that there exists on that computer the same or a more recent version of IBM SPSS that created the file. Note that if for some reason we are using a computer that does not have IBM SPSS (e.g., a home computer), or one that contains an earlier version of the software, then trying to open the output file is not possible.
6.3 Saving an Output File in Other Formats
IBM SPSS allows for an output file to be saved in a myriad of formats including HTML, Excel, PowerPoint, Microsoft Word/RTF, and PDF. We discuss how to do this for PDF, but our description can also be generalized to the other formats.
A PDF document is a type of file that is in Portable Document Format. It is a faithful copy of the original but it is not editable unless it is opened in the full version of Adobe Acrobat or some comparable application. When PDF documents are printed or viewed on the screen, they mirror what was on the screen originally even though the current computer may not have the fonts that were used in the document; that is what makes them portable—the PDF contains within it all the information necessary for the document to be displayed on the screen or printed.
IBM SPSS is capable of saving a PDF version of the output file. This is an ideal way to view the full set of results (view a copy of the output file) when we cannot or choose not to access IBM SPSS.
To instruct IBM SPSS to save an output file as a PDF document, the file should be the active window (click its banner to make sure that it is active). Then select from the main menu File → Export. This opens the Export Output screen shown in Figure 6.1. Select Portable Document Format (*.pdf) from the File Type drop-down menu as shown in Figure 6.2.
Figure 6.1 The Export Output screen.
Figure 6.2 Some of the choices in the File Type drop-down menu on the Export Output screen.
With the File Type specified as PDF, select Browse and navigate to the location where the file is to be saved. Name the file. Figure 6.3 shows the result of this browsing and naming process. Then click Save. This returns us to the Export Output screen. Click OK and wait for the creation process to be finished.
Figure 6.3 Name the file in a way that suggests its context and contents.
6.4 Using Operating System Utilities to Copy an IBM SPSS Table to a Word Processing Document
As noted in Section 4.3, much of the output in IBM SPSS is presented in the form of tables. It is not uncommon for users to save (copy) only certain output tables to a word processing program (e.g., Microsoft Word). The two most useful forms to save such copies are screenshots (images) and word processing tables. Both PCs and Macs have utility programs as a part of their operating systems that can generate such copies.
