Student Study Guide

With IBM® SPSS® Workbook

for Research Methods,
Statistics, and Applications
Third Edition
Sara Miller McCune founded SAGE Publishing in 1965 to support
the dissemination of usable knowledge and educate a global
community. SAGE publishes more than 1000 journals and over
800 new books each year, spanning a wide range of subject areas.
Our growing selection of library products includes archives, data,
case studies and video. SAGE remains majority owned by our
founder and after her lifetime will become owned by a charitable
trust that secures the company’s continued independence.

Los Angeles | London | New Delhi | Singapore | Washington DC | Melbourne

Student Study Guide With
IBM® SPSS® Workbook
for Research Methods,
Statistics, and Applications

Third Edition

Kathrynn A. Adams
Guilford College

Eva K. McGuire
Guilford College
FOR INFORMATION: Copyright © 2023 by SAGE Publications, Inc.

SAGE Publications, Inc.

All rights reserved. No part of this book may be reproduced or utilized
2455 Teller Road
Thousand Oaks, California 91320 in any form or by any means, electronic or mechanical, including
E-mail: [email protected] photocopying, recording, or by any information storage and retrieval
system, without permission in writing from the publisher.
SAGE Publications Ltd.
1 Oliver’s Yard
All trademarks depicted within this book, including trademarks appearing
55 City Road
London, EC1Y 1SP as part of a screenshot, figure, or other image, are included solely for the
United Kingdom purpose of illustration and are the property of their respective holders.
The use of the trademarks in no way indicates any relationship with, or
SAGE Publications India Pvt. Ltd. endorsement by, the holders of said trademarks. SPSS is a registered
B 1/I 1 Mohan Cooperative Industrial Area
trademark of International Business Machines Corporation.
Mathura Road, New Delhi 110 044
India
Printed in the United States of America
SAGE Publications Asia-Pacific Pte. Ltd.
18 Cross Street #10-10/11/12 ISBN: 978-1-0718-1789-6
China Square Central
Singapore 048423

This book is printed on acid-free paper.

Sponsoring Editor: Leah Fargotstein

Product Associate: Kenzie Offley
Production Editor: Astha Jaiswal
Copy Editor: Integra
Typesetter: Hurix Digital
Cover Designer: Candice Harman
Marketing Manager: Victoria Velasquez 22 23 24 25 26 10 9 8 7 6 5 4 3 2 1
 BRIEF CONTENTS

Preface xii

Publisher’s Acknowledgments xiv

About the Authors xv

Chapter 1 • Thinking Like a Researcher 1

Chapter 2 • Build a Solid Foundation for Your Study Based on Past Research 9
Chapter 3 • Measuring Your Variables 18
Chapter 4 • The Cornerstones of Good Research: Reliability and Validity 27
Chapter 5 • Basics of Research Design: Description and Sampling 37
Chapter 6 • Describing Your Sample 44
Chapter 7 • Beyond Descriptives: Making Inferences Based on Your Sample 52
Chapter 8 • Comparing Your Sample to a Known or Expected Score 62
Chapter 9 • Examining Relationships Among Your Variables:
Correlational Design 71
Chapter 10 • Examining Causality 86
Chapter 11 • Independent-Groups Designs 95
Chapter 12 • Dependent-Groups Designs 113
Chapter 13 • Factorial Designs 132
Chapter 14 • Nonparametric Statistics 143
Chapter 15 • Focusing on the Individual: Case Studies and Single N Designs 157
Chapter 16 • How to Decide? Choosing a Research Design and
Selecting the Correct Analysis 165
IBM® SPSS® Workbook Part 1 • Fundamentals of Data Analysis 173
IBM® SPSS® Workbook Part 2 • 
Analyses to Examine One Variable
at a Time 193
IBM® SPSS® Workbook Part 3 • 
Analyses to Examine Relationships
(No Groups) 212
IBM® SPSS® Workbook Part 4 • 
Analyses to Compare Independent
Groups221
IBM® SPSS® Workbook Part 5 • 
Analyses for Dependent Groups and
Mixed Designs 248
 TABLE OF CONTENTS

Prefacexii

Publisher’s Acknowledgments xiv

About the Authors xv

Chapter 1 • Thinking Like a Researcher 1

Learning Objectives 1
Chapter Summary 1
Chapter 1 Quiz 1
Chapter 1 Exercises 3
Exercise 1.1: Critical Thinking 3
Exercise 1.2: The Scientific Approach 4
Exercise 1.3: The Research Process (Aka The Scientific Method) 4
Exercise 1.4: Thinking Critically About Ethics 6
Exercise 1.5: The Big Picture: Proof and Progress in Science 7
Your Research 8
Take a Scientific Approach to Identify a Research Topic 8

CHAPTER 2 • Build a Solid Foundation for Your Study Based on

Past Research 9
Learning Objectives 9
Chapter Summary 9
Chapter 2 Quiz 10
Chapter 2 Exercises 11
Exercise 2.1: Types of Sources 11
Exercise 2.2: Strategies to Identify and Find Past Research 13
Exercise 2.3: Reading and Evaluating Primary Research Articles 13
Exercise 2.4: Crediting Sources 15
Exercise 2.5: The Big Picture: Use the Past to Inform the Present 16
Your Research 17
Find Research on Your Topic 17
Read, Evaluate, and Reference a Primary Research Article on Your Topic 17

CHAPTER 3 • Measuring Your Variables 18

Learning Objectives 18
Chapter Summary 18
Chapter 3 Quiz 18
Chapter 3 Exercises 20
Exercise 3.1: Constructs and Operational Definitions 20
Exercise 3.2: Scales of Measurement 21
Exercise 3.3: Self-Report Measures 22
Exercise 3.4 Behavioral and Physiological Measures 23
Exercise 3.5: Archival Research 25
Exercise 3.6: The Big Picture: How to Choose Measures for Your Study 25
Your Research 26
CHAPTER 4 • The Cornerstones of Good Research: Reliability and Validity 27
Learning Objectives 27
Chapter Summary 27
Chapter 4 Quiz 28
Chapter 4 Exercises 29
Exercise 4.1: Reliability and Validity 29
Exercise 4.2: Assessing the Reliability of Measures 30
Exercise 4.3: Assessing the Validity of Measures 31
Exercise 4.4: Reliability and Validity at the Study Level 32
Exercise 4.5: The Big Picture: Consistency and Accuracy 33
Your Research 36

CHAPTER 5 • Basics of Research Design: Description and Sampling 37

Learning Objectives 37
Chapter Summary 37
Chapter 5 Quiz 37
Chapter 5 Exercises 39
Exercise 5.1: When Is a Descriptive Study Appropriate? 39
Exercise 5.2: Validity in Descriptive Studies 40
Exercise 5.3: Defining the Population and Obtaining a Sample 40
Exercise 5.4: Probability Sampling 41
Exercise 5.5: Nonprobability Sampling 42
Exercise 5.6: The Big Picture: Choosing a Sampling Method 43
Your Research 43

CHAPTER 6 • Describing Your Sample 44

Learning Objectives 44
Chapter Summary 44
Chapter 6 Quiz 44
Chapter 6 Exercises 46
Exercise 6.1: Ethical and Practical Issues in Describing Your Samples 46
Exercise 6.2: Descriptive Statistics 46
Exercise 6.3: Choosing the Appropriate Descriptive Statistics 48
Exercise 6.4: Comparing Interval/Ratio Scores With z Scores and Percentiles 50
Exercise 6.5: The Big Picture: Know Your Data and Your Sample 51
Your Research 51

CHAPTER 7 • Beyond Descriptives: Making Inferences Based on Your Sample 52

Learning Objectives 52
Chapter Summary 52
Chapter 7 Quiz 53
Chapter 7 Exercises 54
Exercise 7.1: Inferential Statistics 54
Exercise 7.2: Hypothesis Testing 55
Exercise 7.3 Errors in Hypothesis Testing 56
Exercise 7.4: Effect Size, Confidence Intervals, and Practical Significance 57
Exercise 7.5: The Big Picture: Making Sense of Results 58
Your Research 60

CHAPTER 8 • Comparing Your Sample to a Known or Expected Score 62

Learning Objectives 62
Chapter Summary 62
 Chapter 8 Quiz 62
Chapter 8 Exercises 64
Exercise 8.1: Choosing the Appropriate Test 64
Exercise 8.2: One-Sample t Tests 65
Exercise 8.3: Calculating an Effect Size 67
Exercise 8.4: Calculating a Confidence Interval 67
Exercise 8.5: The Big Picture: Examining One Variable at a Time 68

CHAPTER 9 • Examining Relationships Among Your Variables:

Correlational Design 71
Learning Objectives 71
Chapter Summary 71
Chapter 9 Quiz 72
Chapter 9 Exercises 73
Exercise 9.1: Correlation Design 73
Exercise 9.2: Relationship Between Two Interval or Ratio Variables 74
Exercise 9.3: Regression 81
Exercise 9.4: The Big Picture: Correlational Design Versus Correlational Analysis 84
Your Research 85
Find and Evaluate a Nonexperimental Study on Your Topic 85

CHAPTER 10 • Examining Causality 86

Learning Objectives 86
Chapter Summary 86
Chapter 10 Quiz 86
Chapter 10 Exercises 88
Exercise 10.1: Testing Cause and Effect 88
Exercise 10.2: Eight Key Threats to Internal Validity 89
Exercise 10.3: Basic Issues in Designing an Experiment 91
Exercise 10.4: Validity in Experiments 92
Exercise 10.5: The Big Picture: Benefits and Limits of Experimental Design 93
Your Research 94
Find and Evaluate an Experiment on Your Topic 94

CHAPTER 11 • Independent-Groups Designs 95

Learning Objectives 95
Chapter Summary 95
Chapter 11 Quiz 96
Chapter 11 Exercises 97
Exercise 11.1: Designs With Independent Groups 97
Exercise 11.2: Designing a Simple Experiment 98
Exercise 11.3: Independent-Samples t Tests, Effects Sizes, and Confidence Intervals 99
Exercise 11.4: Designs With More Than Two Independent Groups 105
Exercise 11.5: Analysis of Multiple Independent-Groups Designs 105
Exercise 11.6: The Big Picture: Selecting Analyses and Interpreting Results
for Independent-Groups Designs 110
Your Research 112

CHAPTER 12 • Dependent-Groups Designs 113

Learning Objectives 113
Chapter Summary 113
Chapter 12 Quiz 114
 Chapter 12 Exercises 115
Exercise 12.1: Designs With Dependent-Groups 115
Exercise 12.2 Advantages and Drawbacks of Dependent Designs 116
Exercise 12.3: Analysis of Dependent Two-Group Designs 119
Exercise 12.4: Designs With More Than Two Dependent Groups 122
Exercise 12.5: Analysis of Dependent Multiple-Groups Designs 124
Exercise 12.6: The Big Picture: Selecting Analyses and Interpreting Results
for Dependent-Groups Designs 128
Your Research 131

CHAPTER 13 • Factorial Designs 132

Learning Objectives 132
Chapter Summary 132
Chapter 13 Quiz 132
Chapter 13 Exercises 134
Exercise 13.1: Basic Concepts in Factorial Design 134
Exercise 13.2: Rationale for Factorial Designs 135
Exercise 13.3: 2 × 2 Designs 135
Exercise 13.4: Two-Way Between Subjects ANOVA 138
Exercise 13.5: Beyond the 2 × 2 Independent-Groups Design 140
Exercise 13.6: The Big Picture: Embracing Complexity 141
Your Research 142

CHAPTER 14 • Nonparametric Statistics 143

Learning Objectives 143
Chapter Summary 143
Chapter 14 Quiz 143
Chapter 14 Exercises 145
Exercise 14.1: Parametric and Nonparametric Statistics 145
Exercise 14.2: Chi-Square Goodness of Fit 146
Exercise 14.3: Chi-Square Test for Independence 150
Exercise 14.4: Analyses for Dependent-Groups Designs with Nominal Data 153
Exercise 14.5: Spearman’s Rho 153
Exercise 14.6: Analyses for Independent- and Dependent-Groups Designs with
Ordinal Data 155
Exercise 14.7: The Big Picture: Selecting Parametric Versus Nonparametric Tests 155
Your Research 156

CHAPTER 15 • Focusing on the Individual: Case Studies and

Single N Designs 157
Learning Objectives 157
Chapter Summary 157
Chapter 15 Quiz 157
Chapter 15 Exercises 159
Exercise 15.1: Samples Versus Individuals 159
Exercise 15.2: Case Studies 160
Exercise 15.3: Qualitative Analyses 160
Exercise 15.4: Single N Designs 161
Exercise 15.5: The Big Picture: Choosing Between a Sample, Case Study, or
Single N Design 163
Your Research 164
Is a Case Study or Single N Design a Good Choice for Your Topic? 164
CHAPTER 16 • How to Decide? Choosing a Research Design and
Selecting the Correct Analysis 165
Learning Objectives 165
Chapter Summary 165
Chapter 16 Quiz 165
Chapter 16 Exercises 167
Exercise 16.1: Choosing a Research Design 167
Exercise 16.2: Selecting Your Statistical Analyses 169
Exercise 16.3: The Big Picture: Beyond This Class 171
Your Research 172

IBM® SPSS® Workbook Part 1: Fundamentals of Data Analysis 173

Before You Begin 173
Data Entry 174
Set up Your Data Set 174
Enter Scores 177
Check for Data Entry Errors Using SPSS 178
Create Scale Scores 180
Recoding Variables 180
Reliability Analysis (Cronbach’s Alpha) 182
Compute a New Variable 185
Managing Data 186
Creating Groups 186
Split the Data File, Select Groups, or Select a Random Sample 188
Part 1 Practice Exercises 189

IBM® SPSS® Workbook Part 2: Analyses to Examine One Variable at a Time 193
Descriptive Statistics 193
Nominal Variables 193
Interval or Ratio Variables 194
Writing up Results 199
Descriptive Statistics Practice Exercises 202
Confidence Intervals 204
Data Entry 204
Calculating the Confidence Interval for the Mean 204
Writing up Results 205
Confidence Intervals Practice Exercise 206
One-Sample t Test 207
Writing up Results 208
One-Sample t Test Practice Exercise 208
Chi-Square Goodness of Fit 210
Writing up Results 210
Chi-Square Goodness of Fit Practice Exercise 211

IBM® SPSS® Workbook Part 3: Analyses to Examine Relationships

(No Groups) 212
Pearson’s r and Spearman’s Rho 212
Data Entry 212
Create a Scatterplot 212
Calculate and Interpret a Correlation Coefficient 213
Conducting Multiple Correlation Coefficients 214
Writing up Results 215
Linear Regression 217
Part 3 Practice Exercises 219
IBM® SPSS® Workbook Part 4: Analyses to Compare Independent Groups 221
Independent-Samples t Test 221
Data Entry 222
Conducting an Independent-Samples t Test 223
What Is Levene’s Test? 224
Effect Size for a Two Independent-Groups Design 225
Writing up Results 227
Independent-Samples t Test Practice Exercises 228
One-Way Between-Subjects ANOVA 231
Data Entry 231
Conducting a One-Way ANOVA 231
Writing up Results 235
One-Way Between-Subjects ANOVA Practice Exercises 236
Two-Way Between-Subjects ANOVA (or Three-way, etc.) 238
Data Entry 238
Conducting a Two-Way Between-Subjects ANOVA 239
Writing up Results 240
Two-Way Between-Subjects ANOVA Practice Exercises 242
Nonparametric Statistics 245
Conducting a Chi-Square Test for Independence 245
Mann-Whitney U and Kruskal-Wallis H Tests 247

IBM® SPSS® Workbook Part 5: Analyses for Dependent Groups and

Mixed Designs 248
Dependent-Samples t Test 248
Data Entry for Dependent Designs 249
Conducting a Dependent-Samples t Test 250
Writing up Results 251
Dependent-Samples t Test Practice Exercises 251
One-Way Within-Subjects ANOVA 255
Data Entry 256
Conducting a One-Way Within-Subjects ANOVA 256
Writing up Results 259
One-Way Within-Subjects ANOVA Practice Exercises 259
Two-Way Within-Subjects ANOVA and Mixed ANOVA 262
Nonparametric Statistics 262
McNemar and Cochran Q Tests 262
Wilcoxon T and Friedman χ2 Tests 262
 PREFACE

The Student Study Guide and IBM® SPSS® Workbook is a companion for the textbook Research Methods,
Statistics, and Applications. Students can use the study guide and workbook as a self-guided tool to
reinforce and apply concepts from the textbook, or it can be used as an in-class or in-lab workbook, or
both. Professors also may wish to assign exercises from the study guide and workbook as homework.
In the textbook, we included in-chapter practice and applications. The study guide and workbook
provides students with additional opportunities to review, practice, and apply essential knowledge and
skills related to research methods and statistics.
Whereas many study guides repeat information from the textbook, we designed the study guide and
workbook to supplement, rather than supplant, the textbook, and therefore we keep chapter summaries
very brief. Additionally, the study guide and workbook goes beyond the textbook by including step-
by-step directions for using the data analysis program, IBM® SPSS®, interpreting output, and writing
up results.
Many study guides focus on recognition and recall of material, and we acknowledge that gaining basic
knowledge is an essential initial step in the learning process. We provide students with recognition
activities via multiple-choice quizzes at the beginning of each study guide chapter, with the answers
appearing immediately afterward so that students can easily check their understanding of basic concepts.
Additionally, we include review exercises to encourage deeper processing by having students recall,
rather than merely recognize, key terms.
A unique feature of this study guide and workbook is that it encourages students to build on their
foundational knowledge. The majority of the exercises require students to think critically and actively
engage with the material. These application exercises require more effort and involvement than
recognition or recall tasks. Such exercises encourage students to make meaningful connections and
are more likely to have enduring effects on their understanding and retention of research processes.
Additionally, most chapters of the study guide include a “Your Research” exercise so that students can
apply key terms and concepts to their own research projects.
Chapters in the Student Study Guide and IBM® SPSS® Workbook contain the following sections:

—— Learning Objectives
—— Chapter Summary
—— Chapter Quiz
—— Chapter Exercises
—— Your Research

Learning Objectives are listed at the beginning of each chapter in both the textbook and the study
guide and workbook. The learning objectives are designed to help students reorient to the textbook
chapter. Moreover, a reminder of the learning objectives encourages students to consider what they have
learned and what areas require additional review.
The Chapter Summary provides students with a very brief overview of the material covered in the
textbook chapter. Like the learning objectives, the summary is designed to reorient the students to the
textbook chapter.
The Chapter Quiz includes 10 multiple-choice questions. It is designed to be a review of key concepts.
Answers are provided directly following the quiz to ensure that students adequately understand key
concepts before going on to the chapter exercises, which require deeper processing and application.

xii
 Preface  xiii

The Chapter Exercises match the organization of the textbook chapters and are designed to help
students achieve the learning objectives. The exercises designated as “Review” require students to
recall key terms and concepts. The exercises designated as “Application” go beyond simple definitions
and require students to apply key terms and concepts. These application exercises lend themselves
particularly well to in-class or in-lab activities and discussion, especially when students are required to
complete the fill-in-the blank sections prior to class or lab.
Students: Instructors have access to the answer keys.
Instructors: You can download the complete answer keys via the Instructor’s Resource page on the
textbook website and then share with your students as you see fit. There are also answer keys for only
the odd-numbered answers. Some professors find it useful to provide all or half the answers to students
ahead of time so that they may check their own work. Other professors assign study guide exercises as
graded or ungraded homework assignments or use the exercises in class and do not wish the students to
have advance access to the answers. We have found assigning exercises from the study guide as ungraded
homework (checked simply as complete or incomplete) helps to ensure students come to class prepared
and with good questions about the material. We then share the answers during or after class.
In the Your Research section, students apply key concepts and skills from the chapter to a research
topic of their choice. We have found that encouraging students to think about and develop their own
area of research promotes deeper understanding and integration of the material. The exercises in this
section may be used as homework assignments, and they are especially useful in classes that require a
semester- or year-long research project. Having students consider how different concepts apply to a
topic of their choice can also be useful in the absence of such a requirement. The questions about the
research project are worded to apply to a great range of topics that might be selected by students or
professors. In addition, the format allows students to compare their responses when working as a group
on a research project.
The IBM® SPSS® Workbook provides step-by-step directions for data analysis and interpretation. There
are practice exercises to help students gain competence using the program, interpreting output, and
writing up results. As with the chapter exercises, instructors have online access to the answer keys. The
workbook is designed for students brand new to SPSS. We also hope that it will be a useful resource that
students can keep and refer to should they use the program in future coursework, graduate programs,
or careers.
New to this edition: For the third edition, the IBM® SPSS® Workbook is a separate section divided into
five parts. We believe this will make finding the directions for specific analyses easier, and students
can easily refer back to the workbook throughout their academic careers and beyond. The workbook
organization matches the guidelines on selecting statistical analyses and includes similar flowcharts
found in the final chapter of the textbook, “Putting it all Together”. We also provide directions to an
alternative data analysis program, jamovi, on the Instructor Resource site.
 PUBLISHER’S ACKNOWLEDGMENTS

SAGE wishes to acknowledge the valuable contributions of the following reviewers.

Michael D’Addona, The College of New Jersey

Chia-Fang (Sandy) Hsu, University of Wyoming
Michael Friedline, Piedmont University
Ahmed Ibrahim, Johns Hopkins University
Adam J. Newmark, Appalachian State University
Jaesin Sa, Touro University
Heo Wookjae, South Dakota State University

xiv
ABOUT THE AUTHORS

Kathrynn (Kathy) A. Adams earned her PhD in general experimental

psychology from the University of Alabama in 1977. She was a Charles
A. Dana Professor of Psychology at Guilford College when she retired in
2017 after 37 years of teaching. Her professional interests include gender
issues, relationships, and teaching pedagogy. She worked with the Preparing
Future Faculty Program for 20 years and helped establish the Early College
at Guilford, a nationally ranked high school. In her spare time, she spends as
much time as possible outdoors, practices yoga, and bakes chocolate desserts.

Eva K. McGuire earned her PhD in clinical psychology from Virginia

Commonwealth University in 2002. She is a Charles A. Dana Professor
of Psychology at Guilford College, where she has taught since 2003. Her
research interests include environmental psychology and computer-mediated
communication. Eva enjoys walking, yoga, and bike riding, and she loves to
listen to live music.

xv
 1
THINKING LIKE A RESEARCHER

LEARNING OBJECTIVES

• Explain the connection between thinking critically and thinking like a researcher.
• Describe the scientific approach, including the challenges and benefits of taking a scientific approach.
• Outline and synthesize the steps of the research process.
• Identify and apply ethical principles and standards.
• Formulate a big picture of the nature of science and how science progresses.

CHAPTER SUMMARY
Students in the social and behavioral sciences are often curious about personal and social phenomena, and learning about
research methods and statistics will give you the tools necessary to engage your curiosity on a deeper level. Critical thinking is
key to research, and researchers put critical thinking to action throughout the entire process of carrying out a research study.
Additionally, researchers think by taking a scientific approach. You are encouraged to consider why the scientific approach
is important and to consider the scientific method as a process rather than a list of steps that must be followed in order. A
complete overview of the research process from start to finish is included in this chapter. The goal is to introduce you to
key concepts in the context of the overall scientific process, and later chapters detail the steps and concepts in more depth.
The next part of the chapter focuses on ethics. The chapter provides a brief history of ethical violations that led to the
development of current ethics codes. Key ethical principles and standards are discussed, and you are encouraged to
consider how ethical standards help to uphold broader ethical principles. The chapter ends with a discussion of proof and
progress in science, debunking the notion that any single study leads to proof and, instead, encouraging you to consider
how multiple studies move the field forward.

CHAPTER 1 QUIZ
Test your knowledge of the key terms from the chapter.
1. Which of the following might be a variable in a study?
a. male
b. participants
c. gender
d. method
1
2   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. Circle the two key characteristics of a testable hypothesis:

a. It is based on past research or theory
b. It is unique and completely original
c. It can be proven
d. It can be disproven
3. If you were to test if the flavor of ice cream impacts motivation, flavor of ice cream is the
a. independent variable
b. dependent variable
c. quasi-experiment
d. hypothesis
4. What is the dependent variable from question 3?
a. ice cream
b. motivation
c. participants
d. flavor
5. Which of the following research designs is best able to demonstrate causation?
a. descriptive research
b. correlational research
c. quasi-experimental research
d. experimental research
6. The Institutional Review Board helps ensure that a study
a. meets the requirement for a class
b. is well written
c. shows causation
d. adheres to ethical principles and standards
7. represents the ideals, whereas are the specific rules designed to
uphold those ideals.
a. Ethical principles; ethical standards
b. Ethical standards; ethical principles
c. Ethics codes; ethic review boards
d. Ethic review boards; ethic codes
8. In the 1960s, Stanley Milgram conducted several studies on obedience in which an
experimenter ordered participants to administer increasingly powerful shocks to a
“learner.”
The participants learned afterward that the “learner” was a confederate who was working
with the experimenter and was never actually shocked. In other words, Milgram included
but not in his study.
a. informed consent; debriefing
b. debriefing; informed consent
c. confidentiality; informed consent
d. confidentiality; debriefing
 Chapter 1 ■ Thinking Like a Researcher   3

9. When the participants in Milgram’s obedience studies asked to stop, the experimenter ordered
them to continue. This violates the ethical standard of
a. confidentiality
b. the right to withdraw
c. informed consent
d. noncoercive incentives
10. Which of the following must be a part of all research studies?
a. informed consent
b. debriefing
c. confidentiality
d. anonymity

Quiz 1 Answer Key

1. c. 2. a, d. 3. a. 4. b. 5. d. 6. d. 7. a. 8. b. 9. b. 10. c.

CHAPTER 1 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 1.1: Critical Thinking

Learning Objective: Explain the connection between thinking critically and thinking like a researcher.
Review
1. List the similarities between thinking critically and thinking like a researcher

2. How is thinking like a researcher a unique type of critical thinking?

Application: Think critically about the information you have received about this research methods
course and consider how you might gather new evidence to evaluate such information.
1. Identify one thing you heard about this class: _____________________________________

2. Who or what was the source of this information? __________________________________

3. What biases might this source have? ____________________________________________

4   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

4. What additional evidence would help you evaluate the information, and how might you go
about gathering this evidence?

Exercise 1.2: The Scientific Approach

Learning Objective: Describe the scientific approach, including the challenges and benefits of taking a
scientific approach.
Review
Which of the following are associated with the scientific approach?

a. complexity b. advice c. improved decision making

d. comfort e. knowledge base f. gut reaction
g. method h. belief i. minimizing bias
Application
1. Recall a time that you made a decision or responded to a friend or family member based
on personal beliefs or personal experiences rather than taking a scientific approach. Briefly
outline that situation.

2. How might you have instead taken a scientific approach? What would be the benefits and
challenges to taking a scientific approach in this situation?

Exercise 1.3: The Research Process (Aka the Scientific Method)

Learning Objective: Outline and integrate the steps of the research process.
Review
1. The steps in the scientific method are
Step 1:
Step 2:
Step 3:
 Chapter 1 ■ Thinking Like a Researcher   5

Step 4:
Step 5:
Step 6:
Step 7:
2. Replace the crossed out words to summarize how the steps are synthesized within the
scientific method:

The steps are NOT… The steps ARE…

about proof about
linear
isolated

Application A: Choose a research design

1. Circle the questions that can be examined with an experiment:
a. Do students who participate in college sports study more or less than students who do not
participate in college sports?
b. Can political campaigns raise more money using negative campaign ads?
c. Are attractive people perceived as more or less intelligent than not-so-attractive people?
d. How do individuals perceive their local police department?
e. Does exposure to violence increase risk of heart disease?
f. Do people high in the personality trait narcissism take more “selfies” than those low in
this trait?
g. Is humanity becoming more or less violent?
h. Can daily statements of gratitude improve one’s well-being?
2. For those questions you identified in question 1 that could be examined with an experiment,
identify the independent and dependent variable.

3. For those questions you identified in question 1 that could not be examined with an
experiment, identify the most appropriate type of research to help answer the question.
6   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application B: Communicate Without plagiarizing

The following excerpt was taken directly from Cash and Whittingham (2010, p. 180):

We found that Nonjudge, the ability to refrain from judging one’s own cognitions, emotions,
and bodily sensations, predicted lower levels of depression, anxiety, and stress. Furthermore,
Act-aware, the ability to maintain awareness of daily activities, predicted lower levels of
depression.

Imagine that three students summarized this excerpt. Consider whether each student plagiarized Cash
and Whittingham (2010) and explain your answer.

Student 1
Cash and Whittingham (2010) found that Nonjudge, the ability to refrain from judging one’s own
cognitions, emotions, and bodily sensations, predicted lower levels of depression, anxiety, and stress.
Furthermore, Act-aware, the ability to maintain awareness of daily activities, predicted lower levels of
depression.
plagiarism or not plagiarism
because .

Student 2
Cash and Whittingham (2010) found that participants who were most aware during their daily activities
reported few symptoms of depression and those who were most accepting of their own thoughts,
feelings, and sensations reported fewer symptoms of both depression and anxiety.
plagiarism or not plagiarism
because .

Student 3
“Nonjudge, the ability to refrain from judging one’s own cognitions, emotions, and bodily
sensations,” and “Act-aware, the ability to maintain awareness of daily activities,” are linked to
improved mood.
_____ plagiarism or ____ not plagiarism
because _________________________________________________________.

Exercise 1.4: Thinking Critically About Ethics

Learning Objective: Identify and apply the ethical principles and standards of your discipline.
1. In the Stanford prison experiment (Zimbardo, 1972), some participants assigned to be guards
acted cruelly toward the participants assigned to be prisoners. Some of those assigned to be
prisoners became depressed and withdrawn. Identify and explain one ethical principle this
experiment violated (there are multiple answers).
 Chapter 1 ■ Thinking Like a Researcher   7

2. A researcher wants to observe interactions in a social media group. The group members are
anonymous, and the group is open, in that anyone can log on and view conversations. Imagine
you are a member of the Institutional Review Board. List three questions that you think are
most important to ask and identify why each is important, based on ethical principles and
standards.

3. A researcher wants to investigate the relationship between depression and exposure to

neighborhood violence, which includes hearing gunshots, witnessing verbal abuse, witnessing
a physical assault, or witnessing a murder. The researcher plans on recruiting adult participants
at various community meetings and then asking them to complete a questionnaire. Imagine
you are a member of the Institutional Review Board. List three questions that you think are
most important to ask and identify why each is important, based on ethical principles and
standards.

Exercise 1.5: The Big Picture: Proof and Progress in Science

Learning Objective: Formulate a big picture of the nature of science and how science progresses.
1. A study found that mindfulness training was effective in reducing stress. Does this prove that
mindfulness is an effective stress-reduction strategy? Explain.

2. Research consistently demonstrates the effectiveness of mindfulness in reducing stress. What

might we conclude?

3. What might be some of the next steps we take to examine the relationship between
mindfulness training and stress?
8   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

4. If someone asked you how science helps us understand something, how might you answer?

Check with your professor for answers to the chapter exercises.

Your Research

Take a Scientific Approach to Identify a Research Topic 3. Now evaluate your list and cross out or modify any that

1. Identify a topic that interests you: • are biased, in that you think you already know the
answer.
2. Come up with a list of 7–10 questions on this topic. Try to
build these questions from textbooks, research articles, • are not testable (you cannot disprove them).
or previous coursework, but you can also include a few
• necessitate equipment or a population to which
questions based on observations or experiences.
you do not have easy access.
a.
• do not fit the criteria, if any, laid out by your
b. professor (e.g., your professor may require you
to do an experiment, and not all questions can be
c.
experimental).
d.
4. What questions remain? These might serve as a
e. good starting point. Keep in mind that your research
question will change as you read more research in this
f.
area.
g.

h.

i.

j.
 2
BUILD A SOLID FOUNDATION FOR
YOUR STUDY BASED ON PAST
RESEARCH

LEARNING OBJECTIVES

• Distinguish types of sources by author, purpose, content, and utility.

• Apply strategies for identifying and finding past research.
• Evaluate primary research articles.
• Format citations and references in APA style.
• Formulate a big picture of how to build on past research.

CHAPTER SUMMARY
Chapter 2 focuses on finding and evaluating past research on a topic. The chapter is designed to help you discern
different sources, including the difference between primary and secondary sources, the difference between scholarly
and popular sources, and the various types of scholarly work that you might find in an academic journal (i.e., primary
research articles, literature reviews, and commentaries) as well as other scholarly sources (conference papers and
posters, unpublished manuscripts, books, theses and dissertations, etc.). You also learn the various ways to search for
relevant research.
How to read and evaluate primary research articles is another major focus of this chapter. The key sections of a primary
research article are outlined, and their purpose discussed. An article on the topic of academic achievement is used
to illustrate the different parts of a research article, and finding and referring to this article will greatly enhance your
understanding of key concepts. Specific guidelines for citing and referencing sources are provided. The chapter ends with
a discussion of building on past research and some specific strategies for doing so.

9
10   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

CHAPTER 2 QUIZ
Test your knowledge of the key terms from the chapter.
1. A may have been written by an expert or by a layperson and is
designed to educate or entertain.
a. scholarly source
b. popular source
c. meta-analysis
d. primary research article
2. Which of the following is considered a scholarly source?
a. articles published in academic journals
b. articles published in newspapers
c. textbooks
d. educational pamphlets
3. Which of the following is considered a primary research source?
a. any article in an academic journal
b. textbooks
c. reports of original research
d. literature reviews
4. A review of past research would be found in
a. a meta-analysis.
b. a literature review.
c. a primary research article.
d. All of the above
5. A review of past research without a report of new, original research is always a .
a. secondary research source
b. primary research source
c. popular source
d. scholarly source
6. When researchers collect and report on new data, those data are
a. primary data.
b. secondary data.
c. either a or b.
d. none of these.
7. A primary research source analyzes .
a. primary data
b. secondary data
c. either a or b
d. none of these
 Chapter 2 ■ Build a Solid Foundation for Your Study Based on Past Research   11

8. The “peers” in the peer review process for journals are

a. researchers who work at the same institution as the author.
b. students or interns.
c. experts in the field of study.
d. anyone the authors ask to review their work.
9. What process is used to reduce the likelihood that results from a study were due only to
chance?
a. peer review
b. statistical significance testing
c. power
d. external validity
10. refers to the generalizability of the results of a study.
a. Statistical significance
b. Power
c. External validity
d. Internal validity

Quiz 2 Answer Key

1. b. 2. a. 3. c. 4. d. 5. a. 6. a. 7. c. 8. c. 9. b. 10. c.

CHAPTER 2 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 2.1: Types of Sources

Learning Objective: Distinguish types of sources by author, purpose, content, and utility.

Application
1. The comedian Aziz Ansari wrote a book called Modern Romance. In addition to sharing
personal anecdotes, he reports results of interviews and focus groups that he and his research
partner conducted.
a. Is this a popular or scholarly source?

b. Are the reports of interviews and focus groups a primary or secondary research source?

2. Paul Eastwick, PhD, and colleagues conducted a meta-analysis examining the role of physical
attractiveness in dating preferences and actual dating behaviors. Their findings were published
in 2014 in the academic journal Psychological Bulletin.
a. Is this a popular or scholarly source?

b. Is their article a primary or secondary research source?

12   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. René Dailey, PhD, and colleagues conducted a study about on-again/off-again dating
relationships. Their findings were published in 2020 in the academic journal Personal Relationships.
a. Is this a popular or scholarly source?

b. Is their article a primary or secondary research source?

4. “Better than you think: The impact of TV on your relationship” is an article on the website luvze.
com. The author of the article, Gary Lewandowski, PhD, summarizes the results of two research
studies (neither of which he conducted) and explains what the results mean to the average person.
a. Is this a popular or scholarly source?

b. Is this article a primary or secondary research source?

5. Use Figure 2.1 in the textbook to help you answer the following questions.
a. Which of the sources from questions 1 to 4 above might you use to generate ideas for a
research topic but not include in your written report?

b. Which of the sources from questions 1 to 4 above might you use to identify patterns and
gaps in past research and to help you identify original research sources?

c. Which of the sources from questions 1 to 4 above represents the type of source that
should make up the majority of your sources for a research project?

6. Imagine you were using the online search engine Google (likely against your professor’s
recommendations) and you came across a report of original research.
a. How would you determine whether or not the source was a scholarly one?

b. Suppose that you determine the source is, in fact, scholarly. How else should you evaluate
it to determine whether it is appropriate to use as a source for a research project?
 Chapter 2 ■ Build a Solid Foundation for Your Study Based on Past Research   13

Exercise 2.2: Strategies to Identify and Find Past Research

Learning Objective: Apply strategies for identifying and finding past research.
Application A
Use the Database PsycINFO to answer these questions:
1. Do a search for the article titled “Income inequality and happiness: An inverted U-shaped curve.”
a. Who are the authors of this article?
b. What is the publication year?
c. Which journal published the article?
d. What is the volume number?
e. Is this a primary or secondary research article?
f. Find and download the pdf of the full article. As verification that you are able to do this,
what is the last word in the text on page 2?
2. Conduct a search for income inequality limiting publication year to between 2015 and 2020.
How many search results are there?
3. Conduct the same search as in question 2, but also limit the publication type to peer-reviewed
journals. How many search results are there now?
Application B
Go to the Directory of Open Access Journals (doaj.org)
1. Do a search for income and happiness. How many available articles are there?
a. Less than 150 b. 150 to 300 c. 300 to 450 d. over 450
2. Find the article titled “Is there a kink in the happiness literature” from the search results. Is
this a primary or secondary research article? How do you know?

3. Find the article titled “The biometric antecedents to happiness” from the search results. Is this
a primary or secondary research article? How do you know?

Exercise 2.3: Reading and Evaluating Primary Research Articles

Learning Objective: Evaluate primary research articles.
Review
Complete the following table listing the key parts of a primary research article in order, and each
section’s purpose. We have completed some of the table for you.
14   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Parts of a Primary
Research Article Purpose

Very briefly describes the study, typically includes key variables

Identifies the researchers who conducted and wrote up the study

Abstract

Explains importance of topic, reviews past research and theory to build a

rationale for the current study, and identifies hypotheses or research questions

Method

Lists all the sources cited in the paper

Application
Look at the example published article in Appendix B in the textbook.
1. What is the title of the article?
2. Who are the authors?
3. Is this a primary or secondary research source?
4. Can you tell from reading the abstract whether or not the authors conducted an experiment?
Why or why not?

5. Summarize, in your own words, the key point of the first paragraph of the Introduction.

6. Summarize, in your own words, the key point of the second paragraph of the Introduction.

7. What is the purpose of the last paragraph in the Introduction?

8. In your own words, summarize the method of the study.

9. In your own words, summarize the results of the study.

 Chapter 2 ■ Build a Solid Foundation for Your Study Based on Past Research   15

10. What is the first limitation noted by the authors? How might a future research study address
that limitation?

11. What questions does this study generate for you?

Exercise 2.4: Crediting Sources

Learning Objective: Format citations and references in APA style.
Application A
1. Look at the example published article in Appendix B in the textbook.
a. Imagine you were going to include this as a source for your own paper about how to
reduce plagiarism. Write a sentence you would include in your paper, being sure that
__the sentence accurately summarizes one or more finding from the study
__the sentence is in your own words
__you cite the correct source using APA format
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
b. Write an APA-style reference for the source you cited above. Use cursive or underline to
indicate italics.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
2. Look at the articles you found for Exercise 2.2.
a. How would you cite these articles?

Article title Narrative citation Parenthetical citation

Income inequality and happiness:

An inverted U-shaped curve

Is there a kink in the happiness

literature?

The biometric antecedents to

happiness

b. On a computer, type up an APA-style References section with these sources. Save your
document and check with your professor on how to turn it in, if applicable.
16   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application B
1. Find and correct the five APA-format citation errors in the paragraph:
Many of us like to believe that “beauty comes from within” and that that inner beauty is more
important than physical beauty. However, research suggests that physical attractiveness is the single
most important factor that impacts our desire to date someone (Kurzban and Weeden, 2005; Li et al.,
2013). An early cross cultural survey conducted by Buss & Schmitt found that across the world physical
attractiveness of one’s partner was more important for men than it was for women (1993). A more
recent meta-analysis found that gender differences only occur in ratings of ideal romantic partners and
ratings of hypothetical targets. Both men and women generally find physical attractiveness equally
important in actual face-to-face interactions (Eastwick, Luchies, Finkel & Hunt, 2014). Moreover,
although we might all desire an attractive partner, we tend to wind up with someone who matches our
level of physical attractiveness (Miller, R. S., 2012). Thus, our stated preferences do not fully predict our
actual dating behaviors (Eastwick et al., 2014).
2. Find and correct the five APA-format errors in the references.

References
   Miller, R. S. (2012). Intimate relationships (7th ed.). M
­ cGraw-Hill.
   Buss, D. M., and Schmitt, D. P. (1993). Sexual strategies theory: An evolutionary perspective on
human mating. Psychological Review, 100(2), 204–232. https://doi.org/10.1037/0033- 295X.100.2.204
   Eastwick, P. W., Luchies, L. B., Finkel, E. J., & Hunt, L. L. (2014). The predictive validity of
ideal partner preferences: A review and meta-analysis. Psychological Bulletin, 140(3), 623–665. https://doi
.org/10.1037/a0032432
   Kurzban, R., & Weeden, J. (2005). HurryDate: Mate Preferences in Action. Evolution and Human
Behavior, 26(3), 227-244. https://doi.org/10.1016/j.evolhumbehav. 2004.08.012
   Li, N. P., Yong, J. C., Tov, W., Sng, O., Fletcher, G. J. O., Valentine, K. A., Jiang, Y.F., Balliet, D.
(2013). Mate preferences do predict attraction and choices in the early stages of mate selection. Journal
of personality and social psychology, 105(5), 757–776. https://doi.org/10.1037/a0033777

Exercise 2.5: The Big Picture: Use the Past to Inform the Present
Learning Objective: Formulate a big picture of how to build on past research.
Review: List the four ways you might build on past research.

1. 
2. 
3. 
4. 

Check with your professor for answers to the chapter exercises.

Apply what you have learned to your own research by completing the “Your Research” exercises.
 Chapter 2 ■ Build a Solid Foundation for Your Study Based on Past Research   17

Your Research
Find Research on Your Topic c. Did the results support the hypothesis you
identified? Explain.
1. What database is most commonly used in your discipline?

2. Using that database, conduct a search for relevant and

recent research articles on your topic. Remember to try
out different keywords and consider searching by author
or looking at the references of key articles.

Doing this search will help you identify appropriate

articles as well as help you refine and revise your
research topic. It should take several hours if you
3. Write an APA-style reference for the article. Use cursive
do it correctly!
or underline to indicate italics.
Identify a few of the most relevant and
interesting articles. Find the full text of these
articles.

Read, Evaluate, and Reference a Primary Research

4. Consider how to build on the study:
Article on Your Topic
a. Identify one suggestion for future research that the
1. Identify and find the full text of a primary research
author(s) identified in the Discussion section of
article on your topic, or revised topic, that you might
the article and that you think you might be able to
model (in other words, you might use a similar
do.
method).

2. Answer the following based on the article, using your own

words (do not copy sentences or phrases directly from the
article). Practice APA formatting by citing the article in
each of your answers.

a. Briefly, what is the rationale for the study? (Hint:

Look in the Introduction.) b. Review Exercise 2.5. Which one of the four ways
to build on past research does your answer to
question 4a best fit under?

b. Identify one hypothesis from the article that

most interests you. (Hint: Hypotheses are usually
toward the end of the Introduction section.)
 3
MEASURING YOUR VARIABLES

LEARNING OBJECTIVES

• Explain how to operationally define constructs using qualitative and quantitative measures.
• Compare the four scales of measurement in terms of their four attributes.
• Identify and describe the advantages and disadvantages of self-report measures.
• Identify and describe the advantages and disadvantages of behavioral and physiological measures.
• Identify and describe the advantages and disadvantages of archival research.
• Formulate a big picture of how to choose measures for your study.

CHAPTER SUMMARY
Defining your variables is a critical early step of research. This step should be accomplished after you have reviewed the
relevant literature for ideas about possible variables you might examine, measure, or manipulate in your study. You must
clearly identify the constructs of interest and how these constructs will be operationally defined (or measured). It is also
important to identify the scale of measurement (nominal, ordinal, interval, ratio) of each variable. You should consider
the type of measure that best fits the study (descriptive, correlational, or experimental) you plan to conduct. Options
range from self-report (questionnaires, interviews) to behavioral and physiological measures (observations, unobtrusive
measures, performance-based, physiological) to archival (secondary data, records and documents). Understanding
the advantages and disadvantages of the different types of measures will assist you in your decision-making. How you
define and measure your variables will affect how you conduct your research, the analyses you can perform, and the
conclusions you can draw about your results.

CHAPTER 3 QUIZ
Test your knowledge of the key terms from the chapter.
1. A precise explanation of a variable in terms of how it is measured or manipulated is called a(n)
a. qualitative measure.
b. operational definition.
c. construct.
d. quantitative measure.

18
 Chapter 3 ■ Measuring Your Variables   19

2. A researcher is interested in children’s attention during reading lessons. She has second grade
teachers write a summary of a different target child’s behavior during the lesson each day for a
week. This measure is an example of a(n)
a. ratio measure.
b. operational measure.
c. qualitative measure.
d. quantitative measure.

3. The four attributes that determine the preciseness of a scale of measurement are
a. rating, ranking, ratio, equal intervals.
b. equal intervals, categories, ranking, uniqueness.
c. true zero, identity, order, equal intervals.
d. sequencing, rating, true zero, identity.

4. Runners rate their energy level on a 10-point scale as they finish the race. This measure
represents the scale of measurement.
a. interval
b. ordinal
c. ratio
d. nominal
5. Self-reports may be inaccurate due to the
a. descriptive nature of surveys.
b. external validity of surveys.
c. social desirability bias.
d. internal validity of surveys.
6. Questionnaire items that use a Likert-type response are one example of a(n)
a. close-ended response format.
b. open-ended response format.
c. scale score.
d. items that will need re-coding.
7. One disadvantage of conducting interview research is
a. participants are less likely to take the research seriously.
b. the response rate is likely to decrease.
c. maintaining anonymity is harder than with questionnaires.
d. the accuracy of responses is likely to increase.
8. In comparison to survey research, observational research
a. examines what people actually do, not what they say they do.
b. is less time consuming than survey research.
c. is less prone to observer bias.
d. requires less training of experimenters.
20   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

9. A observer in a study doesn’t know what the study’s hypotheses are, in

order to reduce observer bias.
a. biased
b. structured
c. covert
d. masked
10. Which of the following is NOT an advantage of archival research?
a. Since the data are already collected, the researcher can save time and resources.
b. Some archives span a large time frame so they allow for analysis of historical trends.
c. The fit between your hypothesis and the data is likely to be quite good.
d. There are often fewer ethical considerations than with other types of research.
Quiz 3 Answer Key
1. b. 2. c. 3. c. 4. a. 5. c. 6. a. 7. c. 8. a. 9. d. 10. c.

CHAPTER 3 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 3.1: Constructs and Operational Definitions

Learning Objective: Explain how to operationally define constructs using qualitative and quantitative
measures.
Review
1. Love is a , or a concept that cannot be directly measured.
2. If you wanted to conduct a research study about love, you would need to decide how to
measure love (acknowledging that it will not be a perfect, direct measure of love).
As researchers, what do we call the definition of a construct?

3. Circle the phrase that best represents an operational definition of plagiarism.

a. Making a higher grade than expected
b. Including a phrase of more than 3 words from another source without citing that source
c. Copying work
d. Not paraphrasing properly
4. Explain why you selected your answer for question 3:

5. Suppose your classmate tells you, “There is only one way to operationally define each
construct we study.” How would you respond to their statement?

6. If you decided to ask people to rate how much they love their partner on a scale from 1 to 10, you
would be operationally defining love using a measure because it is numerical.
 Chapter 3 ■ Measuring Your Variables   21

7. If you interviewed college students and asked them to describe the love they feel for their
partner, you would be operationally defining love using a measure
because it does not use numbers.

Exercise 3.2: Scales of Measurement

Learning Objective: Compare the four scales of measurement in terms of their four attributes.
Review
Fill in the table below:

Property Scales of Measurement With the Property

:
Each number on the scale has a unique
meaning.

:
Numbers on a scale are ordered.

:
The distance between numbers on a scale
is equal.

:
The scale has a fixed zero.

Application
1. Draw a line between each item on the left with the term on the right that best matches it.

1 = Strongly Disagree to 5 = Strongly Agree Absolute zero

Weight Interval scale of measurement

1 to 10 rating scale 5-point Likert-type scale

Rankings (1st place, 2nd place, etc.) Nominal scale of measurement

0 is a fixed point Ordinal scale of measurement

Marital status Ratio scale of measurement

2. Suppose you are studying attitudes toward the Black Lives Matter movement using a
questionnaire. Write four questions for your questionnaire using each of the four scales of
measurement once. List the scale of measurement before each question.
22   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Exercise 3.3: Self-Report Measures

Learning Objective: Identify and describe the advantages and disadvantages of self-report measures.
Review
1. Name two ways to reduce the social desirability bias of self-report measures.

2. What is the difference between a questionnaire and an interview?

3. In the following table, summarize the advantages and disadvantages of the following types of
self-report.

Format Advantages Disadvantages

Mailed questionnaire

Online questionnaire

Structured interview

Unstructured interview

4. Two reasons you should consider using already existing questionnaires are:

Application
1. Look over the items in a questionnaire assessing anxiety (higher scores = higher anxiety):
i. I almost never feel anxious.
Strongly disagree Disagree Neutral Agree Strongly Agree
ii. I worry about everything.
Strongly disagree Disagree Agree Strongly Agree
iii. My schoolwork and my procrastination make me anxious.
Strongly disagree Disagree Neutral Agree Strongly Agree
a. Which of these is a reverse-scored item?
b. Which of these is a double-barreled item?
c. How could you rewrite the item you identified in b?

d. Which of these is a forced choice response?

 Chapter 3 ■ Measuring Your Variables   23

e. After you fix the item you identified in b, explain the next steps you would take to
compute a total anxiety score.

2. Write an open-ended item assessing student anxiety.

3. Would it be better to interview or survey students about their anxiety? Explain the rationale
for your choice.

Exercise 3.4: Behavioral and Physiological Measures

Learning Objective: Identify and describe the advantages and disadvantages of behavioral and physiological
measures.

Review
1. Match the example on the left with the type of measure on the right.

During a sleep study, participants are hooked up to an a. Observation

EEG machine to record brainwaves.
b. Unobtrusive Measure

A researcher has students complete a tracking task on a c. Performance-based measure

laptop to assess their attention span. d. Physiological measure

A confederate washes their hands for either 30 seconds

or 3 seconds in front of others. Researchers covertly
time how long participants wash their hands under each
condition.

A researcher notes the dirt paths at the end of the school

year that were created by students as they walk across
grass spaces on campus.

2. In the table below, summarize the reason(s) you might choose to use each of the following
types of measures.

Measure Reason(s) to Use This Measure

Observation

Unobtrusive

Performance-based

Physiological
24   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. Fill in the table below on identifying and reducing bias:

Description of Bias Name of Bias How to Reduce (list strategies)

Observers focus on behaviors or

interpretations that support their
expectations

Interviewers provide verbal or

nonverbal cues that impact the
participants’ responses

Participants respond based on what

they think is socially acceptable or
how they want to be perceived

Application
Imagine that you are interested in understanding how students at your college or university view the
women’s basketball team (or another sports team, if your college or university does not have a women’s
basketball team).

1. What types of observations might you make that could help you better understand how
students view the team, and where would you make them? Be clear if your observation is
overt or covert, naturalistic or contrived, and if you include a participant observer and/or a
confederate.

2. What would be the advantages and disadvantages of such observations?

Advantages: _______________________________________________________________

Disadvantages: _____________________________________________________________

3. List 2 different behaviors you might record during your observation study using the timing of
a behavior.

4. What steps will you take to avoid observer bias?

5. Add an unobtrusive measure of student views of the team to your observational study.
 Chapter 3 ■ Measuring Your Variables   25

6. Would it be feasible to add a performance-based measure to your study? Explain.

7. Would it be feasible to add a physiological measure to your study? Explain.

Exercise 3.5: Archival Research

Learning Objective: Identify and describe the advantages and disadvantages of archival research.
Review
1. research involves the analysis of existing data or existing records.
2. is existing data that were collected by one researcher or group.

Application
Assume you are still studying student views of the women’s basketball team (from the Application
section of Exercise 3.4).
1. How could archival data help you understand student views of the team as a supplement to
the observational data?

2. Is the archival information secondary data or records and documents?

3. What would be the advantages and disadvantages of conducting archival research on this topic?
Advantages: _______________________________________________________________

Disadvantages: _____________________________________________________________

Exercise 3.6: The Big Picture: How to Choose Measures for Your Study
Learning Objectives: Formulate a big picture of how to choose measures for your study.
Application
Retrieve the article:
Justin, T., Kaszycki, A., Geden, M., & Bunde, J. (2020). Some stress is good stress: The challenge-
hindrance framework, academic self-efficacy, and academic outcomes. Journal of Educational
Psychology, 112(8), 1632–1643. http://dx.doi.org/10.1037/eduedu0000478

Read the introduction and the method sections of the study and use the information provided
to answer the following questions.
1. List the variables in the measures section and the scale of measurement for each one.
26   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. What type of measure (questionnaire, interview, observational, archival, etc.) are each of the
variables?

3. Select one of the variables and describe its operational definition.

4. Based on the literature reviewed in the introduction of the article, how did the researchers
decide on their measures?

Check with your professor for answers to the chapter exercises.

Your Research
Find a primary research article on your topic. Choose 3. What type of measure was the variable (questionnaire,
one variable from the article and answer the following interview, observational, archival, etc.)?
questions.

4. Based on the information in the introduction, what

1. How was the variable operationally defined (or
rationale do the researchers provide for the use
measured)?
of the variable you selected?

2. What scale of measurement (nominal, ordinal, interval,

or ratio) was used to operationally define the variable?
 4
THE CORNERSTONES OF GOOD
RESEARCH
Reliability and Validity

LEARNING OBJECTIVES

• Define reliability and validity as general concepts and specific to measurement.

• Differentiate five types of reliability used to assess measures.
• Distinguish among seven types of validity used to assess measures.
• Discuss how to evaluate the reliability and validity of a study.
• Formulate a big picture of how to evaluate reliability and validity in measures and studies.

CHAPTER SUMMARY
Reliability indicates consistency and validity indicates accuracy.
Chapter 4 introduces these concepts and explains how these terms relate to measurement. The chapter then describes
several ways to calculate and evaluate the reliability of a scale, including internal consistency (coefficient alpha and split-
half ), test-retest, alternative forms, and interrater. Likewise, different ways to assess validity are described, and these can be
conceptualized broadly as face validity versus construct validity. Researchers are more concerned with construct validity.
There are several ways to assess construct validity, including content validity, convergent validity, divergent validity, and
criterion validity (either concurrent or predictive). Examples to illustrate each of the different types of measure reliability
and validity are provided.
Next, reliability and validity at the study level are discussed. Reliability of a study indicates that results have been replicated
in other studies, and validity of a study refers to both internal validity (the amount of control) and external validity (the
generalizability of findings). Researchers should consider the balance between reliability and validity in studies they
conduct as well as in studies they scrutinize before developing their research.

27
28   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

CHAPTER 4 QUIZ
Test your knowledge of the key terms from the chapter.
1. Validity and reliability are dealt with from two perspectives or in regard to
a. personal issues and professional examples.
b. the study level and the measurement level.
c. the manipulation of variables and the data collection.
d. participant confounds and experimenter measures.
2. A scale that measures what it is supposed to measure is
a. reliable.
b. operational.
c. true.
d. valid.
3. The internal consistency of a scale is questioned when
a. α = .80.
b. α < .70.
c. α > .70.
d. α > .90.
4. How is interrater reliability different from the other forms of measurement reliability?
a. It requires ratings by at least two of the participants.
b. It is computed only on forced-choice response formats.
c. It requires that at least two observers or raters provide independent scores.
d. It is measures the global sense of consistency rather than that of individual participants.
5. Researchers are often suspicious of face validity because it is often based on:
a. a single study.
b. multiple studies with different findings but similar procedures.
c. a single person’s view.
d. multiple experts’ views.
6. Divergent validity is supported when a researcher finds a correlation
between the scores of two tests measuring constructs.
a. positive; different
b. positive; similar
c. negative; different
d. no; similar
7. Criterion validity assesses the correlation between
a. a scale and its alternate form.
b. a scale and a behavioral measure.
c. two scales measuring the same concept.
d. the even and odd items in a scale.
 Chapter 4 ■ The Cornerstones of Good Research   29

8. Many studies in psychology use students in lower-level psychology classes as participants. This
fact means the of these studies may be limited to students in these classes.
a. external validity
b. internal validity
c. reliability
d. control
9. A researcher repeats a study in exactly the same way. This is called a and tests
study .
a. literal replication; validity
b. conceptual replication; validity
c. literal replication; reliability
d. conceptual replication; reliability
10. The degree to which we can say that changes in one variable were caused by changes in
another variable is called
a. replication
b. correlation
c. external validity
d. internal validity
Quiz 4 Answer Key
1. b. 2. d. 3. b. 4. c. 5. c. 6. c. 7. b. 8. a. 9. c. 10. d.

CHAPTER 4 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 4.1: Reliability and Validity

Learning Objective: Define reliability and validity as general concepts and specific to measurement.
Review
1. Consistency =
2. Accuracy =
3. In order for a measure or study to have , it must first have
.
Application
1. Imagine that a friend wants to use a stretchy string as a measurement tool to determine
whether a desk will fit in her bedroom. You tell her this is not a good idea because the
stretchiness of the string might lead to different results each time she measures. What
measurement quality is problematic in this situation?
2. Your friend decides instead to use a stick to measure the space and the desk. Based on her
measurements, she determines that the desk will fit. She checks several times and gets the
same result. However, when she tries to put the desk in the space, the desk will not fit.
Explain how this could happen using reliability and validity.
30   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. Imagine that your friend tells you that spinach tastes bad but admits she has never tried it.
You might question the of her conclusion about spinach.
4. Suppose a researcher uses time spent on the website for a research methods class as a measure
of motivation in the class.
a. Is time on the website a reliable measure? Why or why not?

b. Is time on the website a valid measure of motivation? Why or why not?

Exercise 4.2: Assessing the Reliability of Measures

Learning Objective: Differentiate five types of reliability used to assess measures.
Review
1. Complete the following table:

Type of Reliability What Is Correlated?

I.

A. Cronbach’s alpha

B. even and odd items

II.

III. Alternate forms

IV. ratings by two or more observers

Application
1. A political scientist wants to test the measurement reliability of this single item: “On a
scale from 1 to 7, with 7 = very likely, how likely are you to vote in a political election?”
Identify the type of measurement reliability they should assess and explain how they
should assess it.

2. A mental health researcher wants to examine the reliability of a five-item measure of state (or
current) anxiety.
a. What type of measurement reliability would you recommend they assess, and how would
they do it?

b. What type of reliability should the researcher not assess? Explain why not.
 Chapter 4 ■ The Cornerstones of Good Research   31

3. A professor is interested in developing an academic confidence scale.

a. The professor has students complete the academic confidence scale after the first week
of the semester and again in the next class period in the second week. What type of
reliability is the professor assessing?

b. The academic confidence scale has 10 items which are scored on a 5-point Likert-type
scale ranging from 1 = Strongly disagree to 5 = Strongly agree. The professor finds that
Cronbach’s alpha = .80 in the administration during the first week. What would you say
about the internal consistency of the scale?

c. What if the professor had found Cronbach’s alpha = .40? What would you then say about
the scale?

d. How else could the professor measure the internal consistency of the confidence scale?

4. Two observers record whether people avert their gaze when a uniformed police officer enters
the room. What type of reliability should be used to assess the consistency between the two
observers’ scores?
5. Why is it important that the GRE (or LSAT or MEDCAT) have high alternate forms
reliability?

Exercise 4.3: Assessing the Validity of Measures

Learning Objective: Distinguish among seven types of validity used to assess measures.
Review
1. In the following table, outline the specific types of measurement validity and provide a brief
definition.

Measurement Validity Definition

I.

II. Construct validity

A.

B.

C.

D. Criterion validity

i.

ii.
32   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application
1. A political scientist asked, “On a scale from 1 to 7, with 7 = very likely, how likely are you to vote in
a political election?” and followed up with participants to determine whether or not they voted in
the next election. They compared participants’ self-report to their voting behavior 6 months later.
a. What type of validity would be assessed? .
b. Identify a different type of measurement validity the researcher might use to evaluate the
self-report question. Explain how they would evaluate this type of validity.

2. A mental health researcher wants to examine the validity of a five-item measure of state (or
current) anxiety. How could the researcher assess the congruent validity of the measure?

3. A professor is interested in developing an academic confidence scale.

a. How might the professor go about establishing the content validity of the new scale?

b. Suppose the professor finds that students’ scores on the academic confidence scale in the
third week are positively correlated with the number of hours they reported studying that
week. What type of validity has the professor assessed?

Exercise 4.4: Reliability and Validity at the Study Level

Learning Objectives: Discuss how to evaluate the reliability and validity of a study.
Review
1. Replication helps determine the of a study.
2. Explain the difference between literal and conceptual replication.

3. Identify and define the two types of study validity.

4. Which type of validity is especially important in an experiment?

5. In an experiment if there is a variable besides the IV that could have caused changes in the
DV, we say we have a .
6. Which type of validity examines how well results generalize?
7. Having a tightly controlled study improves its but likely impairs its
.
8. A study that takes place in a natural setting and does not involve any manipulation by
researchers has better than .
9. Researchers should always be aware of the balance between and
for their study.
 Chapter 4 ■ The Cornerstones of Good Research   33

Application
List all of the terms below that apply to each of the research examples.
study reliability study validity
literal replication internal validity
conceptual replication external validity
confound
1. A researcher finds that using content-related humor in her class presentations for one section
of a course results in more engagement by students (as measured by attendance and number
of student questions) than in her other section in which she did not include humor. The
researcher isn’t sure that her results were due to the humor she included, and she repeats the
same procedure in her two sessions of the same course the next semester.

2. A colleague in another discipline then decides to try telling jokes in one section of his course
and not in the other. He finds the same result of increased student engagement.

3. Another colleague who tells jokes in one section of his course found no difference
in engagement by the students in the joke versus no joke course sections. In their
end-of-semester feedback, however, students in both sections reported that the
professor sang and danced in all his classes, and the students in both sections rated
him equally funny.

4. A professor tells jokes in an Introduction to Criminal Justice class and the students report that
the professor is funny. The same professor tells the same jokes in an upper-level criminal
justice class and the students report that the professor is not funny.

Exercise 4.5: The Big Picture: Consistency and Accuracy

Learning Objectives: Evaluate studies for evidence of reliability and validity.
Application
Read the following excerpts from two studies and answer the questions following the excerpts. You can
retrieve a copy of the articles if you would like to learn more about either of them.

Torres-Harding, S. R., Andrade, A. L., Jr., & Romero Diaz, C. E. (2012). The Racial Microaggressions
Scale (RMAS): A new scale to measure experiences of racial microaggressions in people of color.
Cultural Diversity and Ethnic Minority Psychology, 18(2), 153–164. http://dx.doi.org/10.1037/
a0027658

Excerpts:

The Racial Microaggression Scale. This scale was developed to assess the themes and
categories of racial microaggressions presented and discussed in the literature. Specific themes
within the larger categories of microinsults and microinvalidations were used as the underlying
basis for the questionnaire items (Sue, Capodilupo, et al., 2007). Members of the research team
developed each item to closely match the themes and categories described in the literature. Then,
members of the research team reviewed items for readability and comprehension. In addition,
for each set of items below, specific statements, thoughts, descriptions, and phrasing expressed or
quoted by participants in the qualitative investigations (cited below) were used to inform item
development. Feedback on the scale was sought from other university students who were familiar
34   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

with the racial microaggressions construct. Similarly, during the administration of the study
measures, feedback was sought regarding reactions to the items from both university students
and community members. (p. 155)

…For each item, respondents were asked to indicate how often they have encountered a
particular racial microaggression on a 4-point Likert-type scale (0 = never, 1 = a little/rarely, 2 =
sometimes/a moderate amount, 3 = often/frequently). (p. 156)

Next, internal consistency of the revised six-factor model was examined using Cronbach’s alphas
for the entire sample (n = 377). The Cronbach’s alphas were found to be very good: Environmental
Invalidations (α = .81); Foreigner/Not Belonging (α = .78); Sexualization (α = .83); Low-Achieving/
Undesirable Culture (α = .87); Criminality (α = .85), and Invisibility (α = .89). (p. 158)

As hypothesized, all of the RMAS (Racial Microaggression Scale, my addition) factor

scores were positively correlated with the subscales of the SRE, a measure of current and lifetime
racial mistreatment and discrimination, which suggests good convergent validity. Similarly, when
comparing scores of White respondents with respondents of color, White persons were less likely
to report experiencing racial microaggressions when compared with persons of color. (p. 161)
1. What is the purpose of the study?
2. What scale of measurement does the RMAS use?
3. Name the type(s) of measurement reliability addressed and describe what Torres-Harding
found.

4. Name the type(s) of measurement validity addressed and describe what Torres-Harding
found.

Pezdek, K., Abed, E., & Cormia, A. (2021). Elevated stress impairs the accuracy of eyewitness memory
but not the confidence–accuracy relationship. Journal of Experimental Psychology: Applied, 27(1),
158–169. http://dx.doi.org/10.1037/xap0000316

Excerpts:

The final sample (N = 106) included 56 participants in the high-stress condition and
50 in the low-stress condition. 2 The mean age did not differ between the high-stress
(mean [M] = 38.1 years, SD = 11.85, range 22– 67) and the low-stress condition (M =
37.5 years, SD = 12.18, range 19 – 64). Both groups were primarily male (male = 57%)
and were primarily Caucasian (76%)...

Materials and procedure. The face stimuli were 48 White male faces selected from a
database of faces used by Meissner, Brigham, and Butz (2005)...

Forty-eight (24 high-stress and 24 low-stress)3 images were sourced from the IAPS (Lang
et al., 2008). IAPS images are normed on arousal and valence. Arousal refers to physical and
 Chapter 4 ■ The Cornerstones of Good Research   35

mental activation and varies from 1 (lowest activation) to 9 (highest activation). Valence refers
to the intrinsic positivity or negativity of an image and varies from 1 (most negative) to 9 (most
positive). Mean arousal was significantly greater for the high-stress (M arousal = 6.47, SD = 0.33)
than the low-stress images (M arousal = 2.96, SD = 0.65), t(66) = 28.10, p < .001, d = 6.81. The
mean valence was also significantly more negative for the high-stress (M valence = 2.34, SD =
0.51) than the low-stress images (M valence = 5.00, SD = 0.08), t(66) = 29.80, p < .001, d = 7.23.
The low-stress IAPS images were specifically selected to be neutral in Experiment 1.

Participants were randomly assigned to either the high-stress or the low-stress

condition. In the presentation phase, participants were presented 24 target slides in
a random order for 8 s each. Each target slide contained an IAPS image (high or low
stress) presented alongside a White male face. Half of the target faces appeared on the
right side, and half appeared on the left side of the screen, and the position of the faces
was counterbalanced so that across participants, each face appeared equally often on
the left and right side of the screen. (p. 160)

Participants were instructed to try to remember both the image and the face because
both would be important later in the study. An old/new recognition-memory test
followed, in which participants viewed 48 test faces (24 old and 24 new). The test faces
were presented in a different random order for each participant, and the 48 faces were
counterbalanced so that across participants, each face served approximately equally often
as a target (old) and foil (new) face. Participants were instructed to respond “old” or
“new” to each test face and to indicate their confidence in each response on a scale of 1
(completely guessing) to 5 (absolutely sure I’m correct). (p. 161)

Summary of Results
Although the study procedure and analysis were much more complicated than this brief description
suggests, the major findings were participants in the high-stress group were significantly less likely than
those in the low-stress group to correctly identify a face. However, correct identification did not differ
between the high- and low-stress groups when participants were highly confident of their choice.
1. What type of study (descriptive, correlation, experiment, quasi-experiment) is described? How
do you know?

2. List each variable in the study and its scale of measurement.

3. What is the operational definition of each variable in the study?

4. Evaluate the study’s external validity based on the information provided.

36   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

5. Evaluate the study’s internal validity based on the information provided.

Check with your professor for answers to the chapter exercises.

Your Research
Find a primary research article on your topic. It may be 4. What evidence, if any, is provided about the measure’s
the article you reviewed for YOUR RESEARCH Chapter 3. validity?
Choose a variable from the article and answer the following
questions.

1. How was the variable operationally defined (or

measured)? Find an article you have recently read that described
an experiment (it might be the study you considered in
questions 1 to 4).

5. Describe how the authors of the article addressed

2. What scale of measurement (nominal, ordinal, interval, or internal and external validity.
ratio) was used to operationally define the variable?

6. What trade-offs did they make in order to achieve an

3. What evidence, if any, is provided about the measure’s acceptable balance between internal and external
reliability? validity?
 5
BASICS OF RESEARCH DESIGN
Description and Sampling

LEARNING OBJECTIVES

• Explain when a descriptive study is appropriate.

• Evaluate the validity of descriptive research.
• Recognize the importance of defining a population and sampling from that population.
• Describe probability sampling and apply three methods to obtain a probability sample.
• Describe nonprobability sampling and apply four methods to obtain a nonprobability sample.
• Formulate a big picture of how to choose a sampling method for your study.

CHAPTER SUMMARY
A descriptive study examines the who, what, when, where, and how, but it does not examine relationships among
them. In this chapter, we first discuss when it is appropriate to use a descriptive study and how to evaluate the validity
of descriptive research. The second half of the chapter focuses on sampling. The chapter explains how to identify a
population and follows with specific procedures for obtaining different types of probability or nonprobability samples
from the population. Issues regarding sample size for probability and nonprobability samples are also discussed. The
chapter ends with an introduction to other types of studies that will be detailed in later chapters.

CHAPTER 5 QUIZ
Test your knowledge of the key terms from the chapter.
1. is how common a behavior, attitude, characteristic, or condition is; a is a
pattern over time.
a. Trend; prevalence
b. Prevalence; trend
c. Population; subpopulation
d. Subpopulation; population

37
38   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. Which of the following types of validity is more important in a descriptive study than in
other types of studies?
a. measurement validity
b. internal validity
c. external validity
d. criterion validity
3. When does a pilot study happen?
a. Before the main study is conducted.
b. After the main study is conducted.
c. While the main study is being conducted.
d. The timing of a pilot study varies.
4. If you are interested in examining how students at your school view a specific sports team, the
for your study is all the students at your school.
a. sampling bias
b. sample
c. population
d. subpopulation
5. In question 4, athletes at your school would be a
a. sampling bias.
b. sample.
c. population.
d. subpopulation.
6. The students who participated in your study on views of the sports team are your
a. sampling bias.
b. sample.
c. population.
d. subpopulation.
7. If athletes are overrepresented in your sample of students, you have a
a. sampling bias.
b. sample.
c. population.
d. subpopulation.
8. A type of probability sampling in which every single member of the population has an equal
chance of being selected for the sample is known as
a. convenience sampling.
b. cluster sampling.
c. random assignment.
d. simple random sampling.
 Chapter 5 ■ Basics of Research Design   39

9. If you rely on participants’ recruiting other participants for you, you are likely using
a. snowball sampling.
b. cluster sampling.
c. random assignment.
d. quota sampling.
10. When a researcher seeks out the full range of extremes in the population for representation in
her sample, this is known as
a. extreme sampling.
b. cluster sampling.
c. maximum variation sampling.
d. wide-range sampling.
Quiz 5 Answer Key
1. b. 2. c. 3. a. 4. c. 5. d. 6. b. 7. a. 8. d. 9. a. 10. c.

CHAPTER 5 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 5.1: When Is a Descriptive Study Appropriate?

Learning Objective: Explain when a descriptive study is appropriate.
Application
1. The following questions are all about the television series Game of Thrones. You may be
familiar with the series or the books upon which the series was based, but you do not need to
know anything about it for this exercise.
Circle all the questions that can be tested with a descriptive study.
a. How many people watched all the seasons of Game of Thrones?
b. Was the first season more popular among younger or older viewers? What about the other
seasons?
c. How did the popularity of the series change from the first to the last season?
d. Who is the most loved character?
e. Does killing off a beloved character cause anxiety among viewers?
f. What did viewers like and dislike about the show?
g. When did most people watch the episodes, as they aired or later?
h. Where did most people watch the episodes, at home or outside the home?
i. What will people think about the show 200 years from now?
2. A study examines work patterns among adults in a community.
a. The study reports the number of people who worked full time from home during July
2020. This is called the of this type of employment.
b. The study also reports the number of people who worked full time from home each July
from 2010 to 2020. This provides the of this type of employment.
40   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. Suppose that a researcher examined weekly exercise routines and mental health issues
including symptoms of depression and anxiety.
a. In the research report, the researcher reported frequency and type of exercise, as well
as frequency and severity of mental health issues within the sample. This represents a
design.
b. In the same research report, the researcher examined the correlation between frequency of
exercise and severity of mental health issues. This represents a design.
c. If the researcher also examined whether or not increased exercise resulted in a decrease in
severity of mental health issues, this would represent a design.
d. Is it reasonable to report results from multiple designs in a single research report? Explain.
_________________________________________________________________________
_________________________________________________________________________

Exercise 5.2: Validity in Descriptive Studies

Learning Objective: Evaluate the validity of descriptive research.
Application
A descriptive study concludes that awareness of racial injustice has increased over the past 5 years.
1. Ask a question about this study that relates to its measurement validity.
_________________________________________________________________________
_________________________________________________________________________
2. Ask a question about this study that relates to its external validity.
_________________________________________________________________________
_________________________________________________________________________
3. Explain why internal validity is not that relevant to this study.
_________________________________________________________________________
_________________________________________________________________________

Exercise 5.3: Defining the Population and Obtaining a Sample

Learning Objective: Recognize the importance of defining a population and sampling from that
population.
Application
Imagine that you are conducting a study on perceptions of racial injustice.
1. Identify a population for your study:
2. Identify a subpopulation for your study:
3. Explain why sampling from the population is a better idea for your study than attempting to
obtain data from every member of your population.
_________________________________________________________________________
_________________________________________________________________________
4. Identify one situation that would make you worry that you have a sampling bias.
_________________________________________________________________________
 Chapter 5 ■ Basics of Research Design   41

Exercise 5.4: Probability Sampling

Learning Objective: Describe probability sampling and apply three methods to obtain a probability
sample.

Review
1. A classmate tells you “random assignment” is essential to probability sampling. Help explain
why the term “random assignment” is inappropriate in this context and identify the correct
term to use instead:
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
2. Explain the difference between random sampling with replacement and random sampling
without replacement.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________

Application
1. Imagine you wanted to interview the actors who were in the first season of the television series
Game of Thrones. You know that it would not be possible to interview all the actors, so you
decide to use probability sampling in order to obtain a representative sample.
You define the population as all the actors who played named characters during the first
season (so you would exclude extras who did not play a character with a name).
You obtain a list of all the named characters from online sources.
a. What would be the next steps if you used simple random sampling?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
b. Suppose you wanted your sample to accurately represent the proportion of female actors
and male actors in the first season. What type of probability sampling would you use and
how would you obtain your sample?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
c. There were 129 named characters on the first season of Game of Thrones.
i. Use either the formula in Chapter 5 of the textbook or the online calculator found at
www.surveysystem.com/sscalc.htm to calculate the minimum sample size you would
need to have a confidence interval of 5 (meaning that the margin of error will be
about 5%) and a confidence level of 95% (meaning that there is a 95% likelihood that
you will achieve your desired margin of error).
ii. If you increased your confidence interval to 12 and kept your confidence level at
95%, what is the minimum sample size you would need?
42   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. Imagine you want to conduct a study about the political attitudes of adults who live in your
neighborhood. You realize you cannot interview everyone in your neighborhood and instead
decide to randomly select streets within the neighborhood.
a. What type of sampling is this?
b. What would be your next steps?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________

Exercise 5.5: Nonprobability Sampling

Learning Objective: Describe nonprobability sampling and apply four methods to obtain a nonprobability
sample.
Application
Imagine again that you want to interview actors on the first season of Game of Thrones, but now you
decide to use nonprobability sampling.
1. What is the advantage of using nonprobability sampling in this example?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
2. How would you obtain a sample using convenience sampling?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
3. How would you obtain a sample using snowball sampling?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
4. Suppose you wanted your sample to accurately represent the proportion of female actors and
male actors in the first season. What type of nonprobability sampling would you use and how
would you obtain your sample?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
5. Suppose you wanted your sample to represent extremes in the population, such as level of
fame, how much screen time the actor got, whether or not the character died, etc. What type
of nonprobability sampling would you use and how would you obtain your sample?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
 Chapter 5 ■ Basics of Research Design   43

Exercise 5.6: The Big Picture: Choosing a Sampling Method

Learning Objective: Formulate a big picture of how to choose a sampling method for your study.
Review
Identify the specific type of probability or nonprobability sampling described.

Description Probability Sampling Nonprobability Sampling

Uses random selection? (yes or no)

Most basic method

Sample proportions of a key characteristic

match population proportions

Identify clusters instead of individuals within not applicable

population
Sample represents extremes within a not applicable
population
Participants recruit others not applicable

Check with your professor for answers to the chapter exercises.

Your Research
Designing a Descriptive Study on Your Topic 3. Choose a specific sampling procedure. Define the
population and then outline how you will obtain a sample
1. Identify one question that you could examine with a
from your population.
descriptive study and that you believe is important to
study (based on what you know about existing research in
this area).

2. Would it be better to use probability or nonprobability

sampling? Explain.
 6
DESCRIBING YOUR SAMPLE

LEARNING OBJECTIVES

• Summarize the ethical and practical considerations of describing your sample.

• Identify and calculate different types of descriptive statistics.
• Recommend and generate appropriate statistics and graphs based on the type of data.
• Compare interval or ratio data using z scores and percentiles.
• Formulate a big picture of how to describe your sample.

CHAPTER SUMMARY
Describing a sample requires critical thinking about both ethical and practical issues in choosing the characteristics of
the sample to describe. Quantitative (numerical) analyses are a popular way to summarize sample characteristics, and
the majority of the chapter focuses on descriptive statistics. Specific descriptive statistics are organized by their overall
purpose: to describe how often a score appears in the sample; to describe the central tendency; and to describe the
variability. Definitions, formulas, and example calculations are included for each type of descriptive statistic. Note
that statistical notations that are most common in published research are used throughout the book (e.g., M and SD).
In this chapter, an emphasis is placed on not only understanding different types of descriptive statistics and their purpose,
but also choosing the appropriate descriptive statistics based on the scale of measurement and, for interval or ratio
variables, the shape of the distribution of scores. The chapter ends with a reminder of the importance of thinking critically
about the quality of your data, the characteristics of your sample, and the meaning of your results.

CHAPTER 6 QUIZ
Test your knowledge of the key terms from the chapter.
1. If you wanted to know what proportion of your sample consisted of left-handed people, you would be
wondering about the
a. frequency.
b. cumulative percentage.
c. mean.
d. percentage.

44
 Chapter 6 ■ Describing Your Sample   45

2. The point in a distribution at which 50% of the scores are below and 50% of the scores are
above is known as the
a. standard deviation.
b. mode.
c. median.
d. mean.

3. The most frequent score in a distribution is the

a. standard deviation.
b. mode.
c. median.
d. mean.

4. When you report the mean, you should also report the that tells you the
degree to which scores differ from the mean.
a. standard deviation
b. median
c. range
d. possible minimum and maximum

5. If you reported that participants in your study had ages between 18 and 64, you would be
reporting the
a. standard deviation.
b. observed minimum and maximum.
c. possible minimum and maximum.
d. mean.

6. Which of the following is NOT a characteristic of a normal distribution?

a. It is shaped like an upside-down bell or U shape.
b. It is symmetrical.
c. The majority of the scores are centered around the middle.
d. The scores taper off at both ends.

7. A narrow and peaked distribution is called

a. kurtosis.
b. leptokurtic.
c. mesokurtic.
d. platykurtic.

8. When scores cluster at one end of the distribution, it is said to be

a. skewed.
b. uniform.
c. bimodal.
d. an outlier.
46   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

9. Which of the following is the best statistic to evaluate an individual score within a distribution?
a. mean score
b. standard deviation
c. observed maximum score
d. z score
10. What is the difference between a percentage and a percentile?
a. A percentage is the proportion of a score in a sample whereas a percentile is the proportion
scores that fall under a set value.
b. A percentage is the proportion of a score in a sample whereas a percentile is the percentage
of a score in the sample between a set interval.
c. A percentage can link directly to a letter grade whereas a percentile cannot.
d. There is no difference.

Quiz 1 Answer Key

1. d. 2. c. 3. b. 4. a. 5. b. 6. a. 7. b. 8. a. 9. d. 10. a.

CHAPTER 6 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 6.1: Ethical and Practical Issues in Describing Your Samples

Learning Objective: Summarize the ethical and practical considerations of describing your sample.
Application
Imagine you are tasked with finding out how satisfied undergraduates at your school are with their major.
1. List steps you would take to maintain the confidentiality of your participants.

2. In addition to asking about satisfaction, what questions would you include to describe your sample?

Exercise 6.2: Descriptive Statistics

Learning Objective: Identify and calculate different types of descriptive statistics.
Application
1. A market researcher asked 20 participants whether they prefer to shop for clothes online, in a
store, or using a paper catalog:

a. What is the frequency for online, store, or catalog preference?

b. What percentage chose shopping online as their preference?

c. What is the mode?
 Chapter 6 ■ Describing Your Sample   47

Participant # Shopping Preference Participant # Shopping Preference

1 Store 11 Online

2 Catalog 12 Online

3 Store 13 Store

4 Store 14 Store

5 Store 15 Online

6 Online 16 Store

7 Store 17 Catalog

8 Online 18 Online

9 Online 19 Store

10 Store 20 Store

2. A highly selective university honors society only recruits incoming students who are in the top
10 of their class in high school. Following are the rankings of their 15 members, in order from
lowest to highest ranking:

7, 7, 5, 5, 4, 3, 3, 3, 2, 1, 1, 1, 1, 1, 1

a. What percentage was ranked first in their high school class?

b. What is the median?
c. What is the possible minimum score (lowest ranking)?
d. What is the observed minimum score (lowest ranking)?
e. What is the range?

3. A professor asks students how many hours they predict they will need to study for an
upcoming exam, with the following results: 1, 2, 2, 2, 3, 3, 3, 3, 3, 4, 4, 5
a. Complete the frequency table:

Hours (predicted) f

b. Calculate by hand and report the mean of these data. Show your work.
c. Calculate by hand and report the standard deviation for these data. Show your
work.
48   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

4. After the exam, the professor asks the students how many hours they actually spent studying,
with the following results: 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 4, 8
a. Complete the frequency table:

Hours (actual) f

b. Calculate the cumulative percentage of those who studied only one or two hours. Show
your work.

Exercise 6.3: Choosing the Appropriate Descriptive Statistics

Learning Objective: Recommend and generate appropriate statistics and graphs based on the type
of data.
Review
Complete the following table:

Descriptive Statistics Descriptive Statistics You

Type of Data You Should Report Might Report If Useful

Nominal data

Ordinal data

Normally distributed interval or ratio data

Skewed interval or ratio data

Bimodal interval or ratio data that can be

split into two distinct and meaningful groups

Bimodal interval or ratio data that cannot

be split into two distinct groups

Uniform distribution

Application
1. Refer to the data and your answers from Exercise 6.2, question 1 (shopping preferences).
a. What type of graph is appropriate to depict this data? Explain why.
__________________________________________________________________
__________________________________________________________________
 Chapter 6 ■ Describing Your Sample   49

b. Draw the graph to summarize the results. Be sure to label your axes.

c. Write a sentence or two describing the results as you would in a Results section of an
APA-style report. ________________________________________________________
______________________________________________________________________
______________________________________________________________________
2. Refer to your answers from Exercise 6.2, question 2 (class rankings). Recall that the data are:
7, 7, 5, 5, 4, 3, 3, 3, 2, 1, 1, 1, 1, 1, 1
a. What scale of measurement is represented?
b. Write a sentence or two describing the results as you would in a Results section of an
APA-style report. ________________________________________________________

3. Refer to your answers from Exercise 6.2, question 3 (predicted study hours). Recall that the
data are: 1, 2, 2, 2, 3, 3, 3, 3, 3, 4, 4, 5
a. What two types of graphs are appropriate to depict this data? Explain why.

b. Draw a graph to summarize the results. Be sure to label your axes.

c. Does the distribution appear to be normally distributed? Explain.

___________________________________________________________________________
d. Write a sentence or two describing the results as you would in a Results section of an
APA-style report. ________________________________________________________
___________________________________________________________________________
___________________________________________________________________________
50   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

4. Refer to the data and your answers from Exercise 6.2, question 4 (actual study hours). Recall
that the data are: 1, 1, 1, 2, 2, 2, 2, 3, 3, 3, 4, 8
a. Draw a histogram to summarize the results. Be sure to label your axes.

b. Does the distribution appear to be normally distributed? Explain.

c. What statistic should you calculate to confirm your assessment of the shape of the
distribution?
d. Assuming you confirm your assessment of the shape of the distribution, what measure of
central tendency and variability should you report?

e. Calculate by hand and report the central tendency:

f. Calculate by hand and report one appropriate measure of variability.

g. Write a sentence or two describing the results as you would in a Results section of an
APA-style report.
______________________________________________________________________
______________________________________________________________________

Exercise 6.4: Comparing Interval/Ratio Scores With z Scores and

Percentiles
Learning Objective: Compare interval or ratio data with z scores and percentiles.
Application

1. Professor K converts all his students’ test scores to z scores. Julie has a z score of 0.5 and Alice
has a z score of –2.7. How would you help Julie and Alice interpret their scores?

2. What is the difference between earning 95% on an exam and being in the 95th percentile on
an exam?

3. If you scored in the 50th percentile on an exam, how would you interpret your performance?
What is your z score?
 Chapter 6 ■ Describing Your Sample   51

Exercise 6.5: The Big Picture: Know Your Data and Your Sample
Learning Objective: Formulate a big picture of how to describe your sample.
Review

1. List five things that might negatively impact your ability to accurately interpret your
descriptive statistics.

2. List at least three considerations to help you be respectful to your sample.

3. How does the sampling technique impact how you interpret descriptive statistics?

Your Research
Find a primary research article on your topic. Look through 3. Did the author(s) report the central tendency and
the Method and Results sections. variability to describe any of the variables? If so, explain
what variables and what measures of central tendency
1. Did the author(s) use frequencies to describe a
and variability the author(s) reported.
characteristic of the sample? If so, explain.



 

 

 

2. Did the author(s) use percentages to describe a 4. Did the authors use z scores or percentiles? If so,
characteristic of the sample? If so, explain. explain.

 

 

 

 
 7
BEYOND DESCRIPTIVES
Making Inferences Based
on Your Sample

LEARNING OBJECTIVES

• Explain how inferential statistics are used to determine whether the finding of a study is unusual.
• Illustrate the hypothesis testing process.
• Describe when to reject or retain a null hypothesis and the types of errors associated with each of these
decisions.
• Distinguish between statistical significance, effect size, confidence intervals, and practical significance.
• Formulate a big picture of how to interpret a study’s results.

CHAPTER SUMMARY
The chapter is focused on inferential statistics that allow us to make inferences about a population from findings with a
sample. It introduces many terms and the process that you will use when analyzing a study’s quantitative data. Sampling
distributions are a distribution of some statistic and provide the basis for deciding whether our sample is representative (or
not) of the population—or when a difference makes a difference. The decision-making process is called hypothesis testing
(or null hypothesis significance testing) and involves multiple steps: stating null and alternative hypotheses, defining the
sampling distribution, setting the criterion level (.05 or .01), computing a statistic, and deciding whether to reject or
retain the null hypothesis. The alternative hypothesis determines the region of rejection in the sampling distribution and
can be one-tailed (directional) or two-tailed (nondirectional). If the results of our analysis fall in the region of rejection,
we reject the null hypothesis and support the alternative hypothesis, but if the statistic falls in the region of acceptance,
we retain the null hypothesis.
We design studies to try to reject a false null hypothesis (called power), but because hypothesis testing is based on
probability, the decision to reject or retain the hypothesis is associated with a probability of error. Type I error occurs
when we reject a true null hypothesis, and Type II error occurs when we retain a false null hypothesis. Ways to decrease
the probability of either type of error, as well as ways to increase power, are described. Finally, effect sizes (magnitude of
an effect), confidence intervals (margin of error), and practical significance (everyday implications) are described as tools
to enhance our understanding of the results of a study.

52
 Chapter 7 ■ Beyond Descriptives   53

CHAPTER 7 QUIZ
Test your knowledge of key terms from the chapter.
1. The purpose of inferential statistics is to allow researchers to determine whether .
a. the computed statistic is reliable
b. a study has internal validity
c. an outcome can be generalized to samples of different sizes
d. the outcome of a study is typical or unusual for a particular sampling distribution
2. The process of determining the probability of obtaining a particular result is called .
a. hypothesis testing
b. gambling
c. theoretical possibilities
d. statistical analysis
3. A sampling distribution is created by .
a. collecting samples of various sizes from a population
b. drawing many samples of the same size from a population and computing the same
statistic for each sample
c. repeating a study multiple times with the same size samples from different
populations
d. drawing different size samples from a population and computing their mean
4. In an experiment the alternative hypothesis predicts .
a. a difference between groups
b. no difference between groups
c. a difference between the sampling distribution and a population
d. the degree of similarity between groups
5. When the results for a study fall in the region of rejection it implies that .
a. our sample does not belong to the sampling distribution as defined by the null hypothesis
and we should reject the null hypothesis
b. our sample does not belong to the sampling distribution as defined by the alternative
hypothesis and we should retain the null hypothesis
c. our sample belongs to the sampling distribution and the researcher should retain the null
hypothesis
d. there was probably a confound in the study and we should reject the alternative hypothesis
6. The critical value of a statistic defines the .
a. top 2.5% of the distribution for a one-tailed hypothesis
b. bottom 2.5% of the distribution for a one-tailed hypothesis
c. extreme 5% of the distribution for a one-tailed hypothesis and extreme 2.5% of the
distribution for a two-tailed hypothesis
d. extreme 2.5% of the distribution for a one-tailed hypothesis and extreme 5% of the
distribution for a two-tailed hypothesis
54   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

7. A Type I error occurs when you .

a. reject a true null hypothesis
b. fail to reject a true null hypothesis
c. retain a false null hypothesis
d. fail to reject a true alternative hypothesis
8. Power is defined as .
a. correctly rejecting a true null hypothesis
b. correctly rejecting a false null hypothesis
c. correctly retaining a true null hypothesis
d. correctly rejecting a true alternative hypothesis
9. The effect size refers to the while the confidence interval refers to the
.
a. significance level; probability of a Type I error
b. probability of a Type II error; power
c. margin of error; strength of the effect
d. strength of the effect; margin of error
10. Practical significance refers to .
a. the magnitude or strength of the effect of a variable
b. the usefulness of our results in everyday life
c. our ability to use the results in another study
d. whether the participants found the study useful
Quiz 7 Answer Key
1. d. 2. a. 3. b. 4. a. 5. a. 6. c. 7. a. 8. b. 9. d. 10. b.

CHAPTER 7 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying the key concepts.

Exercise 7.1: Inferential Statistics

Learning Objective: Explain how inferential statistics are used to determine whether the finding of a study
is unusual.
Review
1. Fill in the table:

Descriptive Inferential

Sample

Parameter

mean (M)

sigma (s)
 Chapter 7 ■ Beyond Descriptives   55

2. Inferential statistics are based on theory, which examines random

events, such as what card you will draw in poker.
3. How does probability theory help a researcher to answer the question, “When does a
difference in scores make a difference?”

Exercise 7.2: Hypothesis Testing

Learning Objective: Illustrate the hypothesis testing process.
Review
1. A describes a distribution of statistics while a
describes a distribution of scores.
2. The hypothesis predicts what you expect to find in your study.
3. The hypothesis predicts no difference between groups.
4. A study is always designed to the null hypothesis.
5. The typical region of rejection is the extreme % of the sampling
distribution, but sometimes we use the stricter extreme % to define
the region of rejection.
6. Statistical significance suggests that your results are not due to and
that the results (do/do not) belong to the sampling distribution
implied by the null hypothesis.
7. Why is a two-tailed test a more conservative test than a one-tailed test?

Application
1. Suppose you are interested in comparing the number of crimes committed by those in
urban areas versus those in rural areas. Name the type of hypothesis for each statement
below, and, if the statement is an alternative hypothesis, identify whether it is directional or
nondirectional.
a. “There will be no difference in the number of crimes committed by those in urban areas
and those in rural areas.”

b. “The number of crimes committed in urban areas will differ from the number committed
in rural areas.”

c. “The number of crimes committed in urban areas will be greater than the number
committed in rural areas.”

2. Define a sampling distribution for the following hypothesis: There is no difference in the
percentage of income spent on housing by elderly households and the percentage of income
spent on housing by all households. Hint: First name the percentage spent by all households
(estimate a reasonable percentage).
56   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. In the figures below, the numbers on the x-axis are standard deviation units away from the
mean. For example, −2 = 2 SD’s below the M, while 1 = 1 SD above the M.

A B

–3 –2 –1 0 1 2 3 –3 –2 –1 0 1 2 3

Figure A: Draw the region of rejection and region of acceptance for a two-tailed test where p < .05.
Figure B: Draw the region of rejection and region of acceptance for a one-tailed test where p < .05.
4. List the seven steps that you would follow to test the following hypothesis: “Males are more
likely than females to physically bully their classmates.” Make your steps relevant to this
specific study (e.g., state the null and alternative hypotheses for this study).

a.

b.

c.

d.

e.

f.

g.

Exercise 7.3 Errors in Hypothesis Testing

Learning Objective: Describe when to reject or retain a null hypothesis and the types of errors associated with
each of these decisions.
Application
1. Which of the following p values would meet the criteria for statistical significance at p < .05?
Circle them.
.001 .05 .006 .50 .047 .20 .02 .07 .70
2. Regardless of your decision to reject or retain the null hypothesis, there is always a chance of
error because statistics are based on .
3. Why can’t you have a probability of both a Type I and a Type II error in a study?

4. You can reduce the probability of a Type II error by (increasing/decreasing)

a. sample size
b. error in your research design
c. strength of the effect
 Chapter 7 ■ Beyond Descriptives   57

5. If you find that, in your sample, girls verbally bully their classmates more often than boys (p = .04),
a. Are the results statistically significant?
b. Would you reject or retain the null hypothesis?
c. What is the probability of a Type I error? Type II error?
d. Name two ways you could reduce your chance of a Type I error in the study.
i.
ii.
6. If you find, in your sample, that males physically bully their classmates more often than girls
(p = .07),
a. Are the results statistically significant?
b. Would you reject or retain the null hypothesis?
c. Is there a probability of a Type I error? Type II error?
7. Two studies were conducted on bullying at an elementary school. Study 1 selected a
convenience sample of 10 boys and 10 girls from grades K through 5. Study 2 selected a
convenience sample of 25 boys and 25 girls from the fourth grade.
Which study is likely to have more power? Why? (Hint: There are two
reasons.)
a.
b.

Exercise 7.4: Effect Size, Confidence Intervals, and Practical Significance

Learning Objective: Distinguish between statistical significance, effect size, confidence intervals, and practical
significance.
Review

1. Draw a line from the term/phrase in first column that best matches the one in the second
column.

Effect size Margin of error

Cohen’s d “These findings have implications for the way people view
their personal responsibility for their health.”
Percentage of variability effect size
Magnitude of an effect
Confidence interval
Effect size measured in standard deviation units
Practical significance
18% of the variability in anxiety was accounted for by stress

2. Fill in the table, but remember that these numbers are guidelines and not designed to be strict
cutoffs.

Effect Size: Proportion of

Interpretation of Effect Size Effect Size: Cohen’s d Variability Accounted For

Small/Weak

Medium/Moderate

Large/Strong
58   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application
1. Interpret each of the following effect sizes as weak, moderate, or strong:
a. 2% of the variability accounted for 
b. 50% of the variability accounted for 
c. d = 0.48 
d. d = 1.20 
e. d = 0.12 
f. 10% of the variability accounted for 
g. d = 0.25 
2. A teacher reports that exam grades ranged from 64% to 98%, with a mean of 79.85
(SD = 10.18), 95% CI [75.09, 84.61]. The confidence interval tells you that you are
confident that the sample mean of represents a population mean that falls
between and .
3. Another teacher reports exam scores based on academic year.
First-year students: M = 75.25, SD = 5.16, 95% CI [71.70, 78.88]
Second-year students: M = 80.33, SD = 7.33, 95% CI [74.11, 86.55]
a. Do the confidence intervals for the population means overlap?
b. A second-year student brags that their year outperformed first-year students. Based on the
confidence intervals, how might you counter this student’s claim?



c. What would you say about the practical significance of these results?




Exercise 7.5: The Big Picture: Making Sense of Results

Learning Objective: Formulate a big picture of how to interpret a study’s results.

Review

What are the benefits of reporting each of the following in describing the results of a study?
1. Statistical significance 

2. Effect size 


3. Confidence interval 


 Chapter 7 ■ Beyond Descriptives   59

4. Practical significance 



Application
1. A study examined whether convenience affected recycling behavior. The researchers
operationally defined recycling behavior based on the weight (in pounds) of material
recycled by each household. They found that households who were supplied a large
recycling bin and weekly pickup (M = 25.00, SD = 5.00, 95% CI [20.25, 29.75]) recycled
significantly more material than the households who were sent flyers encouraging them to
recycle (M = 10.00, SD = 4.00, 95% CI [6.50, 13.50], p = .003). Convenience accounted for
22% of the variance in recycling.
a. Explain what the p value indicates and how you would interpret it.


b. Interpret the effect size.


c. Interpret the confidence intervals.


d. Interpret the practical significance of the study.


2. The researchers replicated the study in a different city and found that those supplied with a
recycling bin (M = 25.00, SD = 7.00, 95% CI [16.20, 33.80]) recycled more than those who
were just given an informational flyer (M = 10.00, SD = 5.50, 95% CI [2.50, 17.50]). This
time, they found that p = .10, Cohen’s d = 0.25.
a. Explain what the p value indicates and how you would interpret it.


b. Interpret the effect size.


c. Interpret the confidence intervals.


d. Interpret the practical significance of the study.


60   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. Which of the two recycling studies

a. Found statistically significant results? Explain your answer.


b. Had a stronger effect size? Explain your answer.




c. Has more practical significance? Explain your answer.







Your Research
Find a primary research article on your topic. If possible, 1. State the research question that interests you and
find an article that has a simple design. If the article explain why the article you selected is related to this
describes multiple studies, pick one of the studies as the question.

focus of this exercise. Before beginning to work on this 


section, look through the entire article.

You will see that in order to answer these questions, you will
2. Focus on the Introduction of the article.
have to read the article very carefully and probably several
a. Identify one hypothesis that most interests you.
times. This practice should give you some idea of the care
(Remember that it is important to learn how
that you should take in digesting research. Even though you
to write about research in your own words. Be
need to take a lot of time to read articles, you may end up
sure that you do not plagiarize in writing the
summarizing the article in your own literature review with
hypothesis!)
a single sentence or by simply citing the article because of 
the measure you use from the study or for the procedure you 
adopt. But by attending to the details of published articles, 
you will learn a great deal about the format for reporting
b. Is the hypothesis directional or nondirectional?
research, different methods for conducting research, various How do you know?
ways to analyze data, and appropriate interpretations of 
statistics. All of this will support you as you learn to design, 
carry out, analyze, and report your own studies. 
 Chapter 7 ■ Beyond Descriptives   61

3. Now examine the Method section. d. Did the authors report an effect size?
What do you learn about the sample and procedure of If yes, list it here and interpret its meaning.
the study that may help the researchers to avoid Type II 
errors (e.g., sample size, error in the research design, 
strength of effect)? You may want to review the section
e. Did the authors include a confidence interval for
“Reducing the Chance of a Type II Error” in Chapter 7
their findings? . If yes, list the confidence
before answering this question.
interval(s) and interpret the meaning.
Sample: 




 

Research design: 5. Reread the Discussion section.

a. Did the authors indicate that they found support


for the hypothesis you selected?

 

Strength of effect:
b. How do they interpret the meaning of the support
 or nonsupport for their hypothesis?
 


4. Review the Results section of the article. Even though 
there may be statistics you do not understand, focus on 
the statistics that you are familiar with: M, SD, range,
c. What do the researchers state or imply about the
possible and observed scores, p, percentage of variability
practical significance of their study? They may not
accounted for (could be noted as r2,rbp2, η2 ), Cohen’s
use the term practical significance; instead, they
d, confidence intervals. Consider the hypothesis you
may discuss the implications or usefulness of their
selected in question 2 and answer the following questions
results.
regarding the analyses used to test this hypothesis.

a. Did the author(s) report any descriptive statistics 
(e.g., M, SD) related to the hypothesis you 
selected? If yes, explain what the 
descriptive statistics tell you about the variable(s)
under examination. d. Researchers also discuss limitations of their
 study in the Discussion. Do you see any limitations
 they mention that you can relate to the possibility
 of Type I or Type II errors? How do the authors
 suggest that future researchers should address or
overcome these limitations?
b. Did the researchers find statistically significant

results for the hypothesis you noted in question 2?

Explain.






As you read other primary research articles on your
c. Is there a probability of a Type I error?
topic, we recommend you use some or all of these
If yes, what is the exact probability? Is
questions to help you evaluate the research.
there a probability of a Type II error?
 8
COMPARING YOUR SAMPLE TO A
KNOWN OR EXPECTED SCORE

LEARNING OBJECTIVES

• Describe the process of choosing the appropriate inferential statistic to compare a sample and
known expected value.
• Compare a sample to a known population value when you have interval or ratio data using a one-sample
t test.
• Compute the two types of effect size when comparing a sample to a known population.
• Compute the confidence interval when comparing a sample to a known population value.
• Formulate a big picture of how to examine one variable at a time.

CHAPTER SUMMARY
The chapter describes the statistics used to compare a sample score to a known or expected score. The appropriate
statistical test depends upon the scale of measurement. The chi-square goodness of fit is the appropriate test for nominal
data, and the one-sample t test is the appropriate test for interval or ratio data. This chapter focuses on the one-sample
t test. Using an example comparing a sample of first-year students’ scores on a citation format knowledge quiz to the
national average for the quiz, the steps of the hypothesis testing process are described. For the statistical analysis, you
learned how to compute a one-sample t test, which is used when you have interval or ratio data. Then two options for the
effect size (eta squared and Cohen’s d) were covered. Finally, you learned how to compute a 95% confidence interval for
the difference between the sample mean and the known mean in a study. This section concluded with instruction about
how to write up Results and Discussion sections for a one-sample study when you have computed these statistics.

CHAPTER 8 QUIZ
Test your knowledge of key terms from the chapter.
1. If we want to find whether the first-year class at our college has higher scores on the SAT writing test than the
national average for the test, we would compute .
a. the effect size
b. practical significance
c. a one-sample t test
d. an alternative hypothesis

62
 Chapter 8 ■ Comparing Your Sample to a Known or Expected Score   63

2. You want to find out whether a significantly greater percentage of students at your university
binge-watch television series than the national percentage for binge-watching. Can you
compute a one-sample t test on these data?
a. Yes, you have sample data and a national value for comparison.
b. No, the data are a nominal scale and a chi-square goodness of fit test is
needed.
c. No, students cannot be legitimately compared to a national sample of all television
viewers.
d. You do not have enough information to make a decision.

3. In order to compute a one-sample t test we must .

a. collect data from a sample of at least 100
b. measure our variable on an interval or ratio scale
c. randomly select our sample
d. compute the expected frequency of our sample

4. The estimated standard error of the means (SEM) is

a. a standardized z score on a normal distribution.
b. an approximation of the standard deviation of a sampling distribution
of means.
c. the population mean minus the sample mean divided by 100.
d. an estimate of the standard deviation of a frequency distribution.

5. The standard deviation of a sampling distribution of means is called the

.
a. eta squared (variability accounted for)
b. variance
c. the standard error of the means (σx)
d. none of the above

6. In computing a one-sample t test, we divide the difference between our sample

mean and the population mean by the estimated standard error of the means because
.
a. the standard deviation of the sample is too small to use
b. we do not know the standard deviation of the sampling distribution and have to estimate
it using the standard deviation of the population
c. we do not know the standard deviation of the sampling distribution and have to estimate
it using the standard deviation of our sample
d. the standard error is greater than the standard deviation of the sample

7. Degrees of freedom represent the .

a. number of scores free to vary in a sample
b. estimated sample size
c. percentage of variability accounted for by the population measure
d. distance between the sample mean and population mean divided by the sample size
64   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

8. The two types of effect size that are computed for a one-sample t test are
and .
a. Cohen’s d; eta squared (η2)
b. estimated sample size (SEM); percentiles
c. percentage of variability accounted for; Cohen’s df
d. eta squared (η2); confidence intervals
9. Typically we compute the confidence interval.
a. 10%
b. 50%
c. 75%
d. 95%
10. A researcher compared the average salary for a sample of first-generation college graduates
with the average for all college graduates one year postgraduation and found t(99) = –1.59,
p = .07, d = .62. Is there a possibility of a Type II error?
a. Yes, the probability of a Type II error is 7%.
b. Yes, but we do not know the exact probability.
c. No, you cannot have a Type II error when the null hypothesis is rejected.
d. No, you cannot have a Type II error when the effect size is moderate or large.

Quiz 8 Answer Key

1. c. 2. b. 3. b. 4. b. 5. c. 6. c. 7. a. 8. a. 9. d. 10. b.

CHAPTER 8 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying the key concepts.

Exercise 8.1: Choosing the Appropriate Test

Learning Objective: Describe the process of choosing the appropriate inferential statistic to compare a sample
and known expected value.
Review
List the appropriate statistical test (z score, one-sample t test, chi-square goodness of fit) for each of the
following situations.
Statistical Test
1. c ompare a sample mean to a known population mean
or to an expected mean
2. c ompare a raw score to a population or very large
sample mean
3. c ompare observed frequencies to known or expected
frequencies
4. v ariables measured on a nominal scale of
measurement
5.
or v ariables measured on an interval or ratio scale of
measurement
 Chapter 8 ■ Comparing Your Sample to a Known or Expected Score   65

Exercise 8.2: One-Sample t Tests

Learning Objective: Compare a sample to a known population value when you have interval or ratio data
using a one-sample t test.

Review
1. The assumptions of a one-sample t test include the following:
a.
b.
c.

2. In order to calculate a one-sample t test, we must have four statistics. Name them and note
their symbol.

STATISTIC SYMBOL

3. The t value is the value that defines the region of rejection for the
sampling distribution if the null hypothesis is true.
4. To be statistically significant, the calculated or computed t value must be
(greater than/less than) the critical t value.

Application
1. Do fathers who have completed a parenting class spend more time each day talking to their
toddlers than the average number of minutes reported in a national survey of fathers?
a. The researcher should compute a (type of statistical test).
b. Should your alternative hypothesis for this study be directional or nondirectional?
Explain why.

c. State an appropriate alternative hypothesis for this study:

d. State an appropriate null hypothesis for this study:

2. A professor finds that the age of students in her class is skewed, and she wants to compare
the median age in her class to that of the university. Why should the professor not use a one-
sample t test to analyze this data?
66   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. A political scientist asks voters leaving the polling center for whom they voted in the city
council election; she plans to compare her data to the actual voting outcome. Why should the
researcher not use a one-sample t test to analyze this data?

4. Suppose you found for the study of fathers in Question 1 the result of t(14) = 3.12.
a. The number of participants in this study is . How do you know?

b. The critical t value for this finding for a two-tailed test at the .05 level is .
(Hint: Use Table C.4 in Appendix C of the textbook.)
c. Would your result be significant for p < .05 for a two-tailed test?
d. How do you know?

e. How do you interpret these results?

5. Suppose you found for a study that t(29) = 2.50.

a. Would this result be significant for p < .01 for a two-tailed test?
How do you know?

b. Would this result be significant for p < .01 for a one-tailed test?
How do you know?

c. Why is there a difference in your conclusions for the one- and two-tailed tests?

6. After surveying drivers at multiple sites throughout the country, the National Transportation
Board reports that drivers report using their cell phones during 60% of their trips in their car.
You think that better-educated people will use their cell phones less frequently while driving.
You ask 25 of your classmates during what percentage of trips in their car (of any duration) do
they use their cell phones. Your classmates report a mean of 52%, with a standard deviation of
10%. Do your classmates use their cell phones significantly less than the general population
when driving?
a. State your null hypothesis.
H0: ___________________________________________________________________

b. State a directional alternative hypothesis.

Directional Ha:

c. df =
 Chapter 8 ■ Comparing Your Sample to a Known or Expected Score   67

d. Calculate the appropriate statistical test.

e. Can you reject the null hypothesis? Why or why not?

______________________________________________________________________
______________________________________________________________________
f. Is there a chance of a Type I error? ; a Type II error?

Exercise 8.3: Calculating an Effect Size

Learning Objective: Compute the two types of effect size when comparing a sample to a known
population.

Application
1. A researcher compares the number of calories consumed daily by a sample of obese
5-year-olds to the recommended calories for this age group. As part of her analysis, she
finds η2 = .15.
a. The name of this test is .
b. She should interpret this finding to mean
which suggests a (weak/moderate/strong) effect.
2. Another researcher compares the amount of time a sample of high school juniors spent on
their phones during a school day to the average time found in a national survey of high school
students. He finds d = .35.
a. The name of this test is .
b. He should interpret this result to mean which
suggests a (weak/moderate/strong) effect.
3. a. Compute the effect size using Cohen’s d for Application question 6 from Exercise 8.2
(about cell phone use while driving).

b. Interpret the effect size.

______________________________________________________________________
______________________________________________________________________

Exercise 8.4: Calculating a Confidence Interval

Learning Objective: Compute the confidence interval when comparing a sample to a known population
value.
68   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application
1. The researcher who assessed the time high school juniors spent on their phones (measured
in minutes) during a school day finds (M – µ) = 4.50, 95% CI [3.30, 5.70]. How would you
interpret this finding?


2. Does a researcher prefer a small or large confidence interval? Why?


3. a.  Compute the confidence interval for Application question 6 from Exercise 8.2 (about cell
phone use while driving).

b. Interpret the confidence interval.

______________________________________________________________________
______________________________________________________________________

Exercise 8.5: The Big Picture: Examining One Variable at a Time

Learning Objective: Formulate a big picture of how to examine one variable at a time.
Review
Complete the following table.

A Results section should include

1.

2. Type of statistical test(s)

3.

4.

A Discussion section should include

1.

2. Interpretation of findings, including how they fit with previous research

3.

4. Possible limitations of the study

5.

Application
1. As part of a complex study of substance users in Australia, the participants’ scores
(M = 53.7, SD = 19.6) on a subjective well being scale were compared with the national
norm (µ = 75.97) for the scale. The analysis found a significant difference, t(131) = 13.1,
 Chapter 8 ■ Comparing Your Sample to a Known or Expected Score   69

p < .001 (Tait, 2018). Using the statistics provided in the report, d = 0.57 and 95% CI
[ –25.65, –18.89].
a. Is this an appropriate use of a one-sample t test? Why or why not?
______________________________________________________________________
______________________________________________________________________
b. Interpret the findings of Tait’s (2018) study.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
2. Retrieve the following results from the study for the drivers using cell phones
a. t test and df (formatted in APA style)
(Exercise 8.2, Application question 6c., d.)
b. Cohen’s d (Exercise 8.3, Application
question 3a.)
c. confidence interval (Exercise 8.4,
Application question 3a.)
d. Write a Results section for your findings. Use APA format and include the descriptive
statistics, type of statistical test and results of the test, effect size, and confidence interval.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
e. Write a Discussion section for your findings. Include the findings, interpretation/
explanation/implication of the findings, limitations of the study, and possible next studies.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
3. The Pew Research Center sampled a large national group on basic science facts using a
12-item quiz. The average correct score on the quiz was 7.9. You decide to see whether
a sample of 16 students in a study group at a local bookstore is more knowledgeable than
the general adult public. The output from a one-sample t test comparing the students to
the national average is below.

One-Sample Statistics

N Mean Std. Deviation Std. Error Mean

Science Facts 16 8.9375 1.73085 .43271

70   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

One-Sample Test

Test Value = 7.9

95% Confidence Interval of the

Difference
Sig. Mean
t df (2-tailed) Difference Lower Upper

Science Facts 2.398 15 .030 1.03750 .1152 1.9598

One-Sample Effect Sizes

95% Confidence Interval

Standardizer a Point Estimate Lower Upper

Cohen’s d 1.73085 .599 .057 1.125

Science Facts Hedges’ correction 1.82385 .569 .054 1.068

a. The denominator used in estimating the effect sizes.

Cohen’s d uses the sample standard deviation.
Hedges’ correction uses the sample standard deviation, plus a correction factor.

a. Write a Results section following APA format. Be sure to check with your instructor for
any specific guidelines that apply for your class.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
b. Write a Discussion section following APA format.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
 9
EXAMINING RELATIONSHIPS
AMONG YOUR VARIABLES
Correlational Design

LEARNING OBJECTIVES

• Describe the advantages and limits of correlational designs.

• Compute and interpret the statistics assessing correlations between interval and ratio variables.
• Demonstrate how to use a relationship to predict scores of one of the variables in the relationship
(regression).
• Formulate the big picture to differentiate correlational design and correlation as a statistic.

CHAPTER SUMMARY
The chapter begins with a discussion of correlation research, which investigates the relationship between variables.
Correlation designs are employed for many reasons, such as when the manipulation of variables is impossible or unethical,
in a pilot study to check whether a relationship exists before testing for causation in an experiment, to test validity or
reliability of measures, or to increase external validity. The major drawback of correlation designs is that they do not meet
the requirements for causation. The ethical interpretation of correlation research and analysis is reviewed throughout the
chapter in terms of focusing on relationships and not implying causation, even when a relationship is strong.

The chapter describes types of correlation statistics that are used to analyze the relationship between two variables. These
statistics may be the primary or secondary analysis in a study, and, thus, correlation statistics are used in experimental as
well as correlational designs. The type of correlation statistic that is appropriate is dependent on the scale of measurement
of the two variables being analyzed. The most commonly used correlation is Pearson’s r, which examines the relationship
between two interval or ratio variables. Other correlation statistics include the point-biserial correlation (rpb) between a
dichotomous and interval/ratio variable (described in this chapter), the chi-square test for independence (χ2) between two
nominal variables, and the Spearman rho (rs) between two ordinal variables (both described in Chapter 14).

When there is a statistically significant Pearson’s correlation, a researcher may choose to compute a regression equation
or the equation for the line of best fit for the data (Y ′ = bX + a). This equation can then be used to predict one of the
variables in the relationship from the other. In addition, multiple regression, an analysis that predicts a variable based on
its relationship to two or more other variables, is briefly described.

71
72   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

CHAPTER 9 QUIZ
Test your knowledge of key terms from the chapter.
1. A correlational design .
a. describes the relationship between two variables
b. is a type of study that tests the hypothesis that variables are related
c. determines causation between variables
d. examines causal relationships between three or more variables
2. Correlational designs can have greater external validity than experiments because
correlational designs .
a. better represent the everyday world where we do not manipulate or
control events
b. study true relationships while experiments only manipulate relationships
c. are more attractive to participants
d. are easier to conduct
3. We can increase the power of a correlational study by
a. using measures with floor or ceiling effects.
b. using measures that are valid and reliable.
c. restricting the range of scores for each measure.
d. only using a very small random sample.
4. A correlation coefficient provides these two pieces of information:
a. strength and direction of a relationship.
b. direction and cause of a relationship.
c. strength and prediction of each variable.
d. direction and effect size of a relationship.
5. The value of Pearson’s r ranges between .
a. 0.0 and +1.0
b. –1.0 and 0.0
c. –1.0 and +1.0
d. 0.0 and 10.0
6. The point-biserial correlation coefficient (r pb) is used to analyze the
a. effect of a nominal variable on an interval/ratio variable.
b. effect of an interval/ratio variable on a nominal variable.
c. correlation between variables measured on interval and ratio scales.
d. correlation between a dichotomous and an interval/ratio variable.
7. You might compute a linear regression equation when you .
a. compute a significant correlation between two variables.
b. find a significant correlation for two interval/ratio variables.
c. find a significant rpb.
d. have one dichotomous and one interval/ratio measure and a significant rpb.
 Chapter 9 ■ Examining Relationships Among Your Variables    73

8. The regression equation is always represented by the formula .

a. X' = Yb + X
b. X' = bX + Y
c. Y' = a – bX
d. Y' = bX + a
9. The average error between predicted Y values and the actual Y values is called
a. line of best fit.
b. standard error of the correlation.
c. standard error of the estimate.
d. estimated standard error of the estimate.
10. The statistic that describes the proportion of variability that is accounted for by knowing the
relationship between two variables is the
a. coefficient of prediction.
b. standard error of the estimate.
c. estimated standard error of the means.
d. coefficient of determination.

Quiz 9 Answer Key

1. b. 2. a. 3. b. 4. a. 5. c. 6. d. 7. b. 8. d. 9. c. 10. d.

CHAPTER 9 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying the key concepts.

Exercise 9.1: Correlation Design

Learning Objective: Describe the advantages and limits of correlational designs.

Review
1. Fill in the chart below.
2. What is the major disadvantage of correlation designs?
_________________________________________________________________________

Reasons we use a correlation design:

74   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. A valid correlation depends on obtaining a variation in the scores for each measure.
a. When the obtained scores all fall at the upper end of the possible scores, the measure may
have a effect.
b. When the obtained scores all fall at the lower end of the possible scores, the measure may
reflect a effect.
Application
1. Why are the following studies appropriate for a correlation design?

Is students’ anxiety is related to test scores in

Research Methods?

Is inadequate nutrition during the first year of

life is related to children’s language abilities
at 3 years of age?

Do visits to a mental health clinic increase

after a town experiences a severe hurricane?

Are people with higher incomes more likely to

own their homes (yes/no)?

2. A student researcher wants to examine risk-taking behaviors among emerging adults in

comparison to their health status. She wants to use a correlation design, but her classmate
argues for an experimental design. Which design would you recommend and why?
_________________________________________________________________________
_________________________________________________________________________
3. A student researcher finds that attractiveness and confidence are positively correlated. He tells
a classmate that his study shows that “high levels of attractiveness cause people to have greater
confidence.” How would you respond to his claim?
_________________________________________________________________________
_________________________________________________________________________
4. A researcher is exploring the relationship between intelligence and work performance, with
both variables rated by the supervisor on a 10-point scale. Does this study meet the criteria for
a powerful correlational design? Explain you answer.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
5. Suppose you find for the work performance ratings in the example above that the ratings
of all the employees range from 1 to 3, although it is possible to score from 1 to 10 on the
measure. Would you be able to find a valid correlation with this measure? Explain.
_________________________________________________________________________
_________________________________________________________________________

Exercise 9.2: Relationship Between Two Interval or Ratio Variables

Learning Objective: Compute and interpret the statistics assessing correlations between interval and ratio
variables.
Review
Match the terms in the first column with the most appropriate term in the second column.
 Chapter 9 ■ Examining Relationships Among Your Variables    75

when the scores of one variable increase as the scores of Pearson’s correlation coefficient
another variable decrease
when the scores of two variables either increase together or perfect correlation
decrease together
r=0 no correlation
r = –.76 positive correlation
r = 1.0 negative correlation
graph that depicts the participant scores on two measures strong correlation
statistic assessing the relationship between 2 interval or scattergram
ratio variables

Application
1. The more alcohol consumed, the poorer a person’s scores on a driving skills simulation. This is
an example of a (positive/negative) correlation.
2. As the number of likes for a news story on Facebook increases, the time spent viewing the
story also increases. This is an example of a (positive/negative)
correlation.
3. Suppose you find that the correlation between hours of study and scores on a statistics exam
is .34, while the correlation between math anxiety and the statistics exam is –.63. Which
relationship is stronger? Explain your answer.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
4. The correlation between hours of study and scores on a statistics exam (r = .34) would
be interpreted as a (weak, moderate, strong) correlation, while the
correlation between math anxiety and statistics exam (r = –.63) would be interpreted as a
(weak/moderate/strong) correlation.
5. Examine the graph below.
a. How would you interpret the graph?
______________________________________________________________________
b. What statistic(s) would you compute to check that your interpretation of the contents of
the graph was correct?

150000.00

125000.00

100000.00
Salary

75000.00

50000.00

25000.00

10.00 15.00 20.00 25.00 30.00

Education
76   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Correlation Between Life Satisfaction and Total Reminiscing

60.00

Total Reminiscing
40.00

20.00

.00
5.00 10.00 15.00 20.00 25.00 30.00 35.00
Total of satisfaction 1, satisfaction 2,
satisfaction 3, satisfaction 4, and satisfaction 5

6. How would you interpret the graph above?

_________________________________________________________________________
_________________________________________________________________________

Application
1. A researcher finds for a group of 30 residents in an assisted living facility that general health status
(assessed on a 20-point scale) and minutes spent exercising per week are correlated, r = .32.
a. State a null hypothesis for the study.
______________________________________________________________________
b. State a directional alternative hypothesis for the study.
______________________________________________________________________
c. What is the df for the study? (Hint: for a Pearson’s r, df = N – 2.)
d. Is the relationship statistically significant for a two-tailed test at p < .05? (Hint: see Table
C.5 in Appendix C of the textbook.) Explain your answer.
______________________________________________________________________
______________________________________________________________________
2. A developmental psychologist observes seven children with their parents, rating levels of
family support and child behavior problems. The data appear below. (Note that this is a very
small sample size for demonstration purposes. If you were conducting an actual study, you
would want a larger sample size.)

Child ID number Ratings of Family Support Ratings of Child Behavior Problems

1 8 4

2 5 4

3 2 7

4 9 3

5 7 4

6 5 7

7 3 5
 Chapter 9 ■ Examining Relationships Among Your Variables    77

a. State a null hypothesis for the study.

______________________________________________________________________
b. State a directional alternative hypothesis for the study.
______________________________________________________________________
______________________________________________________________________
c. Draw a scattergram of the data.

d. Using the formula in Chapter 9 of the textbook, calculate Pearson’s r. (Hint: Add
columns to the table for the needed parts of the formula.)

e. Calculate the df
f. Using Table C.5 in Appendix C of the textbook, what is the critical r for this study for a
two-tailed test at p < .05?
g. Is the relationship statistically significant? In other words, is the r you calculated at or
stronger than the critical r?

Using Data Analysis Programs: Pearson’s r

Application
The data presented below are for demonstration purposes only, and some are based on very small sample
sizes. For these examples, it is likely that a larger N would be needed so that we would have sufficient power.
1. Examine the output below from a survey that students completed regarding the use of
different types of humor in the classroom:
CORRELATIONS
Related Humor Disparaging Humor

Related Humor Pearson Correlation 1 .154*

Sig. (2-tailed) .021
N 232 227
Disparaging Humor Pearson Correlation .154* 1
Sig. (2-tailed) .021
N 227 233

. Correlation is significant at the 0.05 level (2-tailed).

*
78   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

a. What correlation statistic was computed?

b. Name the two variables being correlated.
i. _________________________________________________________
ii. _________________________________________________________
c. Is the relationship statistically significant? How do you know?
______________________________________________________________________
d. Calculate the proportion of variance accounted for: r 2 = .
Interpret this finding.
______________________________________________________________________
e. Describe these findings as you might in a Results section.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
f. Interpret the meaning of the findings.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
2. Examine the output below:

CORRELATIONS
Openness to New
Gender Stereotypes Experiences

Gender Stereotypes Pearson Correlation 1 –.789**

Sig. (2-tailed) .000

N 21 21

Openness to New Pearson Correlation –.789** 1

Experiences
Sig. (2-tailed) .000

N 21 21

. Correlation is significant at the 0.01 level (2-tailed).

**

a. How many participants were in this study?

b. Is the relationship statistically significant? How do you know?

______________________________________________________________________

______________________________________________________________________

c. Calculate the proportion of variance accounted for: .

Interpret this finding.

______________________________________________________________________
 Chapter 9 ■ Examining Relationships Among Your Variables    79

d. The probability of a Type I error = and

a Type II error = .

e. Describe these findings as you might in a Results section.

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

Formulas and Calculations: Point-Biserial r

Application

1. The relationship between living status (live alone or lives with others) and life satisfaction
(rating scale) can be analyzed using the . The statistical notation for this
correlation statistic is .

2. A researcher is interested in the relationship between owning a home (yes/no) and educational
attainment (years of education completed). After surveying 100 individuals at a local farmer’s
market, he finds rpb = .23.

a. What is the df for this study?

b. Is the relationship significant at p < .05 for a two-tailed test? Explain.

______________________________________________________________________

______________________________________________________________________

c. Is the relationship significant at the p < .01 for a two-tailed test? ________ Explain your
answer.

______________________________________________________________________

______________________________________________________________________

Using Data Analysis Programs: Point-Biserial r

1. A gerontologist computes the correlation between pet ownership (yes or no) and happiness
among a group of elderly participants (80 years or older) who live in their own homes. The
output is shown below when Owning a pet = 1 and Not owning a pet = 2:

CORRELATIONS
Pet Owner Happiness

Pet Owner Pearson Correlation 1 –.416*

Sig. (2-tailed) .022

N 30 30

Happiness Pearson Correlation –.416* 1

Sig. (2-tailed) .022

N 30 30

. Correlation is significant at the 0.05 level (2-tailed).

*
80   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

a. Even though the output above says this is a Pearson’s r, what correlation statistic was
computed?
b. Is the relationship statistically significant? How do you know?

c. The probability of a Type I error = and a Type II error = .

d. Describe these findings as you might in a Results section.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
e. What recommendation would you make to those working with the elderly, based on these results?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
f. What variable(s) (other than the two in this study) could be responsible for the
relationship?
______________________________________________________________________
______________________________________________________________________
2. Below is an example output with the interval variables frustration tolerance and self-control,
the ratio variable income, and the nominal variable health status (healthy or unhealthy). Notice
also that each of the correlations is listed twice, so you need to pay attention only to the top
or bottom half of the matrix.

CORRELATIONS
Frustration Annual Health
Tolerance Self-Control Income Status
Frustration Pearson Correlation 1 .704* .673* –.436
Tolerance
Sig. (2-tailed) .023 .033 .208
N 10 10 10 10
Self-Control Pearson Correlation .704* 1 .842** –.748*
Sig. (2-tailed) .023 .002 .013
N 10 10 10 10
Annual Income Pearson Correlation .673* .842** 1 –.529
Sig. (2-tailed) .033 .002 .116
N 10 10 10 10
Health Status Pearson Correlation –.436 –.748* –.529 1
Sig. (2-tailed) .208 .013 .116
N 10 10 10 20
 Chapter 9 ■ Examining Relationships Among Your Variables    81

a. Fill in the blanks.

  T
 he correlation between frustration tolerance and self-control: r = __________,
p = _____________
  T
 he correlation between frustration tolerance and income: r = ____________ ,
p = ____________
  T
 he correlation between frustration tolerance and health status: rpb = __________,
p = _____________
  The correlation between self-control and income: r = ___________, p = __________
  The correlation between self-control and health status: rpb = _______, p = _____________
  The correlation between income and health status: rpb = _________, p = __________
b. Circle the correlations in the matrix that are statistically significant.
c. Indicate the strength of the relationship (weak/moderate/strong) next to each.

Exercise 9.3: Regression

Learning Objective: Demonstrate how to use a relationship to predict scores of one of the variables in the
relationship (regression).

Review
1. You could use the regression equation to predict one variable based on another if
.
2. The variable is the predicted variable in a regression.
3. The is the value used to predict another variable.
4. Sometimes we are interested in using a relationship between three or more variables to predict
a variable and, in this case, we compute a , which is symbolized as
.

Formulas and Calculations: Simple Linear Regression

Review
1. The value that results when you enter a particular X value in a regression equation is called
and is symbolized by .
2. Each X value and its predicted Y value falls on the .
3. The regression equation is always in the format of Y' = bX + a, with b = and
a= .
4. The direction of the slope (positive or negative) is determined by the .
5. The average difference between predicted Y values (Y’) and the actual Y values in the data is
called the and is symbolized by .
6. A small standard error of estimate suggests that .
7. is the proportion of variability accounted for by knowing the
relationship between two variables and is symbolized by .
8. The larger r 2, the (more/less) accurately we can predict Y from X.
82   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Using Data Analysis Programs: Simple Linear Regression

Application
A researcher asked 35 participants to rate the importance of environmental issues and how much time
they spend outside. The output is shown below:

Model Summary
Model R R Square Adjusted R Square Std. Error of the Estimate
1 .381 a
.145 .119 5.47401

a. Predictors: (Constant), I think the environment is the most important political issue today.

ANOVAa
Model Sum of Squares df Mean Square F Sig.
1 Regression 167.847 1 167.847 5.601 .024b
Residual 988.839 33 29.965
Total 1156.686 34

a. Dependent Variable: On average, how many hours do you spend outside each week?
b. Predictors: (Constant), I think the environment is the most important political issue today.

Coefficientsa
Unstandardized Standardized
Coefficients Coefficients 95.0% Confidence Interval for B
Std.
Model B Error Beta t Sig. Lower Bound Upper Bound
1 (Constant) -3.008 4.518 -.666 .510 -12.199 6.183
I think the 2.818 1.190 .381 2.367 .024 .396 5.240
environment is the
most important
political issue today.

a. Dependent Variable: On average, how many hours do you spend outside each week?

On average, how many hours do you spend

outside each week?

20.00
Observed
Linear

15.00

10.00

5.00

.00
3.00 3.50 4.00 4.50 5.00
I think the environment is the most important
political issue today.
 Chapter 9 ■ Examining Relationships Among Your Variables    83

1. Is the relationship statistically significant? . How do you

know?

2. The effect size is . (name the exact value). You interpret this
as meaning . The effect size is .
(strong/moderate/weak).

3. The probability of a Type I error = and a Type II

error = .

4. is the predictor variable and

is the criterion variable.

5. Write the regression equation for the relationship.

6. a.  W hat is the average deviation between actual hours spent outside and the hours predicted
by the regression equation?

b. If a person rates as 4 (mostly agree) that the environment is the most important political
issue, how many hours do you predict he or she spends outside, according to the
regression equation?
7. Describe these findings as you might in a Results section, using APA format.

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________
8. Interpret the findings as you would in a Discussion section.

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________

_________________________________________________________________________
84   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Exercise 9.4: The Big Picture: Correlational Design Versus Correlational

Analysis
Learning Objective: Formulate the big picture to differentiate correlational design and correlation as a
statistic.

1. Complete the following chart:

Correlational Statistics

Scale of Measurement Correlational Analysis Appropriate for What Design(s)?

Two interval/ratio variables

Point-biserial correlation
coefficient

Two nominal variables with Correlational or experimental

two or more categories designs

Spearman rho (rs) Correlational design

2. Name the most appropriate inferential statistic and its statistical symbol for the following
questions:
a. Is a measure of hardiness (scores can range from 20 to 100) reliable? High school students
complete a scale assessing the trait of hardiness and then complete the same scale again
one week later.
______________________________________________________________________
b. Are visits to a mental health clinic (yes/no) related to family stress as assessed by a scale
ranging from 0 to 20?
______________________________________________________________________
c. A school counselor and teacher rank students in the class in terms of frequency of
misbehavior. Are the counselor and teacher rankings correlated?
______________________________________________________________________
d. Is SES (high/middle/low) related to political views (conservative/moderate/liberal)?

e. Is suicide ideation (scores range from 0 to 84) related to work-related traumatic stress
(scores range from 0 to 10) among police officers?
______________________________________________________________________
f. If compulsivity and attention to detail are strongly correlated, what is the score for
attention to detail if compulsivity = 25?
______________________________________________________________________
g. Is participation in after-school activities in high school (yes/no) related to self-reported
acts of delinquency (yes/no)?
______________________________________________________________________
h. What is the line of best fit for the relationship between years on a job and salary?
______________________________________________________________________
 Chapter 9 ■ Examining Relationships Among Your Variables    85

Your Research
Find and Evaluate a Nonexperimental Study on Your Chapters 8 and 9). List them and explain how the authors
Topic used the statistic.

1. Use your library database to find a primary research
article that describes a nonexperimental study on your 
topic (or related to your topic).

2. After you have found an appropriate nonexperimental

study, read the entire article carefully.

3. Look through the entire Results section to identify any
inferential statistics with which you are familiar (from 
 10
EXAMINING CAUSALITY

LEARNING OBJECTIVES

• Describe the key components necessary to test cause and effect.

• Identify the eight key threats to internal validity.
• Illustrate basic issues in designing an experiment.
• Evaluate the validity of experiments.
• Formulate a big picture of the benefits and limits of experiments.

CHAPTER SUMMARY
How do you determine that one variable caused a change in another variable? Chapter 10 begins by outlining the requirements
for causality, including sequence, correlation, and the ability to rule out alternative explanations. Internal validity is the extent
to which a research design ruled out alternative explanations, and the key threats to internal validity are discussed in detail in
this chapter. These threats are inherent in the nonexperimental designs, such as a pre-/posttest or quasi-experimental designs.

An experiment is the only type of design that can demonstrate causality, although simply conducting a study is not
enough to show causality. This chapter discusses basic issues in designing an experiment including practical and ethical
issues in recruiting participants, random assignment, maintaining experimental control, manipulating the independent
variable, and measuring the dependent variable. The chapter focuses on increasing internal validity of experiments while
acknowledging the importance of balancing internal and external validity. The chapter ends with a discussion of the
strengths and limitations of experimental designs.

CHAPTER 10 QUIZ
Test your knowledge of the key terms from the chapter.
1. is the degree to which one can say that one variable caused a change in another.
a. Power
b. Experiment
c. Internal validity
d. Confound

86
 Chapter 10 ■ Examining Causality   87

2. The main difference between a quasi-experiment and an experiment is that a quasi-

experiment does not include
a. manipulation of an independent variable.
b. examination of mediating variables.
c. operational definitions.
d. random assignment to condition.
3. The variable is predicted to affect the variable.
a. mediating; control
b. control; confounding
c. dependent; independent
d. independent; dependent
4. A(n) is someone who pretends to be a participant or bystander but is actually
working with the researcher.
a. confederate
b. control
c. experimenter
d. ally
5. A receives an intervention or treatment only after the completion of the study.
a. confederate
b. confound
c. posttest group
d. wait-list control
6. Which of the following is NOT a requirement for an experiment?
a. random selection of the sample
b. random assignment to condition
c. manipulation of an independent variable
d. measurement of a dependent variable
7. A researcher wants to use simple random assignment in an experiment with two conditions.
Which of the following procedures might he use?
a. assigning every other participant to condition 1
b. rolling dice and assigning those with even numbers to condition 1 and those with odd
numbers to condition 2
c. having all the participants with blonde hair in condition 1 and those with other hair
colors in condition 2
d. match participants on a key variable, then flip a coin to determine which member of the
pair is assigned to condition 1, and which is assigned to condition 2
8. The phenomenon in which a researcher unintentionally treats the conditions differently so
that results support the hypothesis is called
a. the Hawthorne Effect.
b. demand characteristics.
c. experimenter expectancy effects.
d. diffusion of treatment.
88   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

9. Experiments focus primarily on maximizing

a. external validity.
b. internal validity.
c. ecological validity.
d. confound validity.
10. A type of external validity that focuses on real-world application is called
a. predictive validity.
b. divergent validity.
c. ecological validity.
d. face validity.

Quiz 10 Answer Key

1. c. 2. d. 3. d. 4. a. 5. d. 6. a. 7. b. 8. c. 9. b. 10. c.

CHAPTER 10 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 10.1: Testing Cause and Effect

Learning Objective: Describe the components necessary to test cause and effect.
Review

1. What are some terms that researchers use to suggest that one variable caused a change in
another variable?


2. Students sometimes confuse the words “effect” and “affect.” Remember that when discussing
causality, effect is a noun and affect is a verb. Practice by filling in the blanks below with the
word “effect” or “affect”
a. Researchers have examined the of smell on memory.
b. The color of a car might the likelihood of getting into an accident.
c. We hypothesize that spending time in nature will mood.
d. Do you think that taking practice quizzes will have an on your exam score?
3. List the three requirements for causality:


4. What condition(s) for causality are lacking in a correlational design?


 Chapter 10 ■ Examining Causality   89

Application
1. The local animal shelter intends to call people who recently adopted a pet to see if having
a pet has reduced their loneliness. Will they be able to show a causal relationship? Why or
why not?



2. Your friend crammed for an exam but did very well, and then concludes that cramming for
exams is more effective than spacing studying over time. Identify at least one alternative
explanation for causality.



Exercise 10.2: Eight Key Threats to Internal Validity

Learning Objective: Identify eight key threats to internal validity
Review
Fill in the blanks:

Threat to internal validity due to… Name of threat to internal validity

a statistical phenomenon in which extreme scores

become less extreme

occurs when participants withdraw from the study

preexisting differences in the sample

preexisting differences that interact with other

threats

inconsistency in measures, administration, or

scoring.

any event that might impact the outcome of a

study

a pretest

time

Application
1. A political candidate’s approval rating was very low, a mere 20%, prior to unveiling his
economic plan. After the plan was shared, the candidate’s approval rating rose to 28%.
a. What type of design is this? Be specific. ___________________________________
90   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

b. Identify the threats to internal validity that help explain why this design
cannot demonstrate that the economic plan is the reason for the increased
approval ratings:

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

2. A marketing executive wants to know whether a new website might improve sales. The
executive plans to compare the new website to the original one.

a. Identify two ways that the marketing director can minimize the testing threat to internal
validity.

______________________________________________________________________

______________________________________________________________________

b. Why would it be a problem if the marketing executive gave one group up to 30 minutes
to explore the new website, but gave the other group only 10 minutes to explore the old
website? Identify the specific threat to internal validity.

______________________________________________________________________

______________________________________________________________________

c. Why would it be a problem if the marketing executive had one group look at the new
website in pairs and the other group look at the old website individually? Identify the
specific threat to internal validity.

______________________________________________________________________

______________________________________________________________________

d. Why would it be a problem if the marketing executive had all the people who were under
30 explore the new website and had all those who were 30 and older explore the old
website? Identify the specific threat to internal validity.

______________________________________________________________________

______________________________________________________________________

e. Why would it be a problem if the marketing executive asked those in one group to report
the items they would like to purchase and looked at the online shopping carts of those in
the other group? Identify the specific threat to internal validity.

______________________________________________________________________

______________________________________________________________________

f. When would attrition be a threat to internal validity of the marketing executive’s

study?

______________________________________________________________________

______________________________________________________________________
 Chapter 10 ■ Examining Causality   91

Exercise 10.3: Basic Issues in Designing an Experiment

Learning Objective: Illustrate basic issues in designing an experiment.
Application
1. A police department wants to know whether citizens will better comprehend verbal police
instructions if an officer repeats the information twice or if an officer adds listenability
features, such as an introduction (“I am going to tell you three things”) and lists (first . . .
second . . . third). Help them design a simple experiment:

Recruit participants

Prescreen for:

Inform potential participants:

Randomly assign participants to independent variable (IV) conditions

What type of random assignment?

How will you carry out this random assignment?

IV condition (or level) 1: IV condition (or level) 2:

Keep everything else constant (the same) across the two conditions

What will be constant across your IV conditions?

Measure Dependent Variable (DV)

What is the DV? How might you operationally

define (measure) the DV?

2. A researcher plans to conduct an independent-groups experiment (also called a between-

subjects experiment) to examine the effect of drinking 64 ounces of water a day on weight loss.
a. Explain how the researcher should use prescreening and informed consent to ensure an
ethical study.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
92   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

b. What type of sampling would you recommend? Explain.

______________________________________________________________________
______________________________________________________________________
c. Explain how the researcher might use stratified random assignment to make sure that a
specific characteristic is distributed across the conditions (identify a characteristic that is
important and outline the process of stratified random assignment).
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
d. Recommend conditions (or levels) of the independent variable. Explain whether or not
you recommended a control group.
______________________________________________________________________
______________________________________________________________________
e. What manipulation check would you recommend?
______________________________________________________________________
______________________________________________________________________
f. How might the researcher operationally define the DV?
______________________________________________________________________
______________________________________________________________________
3. If the researcher from question 2 does a dependent-groups experiment (also called a within-
subjects experiment), how would the procedures be different?





Exercise 10.4: Validity in Experiments

Learning Objective: Evaluate the validity of experiments.
Application
You plan to conduct an experiment to examine whether protesters are more effective in gaining approval
for a cause if their signs use optimistic versus pessimistic language. The experiment will, of course,
include random assignment, manipulation of the independent variable (IV), and measurement of a
dependent variable (DV).
1. The IV is
The IV conditions (or levels) are: and
2. The DV is
3. How would you improve the internal validity of your study? In particular:
a. What will you keep constant across your IV conditions?
______________________________________________________________________
______________________________________________________________________
 Chapter 10 ■ Examining Causality   93

b. How might you avoid differential attrition?

______________________________________________________________________
______________________________________________________________________
c. What might you do to reduce demand characteristics?
______________________________________________________________________
______________________________________________________________________
d. What might you do to reduce experimenter expectancy effects?
______________________________________________________________________
______________________________________________________________________
e. What might you do to avoid diffusion of treatment?
______________________________________________________________________
______________________________________________________________________
4. Comment on the external validity of your experiment. In particular, focus on ecological
validity.
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________

Exercise 10.5: The Big Picture: Benefits and Limits of Experimental

Design
Learning Objective: Formulate a big picture of the benefits and limits of experiments.
Application
1. What are the benefits of conducting the experiment you described in exercise 10.4 on
optimistic versus pessimistic language in protests?



2. Briefly outline a quasi-experiment on optimistic versus pessimistic language in protests and
explain the advantages of this design compared to an experiment.




3. Briefly outline a correlational design on the topic of optimistic and pessimistic language in
protests and explain the advantages of this design compared to an experiment.




94   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

4. How would you determine which of the designs (experimental, quasi-experimental,

correlational) is most appropriate?



Your Research
Find and Evaluate an Experiment on Your Topic 5. Was there a manipulation check? If so, describe.

Use your library database to find a primary research article that
describes an experiment on your topic (or related to your topic). 6. Identify at least one way, other than random assignment,
Limit your search to experiments published in the last 10 years. that the researchers controlled threats to internal
See Chapter 2 for search strategies. validity.

1. How do you know the study is an experiment?
7. Identify the DV(s). How did the researchers operationally

define the DV(s)?


2. How did the researchers recruit participants (or obtain 
animal subjects)?
8. What, if any, threats are there to the internal validity of

the study? (Hint: review the Discussion section.)


3. What type of random assignment did they use? 
 

4. Identify the IV(s). How did the researchers manipulate the 9. How might you build on this study? Include ways that you
IV(s)? might address any threats you identified above.
 
 
 11
INDEPENDENT-GROUPS DESIGNS

LEARNING OBJECTIVES

• Describe three types of independent-groups designs with two levels.

• Design a simple experiment.
• Analyze and interpret a two-group design using an independent-samples t test, effect size, and
confidence interval.
• Discuss the characteristics and advantages of multiple-groups designs.
• Analyze and interpret data for multiple-groups designs using a one-way ANOVA and post hoc tests.
• Formulate the big picture for identifying and analyzing independent-groups designs.

CHAPTER SUMMARY
In this chapter, you learned about independent-groups designs in which the participants in different groups are unrelated.
These designs can be correlational, quasi-experimental, or experimental studies. Although the statistical tests used
to analyze the three types of independent-groups designs are the same, the interpretation of the results differs. Both
correlational and quasi-experimental results are described as a correlational relationship between the predictor (IV) and
outcome (DV), while experiments are described as a causal relationship between the IV and DV. The chapter reviewed
the controls and considerations you learned about in previous chapters regarding designing powerful independent-groups
studies.
The simplest independent-groups design involves two groups, and you learned how to analyze the data from such studies
using an independent-samples t test. This statistical test assumes that you have two unrelated groups, an IV (predictor)
and DV (outcome), interval/ratio data, normally distributed DV, and similar variability of scores in the two groups
(homogeneity of variance). The independent-samples t test compares the means of the two groups in relation to an
estimate of the standard deviation of the sampling distribution (standard error of the difference between the means). You
also compute an effect size (Cohen’s d or rpb2) and confidence intervals, either by hand or by using a statistical package
(e.g., SPSS, Excel).
You then learned about multiple-groups designs when you have one IV (predictor) with three or more independent
groups. The advantages of multiple-groups designs (in comparison to two-group designs) were described. The findings
of these designs are analyzed using a one-way analysis of variance (ANOVA), which compares the variability between
groups (treatment variance) to the variability within groups (error variance). The assumptions are similar to those for the

95
96   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

independent-samples t test. If you find a significant F (the end statistic for the ANOVA), you compute
a post hoc test, which examines differences between all possible pairs of the groups in the study. Eta
squared (η2) is the effect size appropriate for the ANOVA, and it is interpreted the same way as rpb2. In
reporting your results for a t test or one-way ANOVA, you should make clear the type of relationship
you have examined (correlational or causal) and discuss the practical and statistical interpretation of
your findings.

CHAPTER 11 QUIZ
Test your knowledge of the key terms from the chapter.
1. Correlational designs include a(n) instead of an IV and a(n)
instead of a DV.
a. predictor; estimated value
b. predictor; outcome
c. outcome; predictor
d. estimated effect; estimated value
2. The interpretation of the results for quasi-experimental designs differs from that of an
experiment because
a. quasi-experiments examine relationships while experiments examine causation.
b. quasi-experiments examine causation between two variables while experiments examine
causation between three or more variables.
c. experiments involve the manipulation of a variable while quasi-experiments do not.
d. experiments manipulate an IV while quasi-experiments manipulate a DV.
3. In a simple experiment, the IV is on a(n) scale of measurement and the DV is
on a(n) scale.
a. nominal; ordinal or ratio
b. ordinal; ratio or interval
c. interval or ratio; nominal
d. nominal; interval or ratio
4. In a simple experiment examining the impact of violent games on aggressive thoughts, we
might have participants play a violent game for 45 min rather than 5 min. This decision
reflects our attempt to maximize power by
a. controlling conditions in our experiment.
b. ensuring that our IV is valid and reliable.
c. using a strong manipulation of our IV.
d. using similar participants in our study.
5. To determine where our mean difference falls on the sampling distribution that predicts no
difference we need to know the standard deviation of the sampling distribution. This term is
called .
a. standard error of the means.
b. standard error of the estimate.
c. standard error of the difference between means.
d. standard deviation of the samples.
 Chapter 11 ■ Independent-Groups Designs   97

6. The assumption that variances in populations are the same is called

a. homogeneity of participant populations.
b. homogeneity of variance.
c. similarity of effects.
d. group homogeneity.
7. A multiple independent-groups design examines the effect of
a. more than one IV on a DV.
b. one IV on three or more DVs.
c. one IV with three or more levels on a DV.
d. two or more IVs on two or more DVs.
8. The one-way ANOVA compares
a. within-group variance for each group to within-variance for the total group.
b. error variance for each group to error variance for the total group.
c. between-groups variance and within-groups variance.
d. error and within-groups variance.
9. The mean square within groups represents the
a. average deviation of each participant’s score from the population mean.
b. average deviation of group means from the population mean.
c. average deviation of group means from the total mean of the sample.
d. average deviation within all groups of the study.
10. The effect size for an ANOVA is computed using
a. eta squared (η2) .
b. Cohen’s d.
c. Levene’s test.
d. point-biserial squared.

Quiz 11 Answer Key

1. b. 2. a. 3. d. 4. c. 5. c. 6. b. 7. c. 8. c. 9. d. 10. a.

CHAPTER 11 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 11.1: Designs With Independent Groups

Learning Objective: Describe three types of independent-groups designs with two levels.
Review

1. The three designs with independent groups are ,

, and .
98   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. The independent-groups designs discussed in the chapter compare groups analyzed on

a(n) or a(n) scale of measurement.
3. List the type of design(s) [correlational, quasi-experiments, experiments] that fit the following
characteristics:
Examines one IV with 2 levels and one DV:
Manipulates an IV:
Measures the DV using an interval or ratio scale:
Randomly assigns participants to groups:
Examines relationships:
Examines causation:

Application
1. A study finds that children who read at home every day are more likely to enjoy school than
children who never read at home. This is a/an (type of design) independent-
groups design; reading at home is the variable and enjoying school is the
variable.
2. A researcher interested in environmental issues completes a study on recycling. She arranges
to have one neighborhood’s recycling picked up each week and a second neighborhood’s
recycling picked up each month. The amount of recycling (in pounds of material) for each
resident in the two neighborhoods is measured over 6 months.
a. What type of study is this?
b. What type of variable does timing of recycling represent?
c. What type of variable does amount of recycled material represent?
d. Can we make assumptions about causality regarding this study? Explain.
______________________________________________________________________
______________________________________________________________________
3. How would you make the recycling study an experiment?
__________________________________________________________________________
__________________________________________________________________________

Exercise 11.2: Designing a Simple Experiment

Learning Objective: Design a simple experiment.
Application
Suppose children who are encouraged to gesture while generating creative ways to use objects generate
significantly more novel uses for the objects than those who are not encouraged to gesture.
1. If the children are randomly assigned to one of the instruction types, what type of design is
this?
2. The IV for the gesturing study is ; the DV is
.
3. Would this design allow you to conclude that encouraging gesturing causes an increase in the
generation of novel uses for objects? Explain your answer.
_________________________________________________________________________
_________________________________________________________________________
4. List three ways a researcher could maximize the power of the gesturing study.
 Chapter 11 ■ Independent-Groups Designs   99

a. _________________________________________________________________________

b. _________________________________________________________________________

c. _________________________________________________________________________

5. Name two steps you could take to ensure the reliability and validity of the DV in
this study.

__________________________________________________________________________

__________________________________________________________________________

6. Name two potential confounds that you would want to eliminate from this study.

__________________________________________________________________________

__________________________________________________________________________

7. What ethical concerns might arise from the study and how might you address them?

__________________________________________________________________________

__________________________________________________________________________

__________________________________________________________________________

__________________________________________________________________________

Exercise 11.3: Independent-Samples t Tests, Effects Sizes, and

Confidence Intervals
Learning Objective: Analyze and interpret a two-group design using an independent-samples t test, effect size,
and confidence interval.
Review

1. The inferential statistic used to analyze a two-group independent-samples designs is

__________________________________________________________________________

2. Circle which of the following assumptions should be met in order to appropriately compute
an independent-samples t test.

a. Nominal/ordinal data Homogeneity of variance

b. Interval/ratio data Study is an experiment

c. Three or more groups Normally distributed measurement variable

d. Random assignment to groups Pilot Study

3. An independent-samples t test is used to determine whether the means of two groups differ
more than we would expect by .

4. In an independent-samples t test, the difference between the means of two groups is

compared to a sampling distribution composed of differences between the means of
two groups, when .
100   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

5. Match the symbols on the left (which are associated with analyzing data from two
independent-samples) with the appropriate term or phrase on the right.

µ1 ≠ µ2 a. an effect size that reflects the standard deviation units between the
means of two groups

SE X – X b. Number scores free to vary

SD2 pooled c. Range that we are 95% confident that our mean difference varies

df d. t value that we calculate

t = 0.0 e. a
 n effect size that notes the percentage of variability in the DV accounted
for by the IV

tcrit f. v ariability of the entire sample calculated by dividing variance of each

group its n

tobt g. nondirectional alternative hypothesis

d h. expected t value if the null hypothesis is true

rpb2 i. t value that represents the extreme 2.5 % of a distribution for a two-tailed
test

CI [3.20, 6.89] j. standard error of the difference between means

6. In the formula for an independent-samples t test, the difference between the means is divided
by which is symbolized by .
7. The standard error of the difference between means is (larger or smaller) than the
standard deviation of each group in a study because .
_________________________________________________________________________
8. To find the tcrit for a study, we need the following: , , and
.
9. Suppose for one of our groups SD = 1.50 and for the other group SD =15.00. What
assumption of a t test might we be worried about?
_________________________________________________________________________

Application A
A study investigates the effect of different methods of taking notes in class (entering notes on a laptop or
writing notes in longhand on paper) on student retention of material. Students (N = 27) are randomly
assigned to the two conditions. They hear the same 30-minute lecture and 30 minutes later take a short-
answer test on the material (maximum score = 25).
1. What type of independent-groups design is this study?
Explain what characteristics define the design you selected.
_________________________________________________________________________
_________________________________________________________________________
2. State a directional alternative hypothesis for this study.
_________________________________________________________________________
_________________________________________________________________________
3. For this study: IV = ; levels of the IV are
and .
 Chapter 11 ■ Independent-Groups Designs   101

DV = ____________________
df = ____________ for a two-tailed test at p < .05, tcrit = ___________________
4. Suppose Mlaptop = 18; Mlonghand = 22; t(25) = 5.32. Are these results statistically significant for a
two-tailed test at p < .05? Explain.
_________________________________________________________________________
_________________________________________________________________________
5. The researcher finds rpb2 = .20. Interpret the meaning of this finding.
_________________________________________________________________________
_________________________________________________________________________
6. Suppose 95% CI [2.65, 5.38] for the study. Write a sentence describing the meaning of this result.
_________________________________________________________________________
_________________________________________________________________________
7. In another study exploring the same variables, 95% CI [1.65, 7.32]. Which of these confidence
intervals would a researcher be more satisfied with? . Explain your choice.
_________________________________________________________________________
_________________________________________________________________________
8. A second study examines the effect of the two types of note taking on a multiple-choice test
(rather than a short-answer test). The results show t(38) = 1.98. Are these results statistically
significant for a two-tailed test at p < .05? Explain.
_________________________________________________________________________
_________________________________________________________________________
9. The researcher for this study finds that d = .15. Interpret the meaning of this finding.
_________________________________________________________________________
_________________________________________________________________________

Application B
A developmental psychologist observed seven children with their parents. She recorded whether or not
the parents used any harsh language (yes or no) and rated the level of child behavior problems exhibited.
The data appear below.

Child ID Ratings of Child Behavior

number Use of Harsh Language Problems

1 No 4

2 No 4

3 Yes 7

4 No 3

5 No 4

6 Yes 7

7 Yes 5
102   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

1. Using the formula in Chapter 11 of the textbook or the Computational Formula in Appendix
D.6, calculate t

2. Calculate the df
3. Using Table C.4 in Appendix C, are the results statistically significant for a two-tailed test at
p < .05? For p < .01?
4. Compute an effect size and interpret its meaning.

5. Compute the confidence interval for the mean difference and interpret its meaning.

6. Comment on the practical significance of these findings.

____________________________________________________________________
____________________________________________________________________

Application C
A student is interested in whether faculty who use humor in the classroom are liked more than faculty
who do not use humor. She arranges for a graduate student to be the “guest teacher” in two sections of
the same class. The content of the class is the same but the graduate student uses humor in one section
and omits humor in the other class. At the end of each class, the students complete a survey about the
guest teacher, including how much they liked the class.

1. What type of independent-groups design is this study? Explain what

characteristics define the design you selected.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
2. State a directional alternative hypothesis for this study.
_________________________________________________________________________
3. For this study: IV is ; levels of the IV are and .
a. DV = _________________________________________________________________
b. Use the output in the following tables to answer questions 4–9.
 Chapter 11 ■ Independent-Groups Designs   103

Group Statistics
Teach Group N Mean Std. Deviation Std. Error Mean
Likability Humor 12 8.1250 2.97050 .85751
No humor 12 6.1667 1.19342 .34451

Independent Samples Test

Levene’s Test
for Equality of
Likability Variances t test for Equality of Means
95% Confidence Interval of the
Difference
Sig.
Mean Std. Error Upper
F Sig. t df (2-tailed) Difference Difference Lower
Equal variances assumed 23.288 .000 2.119 22 .046 1.95833 .92413 .04181 3.87485
Equal variances not 2.119 14.461 .052 1.95833 .92413 –.01781 3.93448
assumed

Independent Samples Effect Sizes

95% Confidence Interval

Standardizer a
Point Estimate Lower Upper

Likability Cohen’s d 2.26364 .865 .017 1.696

Hedges’s correction 2.34464 .835 .016 1.637

Glass’s delta 1.19342 1.641 .571 2.668

. The denominator used in estimating the effect sizes.

a

Cohen’s d uses the pooled standard deviation.

Hedges’s correction uses the pooled standard deviation, plus a correction factor.
Glass’s delta uses the sample standard deviation of the control group.

4. Which of the assumptions for an independent-samples t test was not met?

________________________ How do you know?

5. How does this change your interpretation of and conclusions about the output?

6. Interpret the effect size for the study: ________________________

7. Write a few sentences summarizing the findings as you would in a Results section.
104   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

8. What conclusions can you make about the findings?

9. Name a possible flaw in the study and at least one suggestion for future studies.

Application D
Are male presidents of community colleges paid a higher salary than female presidents at these
institutions? A researcher documents the salaries of 22 presidents who were hired within the last 2 years.

1. What type of independent-groups design is this study? __________________ Explain what

characteristics define the design you selected.

2. State a directional alternative hypothesis for this study.

3. For this study, the predictor = ; the outcome = .

Use the output in the following tables to answer questions 4-7. Values/dollars are expressed in thousands.

Group Statistics

Gender N Mean Std. Deviation Std. Error Mean

Salary (in Female 10 170.5000 20.06240 6.34429
thousands) Male 12 198.3333 33.46187 9.65961

Levene’s Test
for Equality of
Variances t test for Equality of Means

95% Confidence Interval of

Salary (in the Difference
thousands) Sig.
Mean Std. Error
F Sig. t df (2-tailed) Difference Difference Lower Upper
Equal 1.907 .183 –2.303 20 .032 –27.83333 12.08756 –53.04753 –2.61913
variances
assumed
Equal –2.408 18.361 .027 –27.83333 11.55673 –52.07895 –3.58772
variances
not assumed
 Chapter 11 ■ Independent-Groups Designs   105

4. Compute Cohen’s d as an effect size:

5. Are the salaries of male and female community college presidents significantly different?
How do you know?
6. Write a Results section, describing the results of this study.

7. Based on these findings, what conclusions can you make about gender and community
college presidents’ salaries?

Exercise 11.4: Designs With More Than Two Independent Groups

Learning Objective: Discuss the characteristics and advantages of multiple-groups designs.
Review

1. Designs with more than two independent groups are called

.
2. The three types of multiple independent-groups design are ,
, and .
3. Circle all of the advantages of multiple independent-groups designs relative to multiple two
independent-groups designs.

Decreased probability of Type I error Can use multiple control groups

Decreased external validity Decreased probability of Type II error

Increased internal validity More efficient

Need fewer participants Allows examination of functional relationships

Easier to analyze More reflective of reality

Exercise 11.5: Analysis of Multiple Independent-Groups Designs

Learning Objective: Analyze and interpret data for multiple-groups designs using a one-way ANOVA and
post hoc tests.
Review
1. A one-way ANOVA compares the variance to the variance.
2. Another name for between-group variance is variance.
3. Another name for error variance is variance.
4. Researchers try to increase the variance and minimize the
variance.
5. Because we can never eliminate individual differences or uniqueness of individuals, the
between-group variance consists of both variance and variance.
106   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

6. F = variance/ variance.
7. If our null hypothesis is true (our treatment has no effect), we would expect F =
.
8. Complete the 5 assumptions for computing a one-way ANOVA:
IV has ______________________groups.
Groups are ________________________.
DV is ____________________________scale.
DV is ____________________________.
Variability in groups is ________________; also called __________________________.
9. The values computed for an ANOVA are presented in a table.
10. The first step in computing an ANOVA is to subtract the mean from each of its scores and
then square the deviation scores and sum them. This value is called the
.
11. We also subtract the total mean from each group mean, square the deviation scores and sum
them. This value is called the .
12. We have two dfs in a one-way ANOVA, one for groups and one for
groups.
13. Using the formulas in Chapter 11 of the textbook, complete the following Summary Table for
an experiment comparing four independent groups:

Sum of Squares df Mean Square F

Between Groups 43.844 ?? ?? ??
Within Groups ?? 28 ??
Total 65.219 31

14. a. Calculate the effect size (η2) using the formula in Chapter 11 of the textbook:
______________________________________________________________________
b. What does this effect size tell you?
______________________________________________________________________
15. Based on the summary table:
a. How many total participants are in the study?
b. If the n is equal in each condition, how many participants are in each condition?

c. Is the result statistically significant at p < .05? How can you tell?
______________________________________________________________________
d. What test should you compute next and why should you do this?
______________________________________________________________________
e. How many paired comparisons are there for this study?

Application A
A social psychologist is interested in the impact of men’s clothing on views of their attractiveness.
College students are randomly assigned to a condition in the study. The output for the study is below.
 Chapter 11 ■ Independent-Groups Designs   107

Descriptives
Attractiveness
95% Confidence Interval for
Mean
Std. Upper
N Mean Deviation Std. Error Lower Bound Bound Minimum Maximum
Jeans & Polo 8 7.1250 .76111 .26909 6.4887 7.7613 5.50 8.00
Jeans & T-shirt 8 6.0500 .84007 .29701 5.3477 6.7523 4.50 6.80
Suit 8 4.5000 1.06904 .37796 3.6063 5.3937 3.00 6.00
Pants & vest 8 7.8750 .79057 .27951 7.2141 8.5359 7.00 9.00
Total 32 6.3875 1.53281 .27097 5.8349 6.9401 3.00 9.00

Tests of Homogeneity of Variances

Levene Statistic df1 df2 Sig.
Attractiveness Based on Mean .810 3 28 .499

ANOVA
Attractiveness
Sum of Squares df Mean Square F Sig.
Between Groups 51.465 3 17.155 22.477 .000
Within Groups 21.370 28 .763
Total 72.835 31

ANOVA Effect Sizesa

95% Confidence Interval
Point Estimate Lower Upper
Attractiveness Eta-squared .707 .449 .790
a. Eta-squared and Epsilon-squared are estimated based on the fixed-effect model.

Dependent Variable: Attractiveness

LSD
95% Confidence Interval
(I) Dress Styles (J) Dress Styles Mean Difference (I–J) Std. Error Sig. Lower Bound Upper Bound
Jeans & Polo Jeans & T-shirt 1.07500* .43681 .020 .1802 1.9698
Suit 2.62500* .43681 .000 1.7302 3.5198
Pants & vest –.75000 .43681 .097 –1.6448 .1448
Jeans & T-shirt Jeans & Polo –1.07500* .43681 .020 –1.9698 –.1802
Suit 1.55000* .43681 .001 .6552 2.4448
Pants & vest –1.82500* .43681 .000 –2.7198 –.9302
Suit Jeans & Polo –2.62500* .43681 .000 –3.5198 –1.7302
Jeans & T-shirt –1.55000* .43681 .001 –2.4448 –.6552
Pants & vest –3.37500* .43681 .000 –4.2698 –2.4802
Pants & vest Jeans & Polo .75000 .43681 .097 –.1448 1.6448
Jeans & T-shirt 1.82500* .43681 .000 .9302 2.7198
Suit 3.37500* .43681 .000 2.4802 4.2698
*. The mean difference is significant at the 0.05 level.
108   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

1. What is the IV and how many groups/levels does it have?

2. What is the DV?
3. What is the probability of a Type I error? Type II error?
4. Which clothing style was rated as least attractive? As most attractive?

5. Was the homogeneity of variance assumption violated in this study? Explain how you know.
_________________________________________________________________________
6. What is the effect size for the study and its strength?
7. List the comparisons that met the criteria for statistical significance at the p < .05. Indicate
the exact p value for each of these comparisons, except report p < .001 when SPSS reports a p
of .000.
_________________________________________________________________________
_________________________________________________________________________
8. Report your findings as if you were writing a Results section. Be sure to identify the IV and
DV, descriptive statistics, and all of the required statistical analyses and their findings. Use
APA format.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
9. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________

Application B
Iam Phuny is investigating the effect of facial expression on ratings of females’ job success. Participants
are randomly assigned to view a picture showing only the head and shoulders of a woman (similar to
one that would appear in an online announcement on a business website or newsletter). They then rate
the female on several characteristics using a 10-point scale (1= not at all to 10 = extremely). Iam is only
interested in ratings of job success. The results of the analysis are below.
Descriptives
Report
Job Success
Facial Express Mean N Std. Deviation
Neutral 7.4667 15 1.95911
Frown 5.7333 15 2.31352
Smile 5.6000 15 2.26148
Total 6.2667 45 2.30020
 Chapter 11 ■ Independent-Groups Designs   109

Tests of Homogeneity of Variances

Job Success
Levene Statistic df1 df2 Sig.
.332 2 42 .720

ANOVA
Job Success

Sum of Squares Df Mean Square F Sig.

Between Groups 32.533 2 16.267 3.411 .042

Within Groups 200.267 42 4.768

Total 232.800 44

ANOVA Effect Sizes

Eta Eta Squared
Job Success * Facial .374 .140
Express

Multiple Comparisons
Dependent Variable: Job Success

LSD

95% Confidence
Interval
Mean
(I) Facial (J) Facial Difference Std. Lower Upper
Express Express (I–J) Error Sig. Bound Bound

Neutral Frown 1.73333* .79735 .035 .1242 3.3425

Smile 1.86667 *
.79735 .024 .2575 3.4758

Frown Neutral −1.73333 *

.79735 .035 −3.3425 −.1242

Smile .13333 .79735 .868 −1.4758 1.7425

Smile Neutral −1.86667 * .79735 .024 −3.4758 −.2575

Frown −.13333 .79735 .868 −1.7425 1.4758

*
The mean difference is significant at the 0.05 level.

1. Name the IV and levels of the IV:

2. Name the DV:
3. What type of design is the study?
4. What scale of measurement does the IV represent?
5. What scale of measurement does the DV represent?
6. State the null and directional alternative hypotheses for the experiment.
H0:_______________________________________________________________________
_________________________________________________________________________
Ha: _______________________________________________________________________
_________________________________________________________________________
110   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

7. Describe two potential threats to internal validity and what the researcher can do to avoid/
minimize each threat.
i. ______________________________________________________________________
______________________________________________________________________
ii. ______________________________________________________________________
______________________________________________________________________
8. What is the total number of participants in the study?
9. What is the probability of a Type I error? Type II error?
10. Graph the results (the group means).
11. Describe the results in APA format.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
12. Comment on the external validity of the study.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
13. What would you recommend to employed females?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
14. Discuss the statistical significance, effect size, and practical significance for this study.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________

Exercise 11.6: The Big Picture: Selecting Analyses and Interpreting

Results for Independent-Groups Designs
Learning Objective: Formulate the big picture for identifying and analyzing independent-groups designs.

1. Identify the flaw in each item below and correct it.

a. Caroline is interested in whether there are more male or more female professors at
her university. She counts the number of male and female faculty in each department
and computes a t test comparing these numbers. The t test is significant and Caroline
concludes that significantly more females than males teach at her school.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
 Chapter 11 ■ Independent-Groups Designs   111

b. Jack is conducting a study on math anxiety. He has participants take the same 10-item
math test, but the groups differ in how long they have to complete the test (10 minutes,
20 minutes, 30 minutes). He assigns the first 25 participants to the 30-minute group, the
next 25 participants to the 20-minute group, and the last 25 participants to the 10-minute
group. He finds that the 30-minute group members score significantly better on the exam
and report lower anxiety, and he concludes that math teachers should allow more time for
all tests because it can cause a decrease in math anxiety.
______________________________________________________________________
______________________________________________________________________
c. Sharon has just completed a study on the effect of different fonts on reading
comprehension. Her analysis shows F(2, 47) = 1.04, p = .45. She reports that she had a
45% chance of a Type I error and no chance of a Type II error.
______________________________________________________________________
______________________________________________________________________
2. For the following studies, name the type of design and what characteristics define the design,
the appropriate statistical analyses, and the appropriate interpretation. We have completed
some of the table for you.

Inferential Statistic
Study Design Characteristics and Effect Size Interpretation

Does choice impact enjoyment? Fifth graders in two two groups, IV independent- samples correlation
one class were assigned a book by their teacher independent manipulated, DV is t test, with either rpb2 or
while students in a second class chose from 10 groups, quasi- interval, no random Cohen’s d as the effect
books. After two weeks, students rated how much experiment assignment size and confidence
they enjoyed reading their book on 10-point scale. interval for the mean
difference

Do people who belong to different political parties multiple

(Democrat, Independent, Republican) differ in independent
their suspicions of politicians, as rated by four groups,
items that sum to a maximum of 20 points? correlational

Do Latinx and Blacks receive harsher sentences three groups, causation

than Caucasians for the same crime? Judges random
read the same case but are randomly assigned assignment, IV
one of three ethnicities for the convicted man and manipulated, DV is
recommend a prison sentence (in years). ratio

Are people more traumatized by hurricanes or two preexisting

tornados? Trauma is measured with a 50-point groups, no
questionnaire, with higher scores representing manipulation,
more trauma. outcome is interval

Do adults who are obese show more attention two

to visual food cues than to attractive nonfood independent
cues? The participants who had just eaten were groups,
randomly assigned to view a series of either experiment
food or non-food cues and gaze (in seconds) was
recorded.

What techniques can undocumented immigrants one-way between-

learn to reduce their anxiety? Groups of groups ANOVA, with η2
immigrants who had been in the United States as effect size, post hoc
for less than a year were taught yoga, meditation, tests if F is statistically
stress reduction techniques, or nothing. Anxiety significant
was measured after a month of sessions.
112   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Your Research
1. State a hypothesis for a simple experiment on your topic 4. What inferential statistic would you use to test your
(one IV, with two levels, and one DV). This should be based hypothesis? Explain.
on past research on your topic.




5. Which of these two designs is more appropriate, given the
2. What inferential statistic would you use to test your state of the research on this topic? Explain.
hypothesis? Explain.





3. Turn your simple experiment into a multiple-groups
design by adding at least one more level to your DV. This
should be based on past research on your topic.


 12
DEPENDENT-GROUPS DESIGNS

LEARNING OBJECTIVES

• Describe the two types of dependent designs—matched and repeated measures.

• Discuss the advantages and drawbacks of dependent designs.
• Analyze and interpret a dependent two-groups design using a dependent-groups t test, effect size, and
confidence interval.
• Distinguish the different types of counterbalancing used with dependent multiple-groups designs.
• Analyze and interpret a dependent multiple-groups design using a dependent-groups one-way ANOVA,
effect size, and post hoc tests.
• Formulate the big picture for identifying and analyzing dependent-groups designs.

CHAPTER SUMMARY
In this chapter, you learned about two types of dependent designs: matching and repeated measures. Their major advantage
is the increased power they provide by decreasing variability caused by the unique characteristics of participants. By
decreasing error variability, you should be able to better see the effect of your IV (in experiments) or predictor variable (in
correlational studies). The two types of dependent designs reduce participant variability in different ways. In a matched
design, you identify a variable that is highly correlated to the dependent (or outcome) variable and match participants on
this variable. You then ensure that one of the matched partners is in each of the study conditions. In a repeated-measures
design, each participant experiences each condition in the study. Dependent designs may involve two groups or multiple
groups, but the rationale and implementation are the same.
As with all designs, the dependent designs have potential disadvantages. Both dependent designs require extra steps
that can complicate the implementation of the study. In a matched design, you must be able to identify an appropriate
matching variable that doesn’t sensitize the participants to your hypothesis or the purpose of the study. Repeated-measures
designs are vulnerable to sensitization, fatigue, practice, and carryover effects and require counterbalancing of the order of
treatments. In addition, you can’t use a repeated-measures design when your IV creates a permanent change.
In addition to learning about the types of dependent designs, you learned about the statistics used to analyze the data
generated from them. The dependent-samples (or paired-samples or repeated-measures) t test examines the differences
between matched pairs, or the same participant under different conditions, and compares this mean difference to the
error within conditions. The dependent-samples (or within subjects or repeated measures) one-way ANOVA is used

113
114   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

to analyze the data from a dependent multiple-groups study. This ANOVA separates error variability
into the variability associated with a particular participant (in a repeated-measures study) or with the
matched participants (in a matched design) and the variability that is error due to other sources (subject
by condition interaction).
As a researcher, you should consider the advantages and disadvantages of different designs before deciding
what type of design (independent or dependent and correlation, quasi-experiment, or experiment) will
best test your hypothesis.

CHAPTER 12 QUIZ
Test your knowledge of the key terms from the chapter.
1. Where does the power of the dependent-groups design come from?
a. the large sample size
b. the stronger manipulation
c. the more sensitive dependent variable
d. the decrease in random error that is created by participant characteristics
2. The matching variable you choose should be ______
a. highly correlated with your dependent variable.
b. the same as your dependent variable.
c. the same as your independent variable.
d. something that increases the variance between groups.
3. Which of the following is true about the relationship between design type and external validity?
a. A matched design with the same variety of participant characteristics as an independent
design will have less external validity than the independent design.
b. A matched design with the same variety of participant characteristics as an independent
design will have the same external validity as the independent design.
c. A matched design with the same variety of participant characteristics as an independent
design will have more external validity than the independent design.
d. There is no relationship between design type and external validity.
4. Counterbalancing eliminates ______
a. attrition.
b. order effects.
c. mortality.
d. power.
5. If the impact of the treatment or measurement lasts longer than the time between different
conditions, this is known as ______
a. boredom.
b. practice.
c. fatigue.
d. carryover.
 Chapter 12 ■ Dependent-Groups Designs   115

6. The average difference between the scores of matched pairs or the scores for the same
participants across two conditions is known as ______
a. the variance.
b. the mean difference.
c. the standard deviation.
d. the standard error of the mean difference.
7. ______ is when participants are randomly assigned to different sequences of conditions
so that each condition is represented in each order an equal number of times but not all
sequences are represented.
a. Complete counterbalancing
b. Partial counterbalancing
c. Latin Square counterbalancing
d. Randomized partial counterbalancing
8. The assumption that the variances of the differences between all the combinations of pairs of
groups are equal is known as ______
a. sphericity.
b. deviance of variance.
c. skewness.
d. counterbalancing.
9. The error associated with participants in the same condition or interaction sums of squares is
symbolized as ______
a. SSs.
b. SSAxB .
c. SSi.
d. Iss.
10. ______ is used to assess the effect size of a dependent-groups ANOVA.
a. Pearson’s r
b. rpb2
c. Partial eta
d. Partial eta squared

Quiz 12 Answer Key

1. d. 2. a. 3. b. 4. b. 5. d. 6. b. 7. b. 8. a. 9. b. 10. d.

CHAPTER 12 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 12.1: Designs With Dependent-Groups

Learning Objective: Describe two types of dependent designs−matched and repeated measures.
116   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Review
1. Put an I by the characteristics of independent-groups designs and D by characteristics of
dependent-groups designs. (Hint: some characteristics fit both types of designs.)
_____________ Can be correlational, quasi-experiments, or experiments
_____________ Can examine causation or correlation
_____________ Are the more powerful design
_____________ May have the same participants in more than one condition
_____________ Achieve equal groups through random assignment
_____________ Decrease random error created by participant characteristics
_____________ More vulnerable to confounds due to participant variables
_____________ More sensitive to changes in the DV created by changes in the IV
_____________ Can be matched or repeated measures designs
2. Distinguish matched and repeated measures designs.


3. a. What is different about the matching process for correlational and quasi-experiments and
the matching process for experiments?


b. How is the matching process the same for the different types of dependent designs?


4. Another name for a repeated-measures design is design.
5. Explain the difference between a repeated-measures design and the use of multiple trials in a study.




Exercise 12.2 Advantages and Drawbacks of Dependent Designs

Learning Objective: Discuss the advantages and drawbacks of dependent designs.
Review
1. How do matched and repeated measures designs each reduce error variance?


2. Name three potential problems or disadvantages of matching designs.
a. _______________________________________________________________________
b. _______________________________________________________________________
c. _______________________________________________________________________
 Chapter 12 ■ Dependent-Groups Designs   117

3. Why do we use counterbalancing in repeated-measures designs?

4. Repeated-measures designs require ___________ (more/fewer) participants than independent-

groups designs or matched-pairs designs.

5. Name three potential problems of repeated-measures designs and how these problems might
occur.

a. _______________________________________________________________________

b. _______________________________________________________________________

c. _______________________________________________________________________

Application: Matched-Pairs Designs

1. Suppose a researcher wants to examine the effect of different language instruction
on the learning of English by immigrants. She decides to match the immigrants
on age.

a. Comment on the effectiveness of this matching variable.

b. What matching variable might you suggest for this study and why?

2. Another researcher decides to replicate the language instruction study but matches
participants on the variable you suggested and on their time since arrival in the United States.
Explain why this might be a good or bad idea.

3. A researcher is examining perceptions of violence in romantic relationships

involving young adults. Participants in the study will read a scenario describing
violence (a slap during an argument) where the perpetrator is either a male or a female.
They rate several aspects of the relationship as well as the violence of the act. The
researcher believes that one’s view of women may influence attitudes toward violence
against and by women, and she decides to match participants on the Attitudes toward
Women scale (AWS) before assigning them to the perpetrator condition. Below are
the scores for the participants on the AWS. Match them so that they can be randomly
118   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

assigned to the two groups, and enter the matched pairs’ participant numbers in the
last column.

PARTICIPANT AWS SCORE PARTICIPANT AWS SCORE MATCHED PAIRS

1 25 11 5

2 18 12 13

3 12 13 14

4 14 14 17

5 21 15 20

6 15 16 23

7 11 17 15

8 16 18 25

9 6 19 7

10 5 20 6

Application: Repeated-Measures Designs

1. A gerontologist studying the effect of mental stimulation on the alertness of those 90 years or
older has 8 participants available. They are randomly assigned to the order of two conditions:
playing a word game for 10 minutes and watching a comedy segment for 10 minutes.
Alertness is assessed after each session. In order not to overtax the energy of the participants,
the researcher has the participants complete both of the 10-minute sessions back-to-back.
What potential confound(s) can you identify in this study and how would you correct it/
them?



2. A high school basketball coach wants to try a different approach to teaching her players how
to make free throws. She has a session one day when she gives her regular instruction, has
players practice free throws for 15 minutes, and then has each team member attempt 10 free
throws. The next day, she follows the same pattern but tries a new method of instruction
before the free throw practice and attempts. She finds that the players make an average of 1.5
more free throws on the second day. Name a potential confound in this exercise and explain
how to correct it.



3. A teacher wants to examine which of two methods of teaching long division is more effective.
Can the teacher use a repeated-measures design? Why or why not?



4. Ima Ipad is a senior psychology major. For her senior thesis, she wants to study the effect
of sleep deprivation on driving skills. After months of negotiating with the engineering
department, she is granted access to the needed driving simulation equipment. She must
 Chapter 12 ■ Dependent-Groups Designs   119

now decide whether to use an independent or dependent design. What would you advise
her and why?




Exercise 12.3: Analysis of Dependent Two-Group Designs

Learning Objective: Analyze and interpret a dependent two-group design using a dependent-groups t test,
effect size, and confidence interval.
Review
1. Dependent two-group designs (matched or repeated measures) are analyzed using a
test.
2. Alternative names for the dependent-samples t test are and
.
3. The independent-samples t test focuses on individual scores while the dependent-samples t test
focuses on differences

4. The major difference in the assumptions for the independent- and dependent-samples t test is
that

5. The numerator for the dependent-samples (or within-subjects or paired-samples) t test is
computed by adding up the differences between the matched scores or repeated measures and
is called the ; it is symbolized by .
6. The denominator for the dependent-samples t test represents the standard deviation of the
for µ D = 0. It is called the and is symbolized by .
7. The standard error of the mean difference is always (larger/smaller) than the
standard deviation for the sample because

8. The formula for df for a paired-samples t test is where N = , rather

than individual scores.

Application A
Suppose we find in a study comparing student retention in an online (M = 7.08, SD = 1.08)
versus in-person (M = 7.83, SD = 1.53) lecture with 24 students paired on their GPA that tobt =
2.46. Retention was measured with a 10-point multiple choice test.
1. df = ________________
2. tcrit for a two-tailed test, p < .05 = ________________
3. Do the two lecture presentation modes result in a statistically significant difference in student
retention? ________________ How do you know?

4. If SED = .30, compute the 95% confidence interval using the following formula:
(SED) (–tcrit) + (MD) ≤ µD ≤ (SED) (+ tcrit) + (MD)



120   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

5. If SDD = 1.06, compute Cohen’s d as the effect size for the study and comment on its strength.
(See the effect size formula for two dependent samples in Chapter 12.)



6. Comment on the practical significance of the study’s findings.




Application B
Suppose you are the gerontologist above who is examining the effect of two types of stimulation (30
minutes of word games or comedy on TV) on the alertness of those who are over 90 years of age. Each
participant is randomly assigned to the order of the two stimulations, which are provided two days
apart. Alertness is measured on a 25-point scale immediately following each stimulation session.

1. What type of dependent design is this? ___________________________________________

2. What ethical concerns should the researcher be sure to address in this study?






3. State the null and a directional alternative hypothesis for the study.
H0: ______________________________________________________________________
Ha: ______________________________________________________________________
The output for the study is below:
PAIRED SAMPLES STATISTICS
Mean N Std. Deviation Std. Error Mean

Pair 1 word game 14.5000 8 5.29150 1.87083

comedy 13.0000 8 4.65986 1.64751

PAIRED SAMPLES CORRELATIONS

N Correlation Sig.

Pair 1 word game & comedy 8 .950 .000

PAIRED SAMPLES TEST

Paired Differences

95% Confidence
Interval of the
Std.
Difference
Error
Mean Std. Deviation Mean Lower Upper t Df Sig. (2-tailed)

Pair 1 word game – comedy 1.50000 1.69031 .59761 .08687 2.91313 2.510 7 .040
 Chapter 12 ■ Dependent-Groups Designs   121

4. df = ______________________
5. How many participants were in the study? ______________________ How do you know?


6. Does mental alertness differ significantly for elderly people after playing word games vs.
watching comedy? Explain your response.


7. Compute the effect size for the results using the formula for dependent samples.

8. Describe the findings in APA format, including all information that is expected in a Results
section.






9. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.







Application C
Are people with harder-to-pronounce names considered more threatening than those with easier-to-
pronounce names? A researcher matches participants (college students) on a scale measuring openness
to new experiences. One of each matched pair of students is assigned to rate people with easy-to-
pronounce names while the other rates people with hard-to-pronounce names with the same number
of letters. All students view the same pictures of a diverse group of 10 adults who are rated on several
characteristics. The findings for ratings of “threatening” are shown below. Threatening was rated on a
7-point scale, with higher scores reflecting more perceived threat.

Paired Samples Statistics

Mean N Std. Deviation Std. Error Mean

Pair 1 easy pronounce 3.7333 15 1.43759 .37118

hard pronounce 4.6667 15 1.34519 .34733

122   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Paired Samples Correlations

N Correlation Sig.

Pair 1 easy pronounce & hard pronounce 15 .025 .931

Paired Samples Test

Paired Differences

95% Confidence Interval

of the Difference
Std. Error Sig.
Mean Std. Deviation Mean Lower Upper t df (2-tailed)

Pair 1 easy pronounce −.93333 1.94447 .50206 −2.01014 .14348 −1.859 14 .084
- hard
pronounce

1. What type of dependent design is this? _________________________

2. State the null and a directional alternative hypothesis for the study.
H0: ______________________________________________________
Ha: ______________________________________________________
3. df = _________________________
4. How many participants were in the study? _________________________ How do you know?
__________________________________________________________________
5. Do perceptions of people with easy-to-pronounce names differ significantly from those with
hard-to-pronounce names in terms of their threat to others? _________________________
Explain your answer.

6. Is there a probability of a Type I error? _________________________ Is there a probability of
a Type II error? _________________________
7. Compute the effect size using the formula for dependent samples.

8. Describe the results using APA format.




9. Interpret the meaning of the results in a few sentences.




Exercise 12.4: Designs With More Than Two Dependent Groups

Learning Objective: Distinguish the different types of counterbalancing used with dependent multiple-groups
designs.
 Chapter 12 ■ Dependent-Groups Designs   123

Review

1. The _________________________ (same/different) advantages occur for two-group and

multiple-group dependent designs in comparison to independent-group designs.
2. Distinguish the different types of counterbalancing used with multiple groups designs and
note when each might be most appropriate.
Complete counterbalancing

Partial counterbalancing

Latin square counterbalancing


Randomized partial counterbalancing

3. Counterbalancing controls for __________ effects but not for _________ effects, so the
researcher should ensure adequate time between conditions.
4. In comparison to two-group matched designs, matching with multiple-groups designs is
__________ (more/less) complicated because

5. With correlational or quasi-experimental designs, matching occurs





Application
1. Suppose a weight loss program was examining the effect of weighing every day (1), every other
day (2), every third day (3), or once a week (4) on participants’ confidence that they could lose
weight. Participants were matched on their age and BMI.
a. The possible number of sequences for this study is _________________________.
b.  If the experimenter randomly assigned participants to all possible sequences, this would be
an example of _________________________ counterbalancing.
c.  The experimenter decided to randomly assign each group of participants to the following
sequences: 1234, 2341, 3412, 4123.
This is an example of _________________________ counterbalancing.
d. Evaluate the matching variables. Can you think of a more effective matching variable?




124   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. A researcher explores the impact of time outside on third graders’ completion of work in the
classroom. Each day of one week each child spends 10, 15, 20, 25, or 30 minutes outside
(every child spends each of the time periods outside over the course of one week). The
completion of work immediately after the child returns to the classroom is measured.
a. What type of design is this? _______________________________
b. What potential confound do you see in the study?
______________________________________________________________________
c. What additional step would the researcher need to take to avoid this confound? Explain
in detail what the researcher would need to do.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
3. A researcher is investigating the impact of verbal reinforcement on 4-year-olds’ scores on a
vocabulary test. The test is individually administered and requires the examiner to read a
question and then have the child choose a picture from several choices. He uses a sample
of children who receive 10 questions followed by verbal reinforcement (“good”) for correct
answers only (C), 10 questions followed by verbal reinforcement for both correct and
incorrect answers (B), and 10 questions followed by no reinforcement (N).
a. How many different orders of the reinforcement conditions are possible?
______________________________________________________________________
b. Show how you would use complete counterbalancing to decide the order of the conditions
(C, B, N) that each child would receive. Remember that the number of orders should
match the number you named in the previous question.
______________________________________________________________________
c. How would the process of assigning children to conditions differ if you used Latin square
counterbalancing? Name the specific orders you might use.
______________________________________________________________________
______________________________________________________________________

Exercise 12.5: Analysis of Dependent Multiple-Groups Designs

Learning Objective: Analyze and interpret a dependent multiple-groups design using a dependent-groups one-
way ANOVA, effect size, and post hoc tests.
Review
1. Dependent multiple-groups studies are analyzed using a - ,
also called a - or a - .
2. Assumptions for the dependent-groups ANOVA include which of the following (circle all
that apply)?

Independent groups A manipulated variable

Causation Three or more conditions or groups

Ordinal or interval data Interval or ratio data

Equal variances for the differences between Measurement variable that is normally
pairs of groups distributed

Dependent groups
 Chapter 12 ■ Dependent-Groups Designs   125

3. The term used to describe equal variances for the differences between groups in a one-way
dependent-groups ANOVA is _____________________.
4. Match the symbols on the left (associated with analyzing data from a within-subjects
ANOVA) with the appropriate term or phrase on the right.

Ha Equal to SSA + SSs + SSAxS

SStot The variability associated with individual participants that

we can extract from the error term

SSw MSA /MSAxS

SSS µ1 ≠ µ 2 ≠ µ 3

dfAxS Effect size for dependent-groups ANOVA

MSA SSA /dfA

Fobt Equal to SSS + SSAxS

ηpartial
2
(k – 1)(kps – 1)

Application A
1. Complete the following summary table.

Source SS DF MS F
Condition (A) 126.20 2 ?? ??

Subject ?? ??

Interaction (A x S) 22.47 ?? ??

Total 224.47 29

2. Are the results of this study statistically significant? _________________ How do you know?

3. Compute the effect size using information from the Summary Table above. _______________
4. Interpret the meaning of the effect size.


5. What next step (in terms of analysis) should you take?

6. How many paired comparisons should you make in the post hoc tests?
7. Suppose the following descriptive statistics are found for the groups in the study.

Descriptive Statistics
Mean Std. Deviation N

Group 1 15.4000 3.62706 10

Group 2 15.5000 4.50309 10

Group 3 19.8000 3.19026 10

Compute LSD and complete the matrix below. Note with an asterisk which paired-comparisons meet
the criteria for statistical significance at p < .05.

126   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

1 2 3
Means (15.400) (15.500) (19.800)

Group 1 — ?? ??

Group 2 — — ??

Group 3 — — —
*
p < .05

Application B
1. State a null and an alternative hypothesis for a study that examined the effect of weighing
every day (1), every other day (2), every third day (3), or once a week (4) on participants’
confidence that they could lose weight.
H0: ___________________________________________________________
Ha: ___________________________________________________________
2. Complete the Summary Table below for the weigh-in study.

Source SS DF MS F
Weighing (A) 24.60 ?? ?? ??

Subject 72.72 36

Interaction (A x S) ?? ?? ??

Total 151.32 39

3. What is Fcrit for p < .01? ___________ Are these results significant at p < .01? ___________
4. Are these results significant at p < .05? ___________ How do you know?

5. Compute the effect size and interpret its meaning.


6. How many paired comparisons should you make in the post hoc tests? ___________
7. The descriptive statistics for the groups are shown below. Compute LSD and draw a matrix to
show which groups differ significantly from one another.

Group Mean (M)

Daily 2.6

Every other day 3.8

Every third day 5.5

Weekly 7.2

Application C
A researcher replicated the weigh-in study described in Application B in a geographic region that is
known for its high rate of obesity and examined the effect of weighing every day (1), every other day
(2), every third day (3), or once a week (4) on participants’ confidence that they could lose weight. The
output for this study is found below.
 Chapter 12 ■ Dependent-Groups Designs   127

Descriptive Statistics
Mean Std. Deviation N
Daily 2.8750 1.24642 8
every other day 3.0000 1.30931 8
every third day 4.0000 1.06904 8
Weekly 6.5000 1.60357 8

Mauchly’s Test of Sphericitya

Measure: weigh-in
Within Subjects Mauchly’s W Approx. Chi-Square df Sig. Epsilonb
Effect Greenhouse- Geisser Huynh- Feldt Lower- bound
factor1 .534 3.587 5 .615 .698 1.000 .333

Tests the null hypothesis that the error covariance matrix of the orthonormalized transformed dependent variables is
proportional to an identity matrix.
a. Design: Intercept Within Subjects Design: factor1
b. May be used to adjust the degrees of freedom for the averaged tests of significance. Corrected tests are displayed in
the Tests of Within-Subjects Effects table.

Tests of Within-Subjects Effects

Measure: weigh-in
Source Type III Sum df Mean F Sig. Partial Eta
of Squares Square Squared
factor1 Sphericity Assumed 67.844 3 22.615 21.679 .000 .756
Greenhouse-Geisser 67.844 2.095 32.376 21.679 .000 .756
Huynh-Feldt 67.844 3.000 22.615 21.679 .000 .756
Lower-bound 67.844 1.000 67.844 21.679 .002 .756
Error (factor1) Sphericity Assumed 21.906 21 1.043
Greenhouse-Geisser 21.906 14.668 1.493
Huynh-Feldt 21.906 21.000 1.043
Lower-bound 21.906 7.000 3.129

Pairwise Comparisons
Measure: weigh-in

95% Confidence Interval for Differenceb

(I) factor1 (J) factor1 Mean Difference (I-J) Std. Error Sig.b Lower Bound Upper Bound
1 2 −.125 .441 1.000 −1.727 1.477
3 −1.125 .479 .308 −2.868 .618
4 −3.625* .420 .000 −5.152 −2.098
2 1 .125 .441 1.000 −1.477 1.727
3 −1.000 .423 .299 −2.536 .536
4 −3.500* .567 .003 −5.561 −1.439
3 1 1.125 .479 .308 −.618 2.868
2 1.000 .423 .299 −.536 2.536
4 −2.500* .681 .048 −4.977 −.023
4 1 3.625* .420 .000 2.098 5.152
2 3.500* .567 .003 1.439 5.561
3 2.500* .681 .048 .023 4.977
Based on estimated marginal means
. The mean difference is significant at the .05 level.
*

b. Adjustment for multiple comparisons: Bonferroni.

128   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

1. Fill in the blanks: F (__________,__________) =__________, p__________ ___________.

2. Timing of weigh-ins accounted for ___________ % of the variance in confidence.
3. Write a Results section for this study, using APA format and including all needed information.






4. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.







Exercise 12.6: The Big Picture: Selecting Analyses and Interpreting

Results for Dependent-Groups Designs
Learning Objective: Formulate the big picture for identifying and analyzing dependent-groups
designs.
1. Identify the flaw(s) in each item below and correct it(them).
a. Your classmate solicits participants who believe that they experience strong test anxiety
and has them watch a video on human causes of global warming and then take a test on
the material. She then has them participate in a meditation designed to relax students and
to have them focus on their strengths as a student. The students then watch a video on
additional causes of global warming and take a test on that material. She finds that the
students score significantly higher on the second test and concludes that the intervention
significantly reduces test anxiety.

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________

______________________________________________________________________
b. A social service director offers clients the option to attend a 6-week job training course
or a 6-week life skills course. Of the 50 clients contacted, 34 agree to participate in the
courses. After matching the participating clients on depression, the director assigns one of
each matched pair to the course taught on the night most convenient for them. Six clients
drop out of the courses over the 6 weeks (4 from job training and 2 from life skills). At
the end of the 6-week period, the remaining matched clients report significantly less
depression after taking the life skills course than the job skills course, t(10) = 3.65, p < .05,
d = .20. The director concludes on the basis of these findings that the agency will now
 Chapter 12 ■ Dependent-Groups Designs   129

require all clients to attend the life skills course because attendance in the course causes a
reduction in depression.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
c. A psychologist is hired to try to increase the productivity of employees at a call center.
He decides to test the effect of different types of appreciation on increasing productivity.
He matches the employees at the call center on extraversion and randomly assigns one
of each group to one of three types of appreciation (a weekly free meal provided at work,
an extra hour during one lunch each week, or a free movie pass each week). The study
runs for one month, after which productivity of each employee in the study is measured.
The researcher finds no difference in the productivity of the groups F(2, 38) = 1.58,
p = .16, η 2partial = .20 . The researcher concludes that different types of appreciation do not
increase productivity.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
d. A trainer who offers diversity workshops is interested in the effect of different techniques
on reducing prejudice toward Black individuals in the workplace. Twenty managers of
small businesses volunteer to take her workshop which consists of 4 two-hour zoom
sessions, offered one a week for four weeks. Each session employs a different focus to
reduce prejudice (1. general information and discussion about prejudice, 2. information
and role-plays about attracting and hiring Black applicants, 3. facts about prejudice and
microaggressions toward Black employees and discussion about how to build positive
relationships, 4. role-playing positive interactions with Black employees). Participants
are randomly assigned to the order of the 4 sessions and complete a different form of the
prejudice scale at the end of each session. The analysis finds no difference in the prejudice
scores at the end of the workshop. Upon further examination, however, the trainer
discovers that participation, questions, and discussion were greatly reduced in the last 2
sessions regardless of their focus. ____________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
130   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. For the following studies, name the type of design and the characteristics that define the
design, the appropriate statistical analyses, and the appropriate interpretation. We have
completed some of this for you.

Inferential Statistic
Study Design Characteristics and Effect Size Interpretation

Does alcohol positively impact video game Dependent two Matched pairs, 1 Dependent- samples Causation
performance? Young adults are matched on matched groups, IV on a nominal t test, Cohen’s d
the amount of time they report playing video experiment scale with random or squared point-
games and randomly assigned to drink either assignment of pairs biserial correlation
12 oz. of water or 12 oz. of alcohol. Game to manipulated as effect size,
scores are recorded after the participants condition, DV on ratio Confidence interval
play a video game for 30 minutes. scale

Is relationship status (married, in a Preexisting multiple One-way within-

committed relationship, not in a committed groups matched on subjects ANOVA,
relationship) related to satisfaction with life? relevant variable, η 2partial as effect size.
Participants are matched on age and income. Predictor variable
If F is statistically
is nominal scale,
significant, post hoc
Outcome is interval
tests with correction
scale
for possibility of
increased Type I error

Do children attend more to an educational Dependent

video with very frequent, somewhat frequent, multiple-groups
or no gesturing by the narrator? Students experiment,
are randomly assigned to the viewing order repeated
of the three videos and their direct gaze is measures
measured.

Is there a difference in the hopefulness of a Dependent-samples

class of graduating high school seniors who t test, Cohen’s d
are randomly assigned to take an introduction or squared point-
to college class and a class who is randomly biserial correlation
assigned to take a life after high school class? as effect size,
Students are matched on high school GPA and Confidence interval
family income.

Are sports fans more likely to attend to ads Multiple conditions Correlation
promoting beer, cars, food, or technology? (4) are manipulated,
Patrons at a sports bar are observed as Participants are
prerecorded ads are shown during a pro exposed to all
football game. The percentage of the relevant conditions but not
ads they watch is recorded. randomly assigned
to their order, IV is
nominal, DV is ratio

Do older adults (75+ years) exercise more on a Dependent two

sunny day than on a rainy day? Adults meeting groups, repeated
the age criterion in assisted living are asked measures
on a sunny day how long they exercised and
asked the same question on a rainy day.
 Chapter 12 ■ Dependent-Groups Designs   131

Your Research
1. In the “Your Research” exercise in Chapter 11 of this b. Briefly describe the procedures you would follow
study guide, you selected either a simple experiment for the design you chose.
or a multiple-group design. Restate the hypothesis for

the design you chose.


c. What inferential statistic would you use to test

your hypothesis? Explain.
2. Is it possible to use a dependent-groups design, either

matching or repeated measures, to test this hypothesis?
If no, explain why not. 

 d. What would be the disadvantages of this design

over an independent-groups design?

If yes, answer the following: 


a. Which would be a better design, matching or
repeated measures? Explain.


 13
FACTORIAL DESIGNS

LEARNING OBJECTIVES

• Describe basic concepts in factorial designs.

• Explain the rationale for conducting factorial designs.
• Hypothesize, graph, and interpret main effects and interactions in 2 × 2 designs.
• Analyze and interpret results for two-way between-subjects designs.
• Identify and compare more complex factorial designs and analyses.
• Formulate a big picture of how to conceptualize and analyze factorial designs.

CHAPTER SUMMARY
Examining the impact of one variable on another might provide an incomplete picture, and often it is important to
understand the interaction of multiple variables. Chapter 13 begins by introducing basic concepts in factorial designs,
including the different types of factorials, factorial notation, and the difference between main effects and interaction
effects. Next is a discussion of the rationale for factorial designs, with an emphasis on the benefits of examining complex
relationships and controlling extraneous or confounding variables.
The second part of the chapter focuses on 2 × 2 designs. The chapter includes information on how to graph these designs
and provides examples of different graphs with and without interaction effects. Strategies for articulating interaction
hypotheses are provided, with three typical interaction patterns outlined along with example graphs demonstrating each
pattern. The chapter provides details on how to analyze independent-groups designs using formulas and provides an
example output generated by SPSS. The chapter also provides an overview of independent designs with more than two
factors or with a nominal outcome or dependent variable, dependent-groups factorial designs, and mixed designs. The
chapter ends by suggesting that the factorial design allows one to avoid overly simple explanations and reminds students
of the importance of ethical and critical thinking when analyzing results.

CHAPTER 13 QUIZ
Test your knowledge of the key terms from the chapter.
1. A(n) is a general name for a predictor or independent variable.
a. IV b. DV
c. factor d. factorial

132
 Chapter 13 ■ Factorial Designs   133

2. A(n) is a comparison of one level of a factor across a level of another factor.

a. cell
b. interaction
c. moderator
d. main effect
3. At minimum, a factorial design must have factor(s) and cells.
a. one; two
b. two; two
c. two; four
d. four; two
4. A factorial design that includes manipulation and random assignment for all the factors is
called a factorial.
a. quasi-experimental
b. experimental (or fully experimental)
c. correlational
d. hybrid
5. A researcher conducts an experiment examining the effect of caffeine on memory. If she
wants to examine participant age as a potential moderating factor, she would have a(n)
factorial design.
a. hybrid
b. quasi-experimental
c. experimental (or fully experimental)
d. correlational
6. A 2 (participant religion) × 4 (participant ethnicity) design is
a. an independent-groups factorial.
b. a correlational factorial design.
c. both a and b
d. None of the above
7. A researcher hypothesizes that participation in extracurricular activities will be associated
with better social skills. What type of hypothesis is this?
a. a main effect hypothesis
b. a null hypothesis
c. a cell mean hypothesis
d. an interaction hypothesis
8. To graph a hypothesized interaction for a 2 × 2 factorial, the graph must have two lines that
a. intersect.
b. are not parallel.
c. are parallel.
d. form an “X.”
134   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

9. A researcher hypothesizes that children will develop a stronger sense of self-esteem when
their families have regular dinners together and that this is especially true for children with
learning disabilities. The best analysis to test this hypothesis is
a. an independent-samples t test.
b. a one-way between-subjects ANOVA.
c. a mixed ANOVA.
d. a two-way between-subjects ANOVA.

10. A researcher conducts a repeated-measures experiment examining the effect of teacher

feedback on exam scores. If she wants to also test whether gender of the teacher moderated
this effect, what type of analysis should she use?
a. an independent-samples t test
b. a one-way between-subjects ANOVA
c. a mixed ANOVA
d. a two-way between-subjects ANOVA

Quiz 13 Answer Key

1. c. 2. a. 3. c. 4. b. 5. a. 6. c. 7. a. 8. b. 9. d. 10. c

CHAPTER 13 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 13.1: Basic Concepts in Factorial Design

Learning Objective: Describe basic concepts in factorial designs.
Review
1. A 3 × 5 factorial design has factors, the first with levels and
the second with levels.
2. A 2 × 2 × 2 factorial design has factors, each with levels.
3. A 2 × 2 has cells, and a 2 × 2 × 2 has cells.
4. Write a definition for an interaction effect, using your own words:
_________________________________________________________________________

Application
Use the following variables to create specific types of factorial designs:

Non-Experimental Variables Experimental IVs

gender (male, female) facial expression (smile, frown)

religion (religious, spiritual, or neither) sexism (not sexist, hostile, benevolent)

year in college (1st, 2nd, 3rd, 4th)

 Chapter 13 ■ Factorial Designs   135

Example: A 2 × 3 correlational factorial: gender × religion

1. A 3 × 4 correlational factorial: ×
2. A 2 × 3 experimental factorial: ×
3. A 4 × 3 hybrid factorial: ×
4. A 2 × 3 hybrid factorial: × OR
×
5. A 2 × 4 × 2 hybrid factorial: × ×

Exercise 13.2: Rationale for Factorial Designs

Learning Objective: Explain the rationale for conducting factorial designs.
Application:
1. Research suggests that parents’ salary strongly predicts the salary their child will eventually
earn. However, there are many examples of individuals from poor backgrounds who go on to
have high salaries and also of individuals from wealthy backgrounds who go on to have low
salaries. Explain how a factorial design might help us understand this complex relationship.
_________________________________________________________________________
_________________________________________________________________________
2. A researcher wants to examine how childhood experiences with police officers impact adult
participants’ current feelings about police. The researcher wants to increase power and is
considering limiting the population to only Black participants. What is an alternative strategy
the researcher might use?
_________________________________________________________________________
_________________________________________________________________________
3. A researcher wants to conduct an experiment to examine the effect of reading versus watching
the news on anxiety. To increase internal validity, the researcher decides that the news story
will be the same one about crime. However, the researcher realizes that this will limit external
validity in that it will be unclear whether results are generalizable to other news stories, such
as those about accidents. Make a case for using a factorial design instead.
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________
4. A classmate collected data for a simple experiment but did not find a statistically significant
effect. The classmate is now considering examining how participant characteristics such as
gender and age might have impacted the relationship between the IV and DV. Under what
conditions might this be a good strategy? When might this be a bad strategy?
_________________________________________________________________________
_________________________________________________________________________
_________________________________________________________________________

Exercise 13.3: 2 × 2 Designs

Learning Objective: Hypothesize, graph, and interpret main effects and interactions in 2 × 2 designs.
Application
1. Identify which of the following are main effects, and which one is an interaction:
a. Checking social media within an hour of going to bed will
negatively impact the quantity and quality of sleep.
136   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

b. Drinking tea before going to sleep will improve the quantity and
quality of sleep compared to drinking water.
c. Distractions will impede performance on the math test but
will improve performance on a creativity test, and the opposite effect will occur in a
distraction-free environment.
d. Compared to a placebo, taking medicine to go to sleep will
increase the quantity of sleep but decrease the quality of sleep.
e. Both television viewing and social media use prior to bedtime will
decrease the quality of sleep.
2. For the interaction you identified above, identify:
a. Factor 1: ; levels:
b. Factor 2: ; levels:
c. Outcome/DV:
3. Which of the following graphs best fit the interaction:

10 9 9
9 8 8
8 7 7
7
Performance

Performance

Performance
6 6
6
5 5
5
4 4 4
3 3 3
2 2 2
1 1 1
0 0 0
Math test Creativity test Math test Creativity test Math test Creativity test
A. B. C.

no distractions distractions

4. Graph the results of a study on happiness:

Pet
none cat
50–64
50—64 4 7
Age

65—80 4 7 65–80

5
Happiness

0
No pet Cat
 Chapter 13 ■ Factorial Designs   137

5. Does the steepness and direction of the lines in the graph indicate
a. A main effect for pet?
b. A main effect for age?
c. An interaction?
6. Graph the results of a study on income:

Degree
Associates Bachelors
English 40,000 55,000
Major

Biology 50,000 77,000 English

Biology
80
Income (in thousands)

70

60

50

40

30
Associates Bachelors

7. Does the steepness and direction of the lines in the graph indicate
a. A main effect for degree? _______________
b. A main effect for major? _______________
c. An interaction? _______________
8. Graph the results of a study on quiz scores:

Current Grade in Class

Study Session

A or B C or less
Attended 9.5 8
Did not attend 7.5 6

10
A or B
9 C or less

8
Quiz Score

4
Attended study session Did not attend

9. Does the steepness and direction of the lines in the graph indicate
a. A main effect for study session? ______________
b. A main effect for current grade? ______________
c. An interaction? ______________
138   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

10. For each interaction you identified above, describe the pattern:
_________________________________________________________________________
_________________________________________________________________________

Exercise 13.4: Two-Way Between Subjects ANOVA

Learning Objective: Analyze two-way between-subjects designs.
Review

1. The two-way between-subjects ANOVA is used when you have:

a. How many factors?
b. The factors are independent or dependent?
c. The DV or outcome is measured on what type of scale of measurement?
or
2. Two additional assumptions of the two-way between-subjects ANOVA are
a. _____________________________________________________________________________
b. _____________________________________________________________________________
3. A researcher examines how race of a suspect and race of a victim impact ratings of seriousness
of a crime.
a. The is the average ratings of seriousness of a crime for the entire sample.
b. are the means used to examine the main effect for race of suspect and
the main effect for race of victim.
c. are the means used to examine the interaction.
4. What effect size should the researcher report for the main effects and interaction?

5. If the researcher found a statistically significant interaction and wanted to determine if the
difference in cell means were statistically significant, the researcher could conduct follow-up
post hoc analyses by conducting . (Identify the specific
type of analysis.)

Application A
Apply the formulas using hand calculations.
A researcher conducts a 2 × 2 factorial with 80 participants and equal numbers in all four cells.
1. Fill in the missing information in the ANOVA summary table:

Tests of Between-Subjects Effects

Source Sum of Squares df Mean Square F
MSB1 / MSw =
Factor 1 137.813 kB1 – 1 = _____ SSB1 / dfB1 = _____
_______
MSB2 / MSw =
Factor 2 2.113 kB2 – 1 = ______ SSB2 /dfB2 = ______
______
SSB1 × B2 / dfB1 × B2 MSB1 × B2 /MSw =
Interaction 25.313 (dfB1)(dfB2 ) = ________
= __________ _____

Error 197.750 C(n – 1) = _____ SSw / df w = ______

Total 362.988 N – 1 = _______

 Chapter 13 ■ Factorial Designs   139

2. Look at Appendix C.6 to determine whether the F value for the main effects and interaction
meet the criteria for statistical significance at p < .05.

Is the F statistically significant at p < .05?

Main effect for factor 1

Main effect for factor 2

Interaction effect

3. Compute the effect size (η 2partial ) for each main effect and the interaction using the formula
ηpartial
2
= SSB / ( SSB + SSw ) .

η 2partial
Main effect for factor 1

Main effect for factor 2

Interaction effect

Application B
Interpret output from a data analysis program.
A researcher hypothesizes that ADHD symptoms will exacerbate the distracting impact of a cell phone
alert during a lecture. Abbreviated SPSS output appears below:

Descriptive Statistics (Dependent Variable: Exam Score)

Cell Phone Condition ADHD Symptoms Mean Std. Deviation N

No Alert Low Symptoms 6.8400 .85049 25

High Symptoms 6.8000 1.19024 25
Total 6.8200 1.02400 50

Alert Low Symptoms 6.4000 1.19024 25

High Symptoms 5.8800 1.09240 25
Total 6.1400 1.16075 50

Total Low Symptoms 6.6200 1.04764 50

High Symptoms 6.3400 1.22241 50
Total 6.4800 1.14133 100

Test of Between-Subjects Effects (Dependent Variable: Exam Score)

Type III Sum of Mean Partial Eta

Source Squares df Square F Sig. Squared
Corrected Model 14.960 a 3 4.987 4.199 .008 .116
Intercept 4199.040 1 4199.040 3536.034 .000 .974
Cell Phone Condition 11.560 1 11.560 9.735 .002 .092
ADHD Symptoms 1.960 1 1.960 1.651 .202 .017
Condition * ADHD 1.440 1 1.440 1.213 .274 .013
Error 114.000 96 1.187
Total 4328.000 100
Corrected Total 128.960 99
140   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Use these results to fill in the blanks below. Round your answers to two decimal places, except when
important information for the p value would be lost with rounding (e.g., p = .004 or p = .045).

1. The grand mean =

2. The group means for ADHD symptoms: low = ; high =
3. The group means for cell phone condition: no alert = ; alert =
4. The cell means:
a. low ADHD and no alert: ; high ADHD and no alert:
b. low ADHD and alert: ; high ADHD and alert:
5. Fill in the blanks for the inferential statistics:
a. Main effect for ADHD symptoms: F( , )= ,
p=
b. Main effect for cell phone condition: F( , )= ,
p=
c. ADHD × condition interaction: F( , )= ,p=
6. a. Is the main effect for ADHD symptoms statistically significant?
b. Is the main effect for cell phone condition statistically significant?
c. Is the interaction statistically significant?
7. Fill in the blanks for the effect sizes:
a. ADHD symptoms accounted for % of the variance in exam scores (η 2partial =
).
b. Cell phone condition accounted for % of the variance in exam scores
(η 2partial = ).
c. The interaction accounted for % of the variance in exam scores (η 2partial =
).

Exercise 13.5: Beyond the 2 × 2 Independent-Groups Design

Learning Objective: Identify and compare more complex factorial designs and analyses.
Review

1. What type of analysis is appropriate when you have three factors and an interval or ratio
outcome variable and…
a. all factors are independent
b. all factors are dependent
c. two factors are independent and the other is dependent
2. If you have five independent factors and an interval or ratio outcome:
a. what type of ANOVA should you conduct?
b. how many main effects would you test?
c. how many interactions would you test? Recall that the formula for the number of
interactions is:
2k – 1 – k
where k = number of factors
 Chapter 13 ■ Factorial Designs   141

Application
1. A researcher designs a study to predict number of calories consumed per week based on
marital status (married vs. cohabitating vs. single) and home ownership (rent vs. own).
a. What type of design is this? Include the factorial notation in your answer.

b. What type of statistical analysis should the researcher conduct?

c. How many main effects will the researcher analyze?

d. How many interaction effects will the researcher analyze?
2. A researcher designs an experiment to test if providing critical thinking education changes the
amount of time one spends on social media. The researcher randomly assigns participants to
one of two conditions (education vs. no education) and collects information on the amount
of social media time at the start and end of the study (pretest vs. posttest). The researcher
also predicts that age group (under 25 vs. 25–35 vs. 36–55 vs. 56 and older) and type of
community (urban vs. suburban vs. rural) will impact results.
a. What type of design is this? Include the factorial notation in your answer.

b. What type of statistical analysis should the researcher conduct?

c. How many main effects will the researcher analyze?

d. How many interaction effects will the researcher analyze?

Exercise 13.6: The Big Picture: Embracing Complexity

Learning Objective: Formulate a big picture of how to conceptualize and analyze factorial designs.
Review
1. What is “fishing” or “data dredging” and why is it a concern?
_________________________________________________________________________
_________________________________________________________________________
2. What type of error might you be making if you fish for statistically significant results?
_________________________________________________________________________
3. If you find statistically significant results through exploratory analyses, what should you do?
_________________________________________________________________________

Application
Indicate which of these tests you should use to analyze the following designs:
a. chi-square test of independence
b. two-way between-subjects ANOVA
c. two-way within-subjects ANOVA
d. mixed ANOVA

1. A study on the relationship between political orientation (liberal or conservative), marital

status (married, cohabitating, or single), and a rating scale assessing attitudes toward same-sex
marriage:
142   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. A study on the relationship between political orientation (liberal or conservative), marital

status (married, cohabitating, or single), and voting history (voted in last national election or
did not vote):
3. A drug trial in which participants are randomly assigned to take either a sugar pill or an
antidepressant medication. They rate their level of depression at the start of the trial (pretest)
and after taking their assigned pill for six weeks (posttest):
4. A study on change in score (percentage of possible points) between the mid-term and the
final exam for those were matched on GPA and then either assigned a study-partner or not.

Your Research
Develop an interaction hypothesis for a 2 × 2 as gender, age, or personality, will moderate the
experimental or hybrid factorial design. effect.

1. State a hypothesis for a simple experiment on your 3. Based on the option you chose and the research
topic (with one IV, with two levels, and one DV). This article(s), develop an interaction hypothesis that
should be based on past research on your topic. suggests one of the following:
 a. A certain condition (option 1) or participant
 characteristic (option 2) will strengthen or weaken
2. Go back to the research on your topic to decide between the effect you stated in your simple experiment
these two options: hypothesis.

1. Option 1 (experimental factorial): Does research b. The simple experiment hypothesis you stated only
suggest that the IV will impact the DV under all occurs under a certain condition or for certain
conditions? If not, find at least one research participants.
article that suggests that a second IV will c. The direction (e.g., positive or negative) of the
moderate the effect. effect of your IV on your DV depends on a certain
condition or participant characteristic.
2. Option 2 (hybrid factorial): Does research suggest

that the IV will impact the DV for all participants?

If not, find at least one research article that
suggests that a participant characteristic, such 4. Graph your hypothesized interaction.
 14
NONPARAMETRIC STATISTICS

LEARNING OBJECTIVES

• Explain the differences between nonparametric and parametric tests.

• Compute and interpret the chi-square goodness of fit test for nominal data.
• Compute and interpret the chi-square test for independence for nominal data.
• Discuss the appropriate analyses for dependent-groups designs with nominal data.
• Compute and interpret the Spearman’s rank order correlation coefficient for ordinal data.
• Describe the appropriate analyses for independent- and dependent-groups designs with ordinal data.
• Formulate a big picture of how to select and analyze parametric versus non-parametric tests.

CHAPTER SUMMARY
Nonparametric statistics are used with nominal or ordinal data, or when the data violate the assumptions required for
parametric statistics. Chapter 14 outlines the key differences between nonparametric and parametric statistics and then
provides information about when to use the various nonparametric tests.
Because chi-square analyses (nominal data) and Spearman’s rho (ordinal data) are relatively common tests in the social
sciences, the chapter details the computations for these tests and also provides examples of how to interpret these analyses
using a data analysis program. For the other nonparametric tests, the computations are included in Appendix D rather
than in the chapter. The chapter focuses on selecting the appropriate nonparametric test based on the type of data
(nominal vs. ordinal) and design (independent vs. dependent groups).

CHAPTER 14 QUIZ
Test your knowledge of the key terms from the chapter.
1. Nonparametric statistics are used when we have ______________
a. ordinal or nominal data.
b. normally distributed variables.
c. homogeneity of variance.
d. the need for a more powerful test.
14 3
144   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. One of the advantages of nonparametric tests is ______________

a. they are appropriate for small samples with a nonnormal distribution of scores.
b. they help to increase external validity.
c. they can quickly test interactions.
d. they are appropriate for normally distributed data.
3. The chi-square test compares ______________
a. observed observations with obtained observations.
b. sum of the observations with obtained observations.
c. observed frequencies with expected frequencies.
d. predicted observations with the expected frequencies.
4. The chi-square goodness of fit is appropriate when we have ______________
a. ranking data.
b. ordinal data.
c. two nominal variables with two or more categories.
d. one nominal variable with at least two categories.
5. The expected frequencies for each category of a chi-square goodness of fit
a. must equal the total N divided by the number of categories.
b. must be equal for each category.
c. can differ according to the null hypothesis.
d. are irrelevant to the calculation of chi-square.
6. Data for a chi-square test for independence are presented in a ______________.
a. bar graph.
b. contingency table.
c. sampling distribution.
d. frequency distribution.
7. The computation of the expected frequency is more complicated for a test for independence
than for a goodness of fit because ______________
a. we have to consider the relationship of the total number of frequencies to the observed
frequency in each cell.
b. the expected frequency is the sum of the probabilities for each variable represented in the
cell.
c. we have to compute the probability of the frequency for each variable represented in the
cell and multiply these probabilities.
d. the expected frequencies for the second variable are unknown.
8. The specific statistic used to calculate the effect size for a χ2 test for independence
depends on ______________
a. the level of significance for χ2obt.
b. the number of variables in the study.
c. the relative size of the contingency table and the frequency distribution.
d. the size of the contingency table and the relationship of rows and columns.
 Chapter 14 ■ Nonparametric Statistics   145

9. One of the objections to the transformation of interval or ratio data to an ordinal

scale is
a. decreased power.
b. decreased internal validity.
c. too few tied ranks.
d. too many tied ranks.
10. We compute a Spearman rho when we want to correlate
a. two groups with nominal data.
b. two variables of ordinal data.
c. two variables of interval/ratio data.
d. one variable with interval/ratio data and one dichotomous variable.

Quiz 14 Answer Key

1. a. 2. a. 3. c. 4. d. 5. c. 6. b. 7. c. 8. d. 9. d. 10. b.

CHAPTER 14 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 14.1: Parametric and Nonparametric Statistics

Learning Objective: Explain the differences between nonparametric and parametric tests.
Review
1. Statistics that are used to analyze interval or ratio data, and that are assumed to have a
normal distribution and have similar variance among groups, are called
.
2. Statistics that are used to analyze nominal or ordinal data and are used when the assumptions
of parametric statistics are violated are called .
3. Fill in the chart below:

Parametric Nonparametric
Interval or ratio data

Distribution free

Tests for interactions

Small samples or pilot studies

Complicated computations

Violates homogeneity of variance

Powerful
146   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Exercise 14.2: Chi-Square Goodness of Fit

Learning Objective: Compute and interpret the chi-square goodness of fit test for nominal data.
Review

1. A chi-square goodness of fit compares frequencies with

frequencies.
2. The frequencies in each category the researcher counts are the ________ frequencies and are
symbolized by .
3. The count we expect in each category is called the frequency and is
symbolized by .
4. We determine the expected frequency by dividing the by the
. The formula for the expected frequency is E = .
5. When we expect twice as many in one category or group, we add to k.
6. Circle the assumptions for a chi-square goodness of fit test:

Independent categories One variable with two or more categories

Dependent categories Two variables with two or more categories

Normally distributed measure Expected frequency of at least 5 in each

category

Each member in the dataset belongs to only Expected frequency of at least 20 in each
one category category

7. In a chi-square goodness of fit, we determine the degrees of freedom by _________________

_________________________________________.
Application

1. A researcher is studying whether people in different age groups (18–30, 31–45, 46–60, 60+)
were equally likely to vote in the last election. He computes chi-square goodness of fit and
finds χ2(3, N = 400) = 5.35.
a. The null hypothesis for this study is
H0: 
   
b. The alternative hypothesis for this study is
Ha: 

c. In this study df = ______________.
d. What is the minimum number of voters that should be included in the sample?
   
e. How many voters were in the sample?
 Chapter 14 ■ Nonparametric Statistics   147

f. He should have completed a chi-square with (equal/unequal) expected

values.
g. The critical χ2 for p < .05 value = .
h. Did the researcher find statistical significance? How do you know?

2. A researcher wants to know whether residents who use a new park match the
percentage of different ethnicities in the county, which is 20% Black, 60%
Caucasian, 20% Latinx.
a. Should the researcher compute a chi-square goodness of fit test with equal or unequal
frequencies? 
b. In this study, k = because
 .
c. The null hypothesis for this study is
H0: 

d. The alternative hypothesis for this study is
Ha:
 .
e. The df = and the critical χ value =
2
.
f. If the researcher finds χ2(2, N = 125) = 7.33, did she find statistical significance?
____________________________
g. What can you conclude about the use of the park by different ethnicities?
 .
3. Is there a favorite snack among elementary school students? Students select from among three
snacks (healthy, moderately healthy, junk food).

Healthy Moderately Healthy Junk Food

10 20 30

a. State your null hypothesis.

H0: 

b. State an alternative hypothesis.
Ha: 

c. For this study: N = ;k= ;E=

d. Calculate a chi-square goodness of fit test.

e. Can you reject the null hypothesis? Why or why not?

148   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

f. What is the probability of a Type I error? ____________ Type II error? ______________

g. Write a Results section for your findings. Use APA format and include the variable
measured and type of statistical test, observed and expected frequencies, whether the
observed frequencies deviated from the expected frequencies, and results of the chi-square
test.






h. Write a Discussion section for your findings. Include the findings, interpretation/
explanation/implication of the findings, and possible next studies.






4. Dr. Enrique is a psychotherapist and wonders if twice as many of his clients show up late as
on time. He collects data for 2 weeks and finds that 44 clients showed up late and 12 clients
showed up on time.
a. What is the appropriate analysis for these data? Explain.


b. What is k for the study?
c. Compute the chi-square test.
d. Are the results statistically significant? Explain.



e. What would you recommend to Dr. Enrique?



5. Suppose a sport studies researcher decides to test whether birth months of adolescents playing
soccer in the Olympic Development Program in the United States demonstrate the Relative
Age Effect (which states that people born earlier in the year are more likely to be selected for
competitive sports teams). She records the birth months of a sample of 50 adolescents and
divides the birth months into quartiles (1 = Jan–Mar; 2 = Apr–Jun, etc.). She then computes a
chi-square goodness of fit test for equal frequencies and the following output results.
 Chapter 14 ■ Nonparametric Statistics   149

BIRTH QUARTILE

Observed N Expected N Residual

1.00 21 12.5 8.5

2.00 14 12.5 1.5

3.00 10 12.5 –2.5

4.00 5 12.5 –7.5

Total 50

TEST STATISTICS

Birth Quartile

Chi-Square 10.960 a

Df 3

Asymp. Sig. .012

a
. 0 cells (0.0%) have expected frequencies less than 5. The minimum expected cell frequency is 12.5.

a. Did the researcher conduct a chi-square test for equal or unequal frequencies?
____________________________ How do you know?

b. Is the result of the analysis significant? ______________ Explain how you know.

c. What is the probability of a Type I error? ______________ Type II error?

d. Did the researcher find support for the Relative Age Effect? ______ Explain your answer.


e. Write a few sentences that you think the researcher should include in her Discussion section.



f. Name at least one limitation that the researcher could mention and a future study that
could address this limitation.



g. What recommendation might the researcher make to coaches of high school soccer teams
(or principals of high schools)?



150   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Exercise 14.3: Chi-Square Test for Independence

Learning Objective: Compute and interpret the chi-square test for independence for nominal data.
Review and Application

1. The statistic used to analyze the relationship between two nominal variables is
, with the statistical notation .
2. When we examine the relationship between two nominal variables, the table that summarizes
the data is called a .
3. The only difference in the assumptions for a chi-square goodness of fit test and a chi-square
test for independence is .
4. Obtaining the expected frequencies for a chi-square test for independence is more
complicated (than for a chi-square goodness of fit test) because ______________
.
5. The chi-square test for independence is analogous to the parametric test.
6. Answer the following questions using the information in the study: Is developmental status
(adolescence vs. emerging adult) related to unhealthy dieting among females? Tenth-grade
and college senior females who self-report that they are dieting describe their diet practice as
moderate (e.g., exercise, balanced diet, reduced fats), unhealthful (e.g., skipping meals, eating
only one food a day), or dangerous (e.g., fasting, vomiting, using laxatives).
a. These data can be summarized in a × contingency table.
b. The minimum sample size for this study is .
c. The appropriate statistic to analyze these data is .
d. The appropriate effect size statistic for these data is because .
e. The df = .
7. Answer questions a–d using the information in the following question: Is race (African
American/Caucasian) related to having been stopped for a traffic violation in the last month
(yes/no)?
a. This is a × contingency table.
b. The appropriate effect size statistics for this analysis is ______________.
c. The df = ______________
d. Suppose the researcher found χ2 = 5.62. Look at Appendix C.7 in the textbook. Is this
relationship significant at p < .05? ______________. Explain.

8. Answer the following questions according to the output:

IMMIGRANT STATUS * EMPLOYMENT CROSSTABULATION

Employment

Unemployed Part time Full time Total

Immigrant Status No Count 6 4 15 25

Expected Count 5.0 6.0 14.0 25.0

Yes Count 4 8 13 25

Expected Count 5.0 6.0 14.0 25.0

Total Count 10 12 28 50

Expected Count 10.0 12.0 28.0 50.0

 Chapter 14 ■ Nonparametric Statistics   151

CHI-SQUARE TESTS

Value df Asymptotic Significance (2-sided)

Pearson Chi-Square 1.876 a 2 .391

Likelihood Ratio 1.905 2 .386

Linear-by-Linear Association .000 1 1.000

N of Valid Cases 50

a
. 0 cells (0.0%) have expected count less than 5. The minimum expected count is 5.00.

SYMMETRIC MEASURES

Value Approximate Significance

Nominal by Nominal Phi .194 .391

Cramer’s V .194 .391

N of Valid Cases 50

a. These data represent a × contingency table.

b. Name the two variables being correlated. 
c. How many participants were in this study?
d. Is the relationship statistically significant? How do you know?

e. The effect size is (name the exact value). You interpret this as meaning
. The effect size is (strong/moderate/weak).
f. There is a possibility of a Type error.
g. Describe these findings as you might in a Results section.



h. Interpret the meaning of the findings, possible flaws, and next studies as you might in a
Discussion section.




9. Sulloway and Zweigenhaft (2010) examined the relationship between birth order and risk
taking among major league baseball players. They found that first borns were less likely to
steal bases than later borns. One of your classmates examined this relationship by watching
one season of baseball games by one college team and recording the birth order of the players
and whether or not the player stole a base during the season.
152   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

BIRTHORDER * STOLENBASE CROSSTABULATION

Stolenbase

Stole at least
No bases stolen one base Total

Birthorder First borns Count 10 7 17

Expected count 12.6 4.4 17.0

Later borns Count 16 2 18

Expected count 13.4 4.6 18.0

Total Count 26 9 35

Expected count 26.0 9.0 35.0

CHI-SQUARE TESTS

Asymptotic
Significance Exact Sig. Exact Sig.
Value df (two-sided) (two-sided) (one-sided)

Pearson Chi-Square 4.137a 1 .042

Continuity Correctionb 2.713 1 .100

Likelihood Ratio 4.311 1 .038

Fisher’s Exact Test .060 .049

Linear-by-Linear 4.019 1 .045

Association

N of Valid Cases 35

a
2 cells (50.0%) have expected count less than 5. The minimum expected count is 4.37.
b
Computed only for a 2 × 2 table.

SYMMETRIC MEASURES

Approximate
Value Significance
Nominal by Phi –.344 .042
Nominal Cramer’s V .344 .042
Number of Valid Cases 35

a. Did the classmate conduct the correct analysis? Explain.


b. Are the results statistically significant? Explain.

c. How would you calculate the proportion of variance in base stealing accounted for by
birth order?

 Chapter 14 ■ Nonparametric Statistics   153

d. Are the results your classmate found consistent with Sulloway and Zweigenhaft’s (2010)
finding? Explain.



Exercise 14.4: Analyses for Dependent-Groups Designs with

Nominal Data
Learning Objective: Discuss the appropriate analyses for dependent-groups designs with nominal data.
Review

1. What nonparametric test for nominal data corresponds to the dependent-groups ANOVA?

2. The appropriate statistic to compute when you have two related-samples with nominal data is
 .
3. Which nonparametric test for dependent groups requires dichotomous data?

4. Young adults matched on their game playing habits play one of two versions of a popular
video game. Half of the students play a 15-minute updated version that uses avatars and
half of the students see an updated version without avatars. The students note whether they
“would play again,” “would not play again,” or “not sure” if they would play the game again.
What test should be computed to determine whether either version of the game is more likely
to be played? 

Exercise 14.5: Spearman’s Rho

Learning Objective: Compute and interpret the Spearman’s rank order correlation coefficient for ordinal
data.
Review and Application

1. , with the statistical notation is the appropriate

correlation statistic to analyze a relationship between two ordinal variables.
2. Spearman’s rho values can vary from ______________ to ______________, just like the
______________ correlation coefficient, and it is interpreted similarly.
3. As part of a larger study about employment at a small company, a researcher explored the
relationship between employees’ job satisfaction and the number of sick days taken over the
last year. Employees (N = 24) were ranked from most satisfied to least satisfied with their job
and ranked in terms of number of sick days taken. The analysis showed rs = .36.
a. Is the relationship between the ranks significant at p < .05 for a two-tailed test? (Hint: see
Appendix C.8 in the textbook.) ______________ Explain.
_____________________________________________________________________
b. Is the relationship between the ranks significant at p < .05 for a one-tailed test?
______________ Explain.


154   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

4. Answer the following questions given the output.

CORRELATIONS

Language
Prematurity Skills

Spearman’s rho Prematurity Correlation Coefficient 1.000 –.582 *

Sig. (2-tailed) . .023

N 15 15

Language Skills Correlation Coefficient –.582 *

1.000

Sig. (2-tailed) .023 .

N 15 15

*
Correlation is significant at the 0.05 level (2-tailed).

a. What statistic was computed? ____________________________

b. What variables were correlated? ____________________________
c. How many participants were in the study? ____________________________
d. Is the correlation statistically significant? ____________________________
e. What is the probability of a Type I error? ____________ Type II error? _____________
f. Describe the relationship between the two variables as you would in a Results section,
using APA format.



5. Ben wants to know whether the rankings for the 10 teams in his fantasy football league at
the beginning and end of the season are related. At the end of a season, he finds the following
rankings:

Team Beginning of season End of season

A 1 3

B 2 2

C 3 4

D 4 1

E 5 5

F 6 7

G 7 6

H 8 8

a. State the null and alternative hypotheses:

H0: _____________________________________________________________________
Ha: _____________________________________________________________________
 Chapter 14 ■ Nonparametric Statistics   155

b. Compute the appropriate analysis by hand, with a data analysis program, or both.

c. Are the results significant at p < .05? Explain.

d. How should Ben interpret his findings?



Exercise 14.6: Analyses for Independent- and Dependent-Groups Designs

with Ordinal Data
Learning Objective: Describe the appropriate analyses for independent- and dependent-groups designs with
ordinal data.
Review and Application

1. One of the objections to the transformation of interval or ratio data to an ordinal scale is
 .
2. Jesse collects data on the average minutes of daily practice on the guitar. He decides to
convert the data to ranks. If his original data are 80, 50, 50, 45, 40, 35, 30, 25, 20, 15, what
would his ranked data be if they represented the same order as the original data?

3. Deedee is testing the effect of a new advising method on the study time of college students.
Twelve students are matched on their GPAs and one of each pair is advised by the new
method and one by the typical method. At the end of a month, all the students are ranked on
their average daily study time. How would Deedee determine whether there is a difference in
the study times of the two groups? (Hint: What test should she compute?)

4. A musician wonders if songs about breakups tend to rank higher than songs about falling in
love on the Billboard top 10 chart for this year.
a. What inferential statistic should the musician conduct? Explain.

b. If the musician wanted to compare rankings of breakup songs versus falling in love for
the Billboard top 100 over the past 50 years, what inferential statistic is now appropriate?
Explain.

5. You should choose between using a Kruskal-Wallis H test and a Friedman χ2 to analyze your
data based on  .

Exercise 14.7: The Big Picture: Selecting Parametric Versus

Nonparametric Tests
Learning Objective: Formulate a big picture of how to select and analyze parametric versus non-parametric
tests.
Fill in the blanks with the appropriate parametric or nonparametric test
156   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Design Parametric Test Nonparametric Test

Correlation/No groups

Comparison of expected and observed

values
Two independent groups

Two related groups (matching or repeated

measures)
Multiple independent groups

Multiple related groups

Your Research
Select a primary research article on your topic. 

1. Look through the entire Method and Results section 

to identify any nonparametric statistics from this 2. List up to three other inferential statistics that you
chapter. List any that you found and explain how the recognize from other chapters and explain how the
authors used them. (It is possible that the authors did authors used them.
not use nonparametric statistics. If this is the case, try 
searching through other articles on your topic.)



 
 

 
 15
FOCUSING ON THE INDIVIDUAL
Case Studies and Single N Designs

LEARNING OBJECTIVES

• Describe the strengths and limitations of sample-based studies.

• Explain what the case study is and its strengths and limitations.
• Identify qualitative analyses used to evaluate case studies.
• Explain what the single N study is, how to plan and conduct different types of single N designs, and the
strengths and limitations of the different designs.
• Formulate a big picture of how to choose between a sample, case study, and single N design.

CHAPTER SUMMARY
Previous chapters have focused on sample-based studies. The primary goal of sample-based studies is to extrapolate results
to the population from which the sample was drawn, but it is not clear whether the results from such studies will be
relevant or helpful to a specific individual. Chapter 15 outlines two alternatives to sample-based designs: the case study
and the single N design. The case study is typically qualitative and the chapter provides an overview of some of the most
common qualitative analyses used with case studies. The single N design is quantitative, and graphs are essential to their
interpretation. The chapter includes explanation of effective use of each of these designs, along with their strengths and
limitations. The chapter ends with a discussion of when to choose a sample, case study, or single N design.

CHAPTER 15 QUIZ
Test your knowledge of the key terms from the chapter.
1. Which of the following statements about sample-based studies is FALSE?
a. Their primary goal is to extrapolate the results from the sample to the population.
b. They can be used to infer if a relationship found in a sample represents a relationship that exists in the population.
c. They can be used to understand a specific person, event, or organization.
d. They can include a large number of participants.

157
158   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. A researcher collects in-depth information on the medical history of an individual using

interviews and observations. This type of research design is called a
a. descriptive study.
b. single N design.
c. small N design.
d. case study.
3. A researcher collects in-depth information from the managers of a company in order to better
understand the company as a whole. This type of research design is called a(n)
a. case study.
b. embedded case study.
c. single N design.
d. descriptive study.
4. Those who use believe that it is possible to use single cases to develop a theory.
a. small N designs
b. single N designs
c. grounded theory
d. case studies
5. Identify the TWO qualitative analyses that rely heavily on past research or theory.
a. thematic analysis
b. a priori content analysis
c. analytic induction
d. analytic generalization
6. is a quantitative design used to examine a cause and effect within a single
case, and a series of these designs is called a .
a. Single N design; small N design
b. Small N design; single N design
c. Case study; embedded case study
d. Embedded case study; case study
7. Single N designs use in order to rule out alternative explanations for
causality.
a. embedded designs
b. qualitative measures
c. multiple manipulations
d. repeated assessment
8. In a single N design, A = and B =
a. assessment; baseline
b. baseline; manipulation
c. assessment; manipulation
d. manipulation; assessment
 Chapter 15 ■ Focusing on the Individual   159

9. A stable baseline is one that

a. has little variability in scores.
b. has no trends.
c. Both a and b
d. None of the above
10. A multiple-baseline across behaviors designs requires
a. conducting a series of single N designs.
b. conducting a series of case studies.
c. introducing the manipulation at the same time across two or more behaviors.
d. introducing the manipulation at different times across two or more behaviors.

Quiz 15 Answer Key

1. c. 2. d. 3. b. 4. c. 5. b, d. 6. a. 7. d. 8. b. 9. c. 10. d.

CHAPTER 15 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 15.1: Samples Versus Individuals

Learning Objective: Describe the strengths and limitations of sample-based studies.
Application
1. Sample-based evaluations of the Drug Abuse Resistance Education (DARE) program consistently
show that it is ineffective. In other words, these studies suggest that it does not reduce the use of
alcohol and drugs among children and adolescents. Explain why school administrators should take
such results seriously and look for alternative programs with demonstrated effectiveness.




2. After explaining the results of sample-based studies that suggest DARE is ineffective, you hear from
one parent who tells you that DARE helped her child stay away from drugs. Explain how it is possible
that DARE might have helped this child, even though sample-based studies suggest it is ineffective.




3. Provide an alternative explanation for why someone might believe that DARE was effective for a
child, even though the program might not have caused a change in the child’s behavior. (Be sure
to incorporate the criteria for causality into your answer. See Chapter 10 if you need a refresher.)




160   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Exercise 15.2: Case Studies

Learning Objective: Explain what the case study is and its strengths and limitations.
Review

1. Explain the strengths of the case study.



2. One should choose a case for a case study because that case is
a. 
b. 
c. 
3. What are the limitations of the case study?
a. 
b. 
c. 

Exercise 15.3: Qualitative Analyses

Learning Objective: Identify qualitative analyses used to evaluate case studies.
Review

1. List at least two ways that one might one analyze the content of a case based on words and
phrases.


2. Outline the steps in thematic analysis.



3. Outline the steps in a priori content analysis.



4. Explain the similarities between analytic generalization and analytic induction.



5. Explain the differences between analytic generalization and analytic induction.



 Chapter 15 ■ Focusing on the Individual   161

Exercise 15.4: Single N Designs

Learning Objective: Explain what the single N study is, how to plan and conduct different types of single N
designs, and the strengths and limitations of the different designs.
Review

1. Why is the single N considered a quasi-experiment?


2. Why is a stable baseline important in a single N design?

3. A ____________________________ introduces the manipulation at different points across
different persons, settings, or behaviors.
4. Match the single N notation with the description

i. AB1B2 a) A multiple-manipulation return-to-baseline design that includes a baseline

followed by the first intervention, then a return to baseline followed by the
ii. ABA
second intervention
iii. AB1AB2
b) The simplest single N design with one baseline phase and one manipulation phase
iv. AB
c) The simplest reversal design with a baseline followed by a manipulation and
v. ABAB then a return to baseline
d) A reversal design that ends with a manipulation
e) A multiple-manipulation design that includes just one baseline

5. What are the unique strengths of single N designs?

a. 
b. 
c. 
6. Identify at least two limitations of the single N design.



Application A

1. Client A keeps track of how many cigarettes she smokes over a 10-day period and finds the
following results: 18, 20, 19, 20, 21, 18, 19, 20, 21, 20. Is this a stable baseline? Explain.

2. Client B keeps track of how many cigarettes he smokes and finds the following results: 2, 5,
10, 7, 1, 0, 5, 8, 12, 9. Is this a stable baseline? Explain.

3. Client C keeps track of how many cigarettes she smokes and finds the following results: 0, 3,
5, 7, 9, 11, 13, 15, 17, 19. Is this a stable baseline? Explain.

4. Client D keeps track of how many cigarettes he smokes and finds the following results: 22,
20, 20, 16, 17, 15, 16, 16, 14, 12. Is this a stable baseline? Explain.

162   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

5. If you wanted to implement and evaluate an intervention to help these clients reduce the amount
they smoked, which two clients would be most ready to begin the manipulation phase? Explain.



Application B

1. A yoga student consistently meditated 10 minutes each day for 5 days. In order to increase the
meditation time, the student tries to add nature sounds to the routine and is able to meditate
for 14 minutes that day, 13 minutes the next day, and 15 minutes the following 3 days.
a. Graph these results.

b. Based on visual inspection, is there sufficient evidence to say that the intervention was
effective? Explain.



c. The yoga student is considering adding a return-to-baseline phase (ABA design).
i. If the intervention is effective, what would the student expect to see? Graph an
example ABA design (make up data for the second baseline):

ii. If the intervention is not effective, what would the student expect to see? Graph an
example ABA design (make up data for the second baseline):

2. An individual is trying to drink less cola and wonders whether a simple note to himself
saying, “Drink water, not cola!” will help. He conducts a multiple-baseline across settings to
test the effectiveness of the note and finds the following:

Day 1 2 3 4 5 6 7 8 9 10 11 12

Number of sodas Baseline: Intervention:

drunk at home
4 5 4 4 3 2 1 1 2 3 2 1

Number of sodas Baseline: Intervention:

drunk at work
6 6 7 6 7 7 5 6 7 6 7 7
 Chapter 15 ■ Focusing on the Individual   163

a. Graph these data:

b. Based on visual inspection, how would you interpret these results?



3. How would you interpret the following results from a multiple-manipulation design testing
the effectiveness of two interventions designed to decrease procrastination?
Hours between getting home and

8 Baseline Treatment 1 Baseline Treatment 2

7
starting homework

6
5
4
3
2
1
0
1 5 10 15 20
Days




4. If an AB design suggests that an individual spent more time meditating during the
intervention stage than the baseline, would it be ethically appropriate to add a return-to-
baseline phase? Explain.



5. If an AB design suggests that an individual consistently drove while intoxicated during
the baseline but stopped during the intervention, would it be ethically appropriate to add a
return-to-baseline phase? Explain.




Exercise 15.5: The Big Picture: Choosing Between a Sample, Case Study,
or Single N Design
Learning Objective: Formulate a big picture of how to choose between a sample, case study, and single N
design.
164   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application
Identify which would be the best design for the following questions. Explain.

1. To determine whether a drug use treatment program is likely to be effective for a population.

2. To determine whether a drug use treatment program is effective for an individual.

3. To determine the process by which your high school principal decided which programs were
beneficial for students and evaluated the cost/benefit ratio of implementing the program.

4. To understand how the director of the film Moonlight felt about winning the 2017 Academy
Award for Best Picture after La La Land was first mistakenly given the award.

5. To understand how viewers felt when the film Moonlight was awarded the 2017 Academy
Award for Best Picture after La La Land was first mistakenly given the award and whether
reactions varied based on gender, age, or race.

6. To determine whether turning off your Internet access at home increases or decreases the
amount of time you study.


Your Research
Is a Case Study or Single N Design a Good Choice for 
Your Topic? 
1. Does your topic cover a unique phenomenon? Or do 
you have a reason to gain in-depth information about 3. If you chose the case study method, explain whether or
one individual or a handful of individuals? not an embedded case study or using grounded theory
• If your answer is no to either of these questions, would be part of your methodology.
stop here. You should stick with one of the sample- OR, if you chose the single N design, identify the
based studies discussed in previous chapters. specific design (AB, reversal, multiple baseline,
• If your answer is yes to either of these questions, multiple manipulation) you would use and why.
answer questions 2 and 3. 

2. Which is a better choice for your topic, a case study or 
single N design? Explain. 
 
 

 16
HOW TO DECIDE?
Choosing a Research Design and
Selecting the Correct Analysis

LEARNING OBJECTIVES

• Synthesize what you have learned about choosing a research design.

• Synthesize what you have learned about choosing appropriate statistical analyses.
• Formulate a big picture of how to be a critical consumer and producer of research.

CHAPTER SUMMARY
Chapter 16 is a review chapter, which means you can take stock of all that you have learned and focus on some key takeaway
points about research methods and analyses. The first part of the chapter includes information about the decision-making
process involved in choosing a research design. As always, you are reminded to base your study on past research, to consider
internal and external validity, and to determine whether your goal is to describe, predict, or explain. Additionally, you should
consider whether or not you will examine groups, and, if so, you will need to decide the number of groups, whether to
conduct an independent- or dependent-groups design, and whether or not to conduct a factorial design.

The second part of the chapter focuses on selecting the appropriate statistical analyses. The focus is on selecting analyses
based on the type of data collected. The tables and figures in the chapter can help guide your decision-making process
for any current or future research project. Remember that a single research study might contain different designs and
different analyses, and the chapter highlights two examples from the research to illustrate this point.
We hope you take a moment to savor all that you have learned about research methods. We also hope that you will take
these lessons with you as you move on to new challenges and opportunities.

CHAPTER 16 QUIZ
Test your knowledge of the key terms from the chapter.
1. generally have stronger external validity than .
a. Correlational designs; descriptive designs c. Experiments; quasi-experiments
b. Correlational designs; experiments d. Experiments; correlational designs
165
166   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. generally have stronger internal validity than .

a. Descriptive designs; experiments
b. Correlational designs; experiments
c. Quasi-experiments; experiments
d. Experiments; quasi-experiments

3. A researcher wants to examine how beliefs in climate change have changed over time. The
researcher should design a(n) study.
a. descriptive
b. correlational
c. quasi-experimental
d. experimental

4. A researcher wants to know if those who live in urban areas are more likely to believe in
climate change. The researcher should design a(n) study.
a. descriptive
b. correlational
c. quasi-experimental
d. experimental

5. A researcher wants to determine what type of educational materials might increase

participants’ belief in climate change. If the researcher can randomly assign participants to
groups, a(n) study is the best choice of studies.
a. descriptive
b. correlational
c. quasi-experimental
d. experimental

6. An independent-groups design is a better choice than a dependent-groups design when:

a. You have concerns about carry-over effects and you do not have a good matching
variable.
b. You have concerns about within-groups error and you do not have a good matching
variable.
c. You have concerns about within-groups error and participants are scarce.
d. You do not have a good matching variable and participants are scarce.

7. A repeated-measures design is a better choice than a matched-groups design when:

a. You have concerns about carry-over effects and you do not have a good matching variable.
b. You have concerns that repeated exposure to the outcome measure will sensitize
participants.
c. You have concerns about within-groups error and participants are scarce.
d. You are worried that your participants will get fatigued and drop out of the study.
 Chapter 16 ■ How to Decide?   167

8. Paired-samples t tests and independent-samples t tests are both

a. used to compare two groups.
b. used to compare more than two groups.
c. used exclusively with independent-groups designs.
d. only used to analyze experiments or quasi-experiments.
9. A one-way between-samples ANOVA and a one-way within-samples ANOVA are both
a. used to compare two groups.
b. used to compare more than two groups.
c. used exclusively with independent-groups designs.
d. only used to analyze experiments or quasi-experiments.
10. ANOVAs and t tests are all
a. only used to analyze experiments or quasi-experiments.
b. non-parametric statistics.
c. only used when the outcome or dependent variable (DV) is ordinal.
d. only used when the outcome or DV is interval or ratio.

Quiz 16 Answer Key

1. b. 2. d. 3. a. 4. b. 5. d. 6. a. 7. c. 8. a. 9. b. 10. d.

CHAPTER 16 EXERCISES
Achieve the learning objectives of the chapter by reviewing and applying key concepts.

Exercise 16.1: Choosing a Research Design

Learning Objective: Synthesize what you have learned about choosing a research design.
Review
1. Once you choose a topic, you should then __________________________________________
 .
2. Fill in the name of the design:
a. designs are used to describe.
b. designs are used to predict.
c. designs are used to explain.
3. ____________________________ designs may or may not examine groups, but
______________ and ______________ designs always examine groups.
4. If you have a group design, you need to determine whether your groups will be
or .
5. Dependent group designs are ______________ and ______________.
6. Two reasons to choose a factorial design are
a.  _
b.  _
168   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application

1. Using Figure 16.1 in the textbook, decide whether a descriptive, correlational, or experimental
design would be the best way to help answer each of the following questions:
a. How much fast food do Americans eat?

b. Do married individuals eat more or less fast food than single individuals?

c. Has television viewing increased, decreased, or stayed the same over the last 10 years?

d. Can watching comedies on television decrease depression? 

e. Does binge-watching a television series in one night versus watching the series over several
weeks have an effect on anxiety? 
f. Is there a relationship between age and binge-watching television? 

2. For the correlational designs above, identify the one(s) that
a. Will definitely examine groups: 
b. May or may not examine groups: 

3. For the question, “Does binge-watching a television series in one night versus watching the
series over several weeks have an effect on anxiety?” what are some reasons for expanding
from the current two-group design to a multiple-groups design?


4. Look at Figure 16.2 in the textbook to answer the following:
a. For the question, “Can watching comedies on television decrease depression?” explain
why you might choose
i. an independent-groups design


ii. a dependent, matched-groups design


iii. a dependent, repeated-measures design


b. For the question, “Do married individuals eat more fast food than single individuals?”
two out of the three designs in Figure 16.2 are possible choices.
i. Which two are they and why?


 Chapter 16 ■ How to Decide?   169

ii. Explain how you would decide between these two choices.





Exercise 16.2: Selecting Your Statistical Analyses

Learning Objective: Synthesize what you have learned about choosing appropriate statistical analyses.
Review
In the first column below, list all the inferential statistics you have learned in your course. Then fill in
the rest of the table. Try this first without looking at your textbook.

Inferential Statistic When to Use? Effect Size? Post hoc?

170   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Application

1. Use the figures in textbook, as indicated below, to help identify the appropriate statistical
analyses for each of the following questions.
Does the question include just one variable? → see Figure 16.4
Does the question examine a relationship (with no groups) → see Figure 16.5
Does the question compare groups? → see Figure 16.6
a. How much fast food do Americans eat? 

b. Do married individuals eat more or less fast food than single individuals?
i. No matching: 
ii. Individuals matched on socioeconomic status: 

c. Does the amount of fast food eaten by married versus single individuals depend on
whether the individual is male or female?
i. No matching: 
ii. Individuals matched on socioeconomic status: 

d. Does binge-watching a television series in one night versus watching the series over several
weeks have an effect on anxiety?
i. Independent groups design: 
ii. Dependent groups design: 
e. Are there differences in mood after watching a comedy, drama, or horror movie?
i. Independent groups design: 
ii. Dependent groups design: 
2. Identify three to five inferential statistics that you covered in your class. Come up with
research questions that you could test with each statistic. We have provided an example
for you to get you started, but this table is designed to help you review inferential statistics
important in your class. We expect that your professor may have skipped or glossed over some
of the statistics in this textbook.

Inferential Statistic A Question That Can Be Tested

Independent-samples t test Are critical thinking skills better among those who have completed
a research methods course than among those who have not yet
completed such a course?
 Chapter 16 ■ How to Decide?   171

Inferential Statistic A Question That Can Be Tested

Exercise 16.3: The Big Picture: Beyond This Class

Learning Objective: Formulate a big picture of how to be a critical consumer and producer of research.
Application

1. Think of a recent news story that captured your attention.

a. Identify a research question you have based on the story.


b. What concepts or theories have you learned in other courses that are related to this
question?


c. What type of research design might be used to investigate your question?


2. What do you plan to do after you graduate? How might research methods, analyses, or both
be important in this next step? Be as specific as possible.
172   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Your Research
1. After reading past research on your topic, what are the 3. If you are evaluating groups, use Figure 16.2 in the
primary questions that deserve further investigation? textbook to help you decide whether your groups should
 be independent or matched or whether you should use
 repeated measures. Note your decision here.
 
 

Choose one of these that interests you the most, you 4. Does it make sense to do a factorial design? Why or why
believe you could objectively study, and you could not?
carry out given your time and resource limitations. Put 
an asterisk (*) next to the question you chose. 

2. Use Figure 16.1 of the textbook to help determine what 5. What inferential statistic would you use to help answer
type of research design you might use to help answer the the question? Explain.
research question you chose. Note your decision here. 
 

 IBM® SPSS®*
WORKBOOK PART 1
FUNDAMENTALS OF DATA ANALYSIS

BEFORE YOU BEGIN

Know Your Data
You may know the famous maxim by Plato, “know thyself.” In data analysis, we have an equally important, but less
famous, maxim of “know your data.” This is true for any data analysis software program. The program makes data analysis
a whole lot easier than hand calculations, but if you don’t have a good sense of your data you are certainly not going to be
able to make sense of the output. Moreover, no data analysis program can tell you the right analyses to run. You have to
make that decision based on the type of data you have and the questions you want to answer.
Another important saying to remember is “garbage in, garbage out.” You can enter any numbers into a data analysis
program and it will run analyses and provide some nice-looking output. If you don’t have quality data, however, that
output will be garbage. It is therefore important that you spend time carefully designing your study and deciding on
the key variables and how you will measure them. Even the best planned study, however, may result in poor data due to
participant or coder error, carelessness, or sabotage.
Before you enter your data, you should first carefully review your raw data. Raw data are the original measures your
participants completed, the observation check lists your observers filled out, etc.

—— You should omit any questionable data. For example, if it is clear that a participant did not answer questions
correctly or if they seemed to be answering haphazardly, you will want to omit that data. You can choose to
omit part of a participant’s data (one question or one scale) by leaving it blank (“.”) in SPSS. Alternatively,
you may decide that all the data from a participant must be discarded. In either case, be sure that you note
any data that you chose not to enter and why. You will include this in the Method section of your research
report.
—— It is unethical to eliminate any piece of data because it is not what you expected or does not fit your hypothesis.
You should discuss with your professor when deciding to eliminate data that seems to suggest the participant
misunderstood what was being asked or did not take the study seriously.
—— It is OK if a participant left some items blank. Deciding not to answer a question is the right of the participant
and a few missing variables is to be expected in most research studies.
—— Although you can store non-numerical responses in the data file as “string” variables, you cannot use them in
SPSS statistical analyses. You will need to code into numbers any open-ended questions or other non-numerical
responses you plan to analyze with SPSS®. For example, instead of entering “male” as a participant’s gender, you
should code male as a number (0, 1, 2, the number doesn’t matter as long as you are consistent in your coding
and record your codes so that you remember them).
—— It is a good idea to put an ID number on each of the measures or observer sheets and then enter that number
into SPSS. This will provide a link between your raw data and your SPSS data file. An ID is particularly
helpful when you find a data entry error in your SPSS file.

*IBM® SPSS Statistics / SPSS is a registered trademark of International Business Machines Corporation

173
174   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Understand the Basics of SPSS®

SPSS® must be installed on the computer you are using. You will open the program through an icon
on your desktop or by finding it in your computer applications. The program is updated yearly and so
be sure you open the most recent version. This workbook is based on version 27. Later versions may
look slightly different but the key commands likely remain the same. There are also slight differences
in how the program looks on a PC versus a Mac.
There are three major components to the program:

—— Data Editor (.sav)

{{ This is the spreadsheet where you will enter and manage your data and run your analyses.
{{ You can import or export other types of data files (e.g., .csv)
—— Output (.spv)
{{ Output displays the results of analyses.
{{ The output window will pop up after you execute commands from the data editor or you
can open a saved output file and add to it.
—— Syntax Editor (.sps)
{{ The editor allows you to enter and edit statistical commands manually.
{{ Because it is a slightly more advanced skill, we do not include directions for editing syntax
in this workbook. Explore this option on your own if you wish.

DATA ENTRY
Every variable in your dataset must have a unique name. If you are entering data for a class research
project, be sure you use the variable names your professor tells you to use. If this is your own data
set, you can name the variables anything you want as long as there are no spaces and the name begins
with a letter (not a number). Versions prior to SPSS 27 do not allow for symbols in the variable name.
For example, variable names for the questionnaire below might be as follows:

ID, learnexp, afraid, lookforward, love, gender, age

Research Methods Questionnaire

ID#
Rate how much you agree or disagree with each statement:

1. Research Methods is a good Strongly Disagree Disagree Agree Strongly Agree

learning experience.

2. I am afraid of Research Methods. Strongly Disagree Disagree Agree Strongly Agree

3. I look forward to Research Strongly Disagree Disagree Agree Strongly Agree

Methods.

4. I love Research Methods. Strongly Disagree Disagree Agree Strongly Agree

Gender: Male Female Non-binary         Age:

Set up Your Data Set

1. Open SPSS® and close out of the welcome window and you should see a window that looks
like the one below. Go to “Variable View.”
 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 175

Go to “Variable View” by clicking the tab

at the bottom of the screen.

2. List the names of your variables in the “Name” column.

Enter the variable names in the Name column.

• Each variable name must be unique.
• There can be no spaces in the name.
• The name cannot begin with a number.

3. Be sure “Type” is Numeric for all your quantitative variables (numbers).

4. In the “Label” column, you can provide more information about what each variable measures.
Adding a label serves as a reminder for you, but you can skip this step if you have a small data
set and the variable names provide enough of a reminder.

The label allows you to add more detail about each variable.
176   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

5. The “Values” column serves as a reminder for how your nominal, ordinal, and interval
variables are coded. You do not have to enter values, but it is a good idea so that you can easily
access data codes without having to return to your raw data or code sheets.

Click on a cell, then click on ... and the Value Labels box will appear.

Enter the value labels

one at a time, clicking
add after each. Click
OK after you have
entered all the value
labels.

The example above is for the first question on our Research Methods questionnaire
(named “learningexp”). In the image, we had already entered value labels for the values 1 through
3 and are in the process of finishing our last value label.
Value → 1 Value Label → strongly disagree; Add
Value → 2 Value Label → disagree; Add
Value → 3 Value Label → agree; Add
Value → 4 Value Label → strongly agree; Add
When you have finished adding all the value labels, you click OK.
Hints:
—— You do not need value labels for ratio data.
—— Your value labels will depend on how your variables are coded.
{{ For nominal variables (such as gender, in our example), you will decide how to code the
categories. For our example, we decided to code male as 1, female as 2, and nonbinary
gender as 3, but we could have reversed that or used different numbers.
{{ Some interval scales might not have value labels for every score, and you can just enter
the lowest- and/or highest-value labels. For example, if we asked participants to rate how
happy they are on a scale from 1 to 10, with 10 being extremely happy, we might just
enter the value label for the value 10 (extremely happy).
—— If you have multiple variables with the same value labels, you can copy and paste.
 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 177

6. The “Measure” column identifies the scale of measurement for each of your variables. Because
interval and ratio data are treated the same for analysis purposes, these are both identified as a
“scale” in SPSS.

Enter Scores

1. Go to the “Data View” screen.

Click on the “Data View” tab at the bottom of

the screen.

2. Enter your participants’ scores.

Be sure to SAVE your data often! Data files are .sav files.

On the Menu bar, click File → Save

—— At this point, you do not need to save the output file (.spo). But the output will be important
once you have run analyses, and you will want to save it then.
—— After you have named and saved a file once, you can click on the second icon from the left in
the tool bar (the blue square) to save updates to the file.
178   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Check for Data Entry Errors Using SPSS

Even if you entered your data carefully, it is a good idea to check that you did not make data entry
errors. Manual checks, or spot-checks if you have a large data set, are essential. You can also have SPSS
check for obvious errors. To do this, run frequencies to ensure that there are no values outside the
possible range. For example, if there is a value of ‘12’ for gender in our dataset, we know this an error
because gender was coded as 1, 2, or 3. To run frequencies:

On the menu bar, click Analyze → Descriptive Statistics → Frequencies

The “Frequencies” box will open:

1. Select
all the 2. Send them to
variables the “variables” box
(either by clicking the
one at a arrow.
time OR
use shift +
left click).

3. Be sure “Display 4. Click “OK.”

frequency tables” is checked.

An output window will appear, as shown below. Inspect your data to determine if any values in the
output do not make sense (e.g., they are out of the possible range of values based on how your data are
coded).
IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 179
180   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

CREATE SCALE SCORES

Remember that some constructs are operationally defined with multiple items, in which case you will
need to create a scale score. To do so, you will follow these steps:

—— Make sure that all items are coded in the same direction and recode variables if
necessary.
—— Check that the scale has good internal consistency (a type of measurement reliability). Here
we will show you how to calculate Cronbach’s Alpha.
—— Compute a scale score.

(NOTE: If the measure has separate subscales, follow these directions for each subscale.)

Recoding Variables
Make sure that all the items are coded in the same direction. Recode any that are coded in the opposite
direction.
For example, the four items below are part of the “Attitudes Toward Research Methods” measurement
scale.

1. Research Methods is a Strongly Disagree Disagree Agree Strongly Agree

good learning experience

2. I am afraid of Research Strongly Disagree Disagree Agree Strongly Agree

Methods

3. I look forward to Strongly Disagree Disagree Agree Strongly Agree

Research Methods

4. I love Research Methods Strongly Disagree Disagree Agree Strongly Agree

A higher score on questions 1, 3, and 4 indicates a positive attitude.

—— However, a higher score on question 2 indicates a negative attitude.
—— We want a higher score on our “Attitudes Toward Research Methods” scale to indicate
more positive attitudes, and we therefore will need to recode question 2 (named
“afraid” in our example data set) so that it is coded in the same direction as the other
variables.

It is best to use “Recode into Different Variables” so that you can keep track of changes you have made
to your data set.
—— When you recode into a different variable, you create a brand-new variable in your data set.
An easy strategy for naming your recoded variables is to add an “R” to the end of the original
name (in our example, we will recode “afraid” into “afraidR”).
—— After recoding, the new recoded variable will show up at the end of your data set. The
original variable will remain in your data set.
 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 181

How to Recode:
On the Menu Bar, click Transform → Recode into different variables
The “Recode” box will open:

1. Click on the variable

you are recoding.

2. Send over to the input output box by clicking

the arrow (OR you can double click on the
variable name).
182   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

A new box will open:

In our example dataset, the original variable was coded on a 4-point scale. To recode, we need to reverse
the codes so that a higher score indicates more positive attitudes:

1. 1 → 4
2. 2 → 3
3. 3 → 2
4. 4 → 1
Hints:
—— Your old and new values will depend on how your original variable was coded.
—— You need to enter new values for ALL your old values. Even if the old and new values are the
same number, you need to enter it (i.e., 3→3).
—— You can recode multiple variables at the same time as long as the old and new values are the
same for all the variables.
—— Your recoded variable will show up at the end of your dataset. If you like, you can click and
drag it to a different position in your dataset.
—— You can check that you did the recoding correctly by looking at the recoded variable in data
view. Select a few participants and compare the original value to the recoded value.
—— Recoding can be used for other transformations as well. For example, you could convert a
4-point scale to two categories (low vs. high).

Reliability Analysis (Cronbach’s Alpha)

On the Menu bar, click Analyze → Scale → Reliability Analysis

The “Reliability Analysis” box will appear:

 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 183

After you click OK, an output window will appear.

184   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Reliability

Scale: ALL VARIABLES

Case Processing Summary

N %
Cronbach’s alpha is the measure of
Cases Valid 10 100.0 internal consistency and tells us how
reliable the measurement
Excludeda 0 .0 scale is.

Total 10 100.0
The higher the alpha, the better
a. Listwise deletion based on all variables in the procedure. reliability. Ideally, the alpha should be
.70 or above, although an alpha as
low as .60 can be acceptable.
Reliability Statistics

Cronbach's In this example, alpha = .76.

Alpha N of Items

.762 4

Item-Total Statistics

Scale Mean Scale Corrected Cronbach's This table allows you

if Item Variance if Item-Total Alpha if Item to examine each item.
Deleted Item Deleted Correlation Deleted The last column tells
you what your alpha
would be if an item
RM good 7.9000 4.100 .605 .683
were deleted (or
learning
omitted).
experience

look forward 8.5000 3.611 .763 .591

to RM
Here we see that
omitting the “love
love RM 8.1000 4.989 .227 .882
RM’’ question would
Recoded- 8.2000 3.733 .740 .607 raise our alpha to .88.
afraidRM

Scale Statistics

Mean Variance Std. Deviation N of Items

10.9000 6.767 2.60128 4

Making Decisions Based on Reliability Analysis

—— If the alpha is below .70, check first that your data were correctly coded and recoded and
you selected the correct variables (e.g., the new recoded items plus the original items). If
so, check to see if omitting an item will raise the alpha. If you cannot obtain an acceptable
alpha, this means that the items are not consistent with each other and should be analyzed
separately.
—— If the alpha is above .70, your measure has good internal consistency. You can still check
individual items to see if omitting an item will raise the alpha.
—— The decision to omit an item should be based on how much it will help to improve the alpha
and a review of the item itself. You may decide to keep an item if you believe it is important to
your overall construct.
 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 185

Note: For scales developed and validated by other researchers, it is usually best to keep all the
items if possible. For such scales, only omit items in cases where the alpha is below .70 and
omitting items will raise the alpha.

—— In the example output, we have a pretty decent alpha with all four items (.76). However, if
we omit the “love RM” question, the alpha will increase to .88, and this is a big increase.
Plus, this question doesn’t seem to fit as well with the other items because it assesses a general
feeling of love. Consequently, we chose to omit it from further analyses.
—— To omit an item, you don’t actually delete it from your dataset. You simply leave it out when
you compute a scale score in the next step.

Compute a New Variable

To compute a scale score, you will create a new variable that represents the total or average of all the
individual items on a measurement scale. You will use this scale score for future analyses.

On the Menu bar, Click Transform → Compute Variable

The “Compute Variable” box will open:

1.Name the computed variable.

2. To compute a scale score,
Create a numeric expression
(see numeric expression
options).

Send variables over with the

arrow or type directly into the
numeric expression box.

Don’t forget to use any

recoded variables instead of
the originals.

3. Click “OK.”

Numeric Expression Options

You can create computed variables using any numeric function. Following are the most common
numeric functions for creating a scale score:
—— Create a sum of the variables by using the + sign.
{{ Example shown: learnexp + lookforward + afraidR
{{ If a participant has missing data for any of the variables included in the numeric
expression, the computed variable will also be missing.
186   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

—— Create an average of the items by using the mean function.

{{ For example: MEAN(learnexp, lookforward, afraidR)
—— To create a sum if there are missing data, calculate the average and then multiply that by the
number of variables in the numeric expression.
{{ For example: MEAN(learnexp, lookforward, afraidR)*3

MANAGING DATA
IBM® SPSS® can help you create groups, select groups, and select a random sample of cases. We will use
a different data set on a study of self-esteem for the examples in this section.

Creating Groups
You might want to identify a subpopulation or group of women or athletes. This is easily accomplished
because gender and athlete status are both nominal variables (they are already categories). But what if
you have interval or ratio data? If you wanted to examine a subpopulation or group who were under 21,
for example, you could convert that ratio data into categories by creating groups in SPSS.

How to Create Groups Based on a Cutoff Score

You may create groups based on a predetermined cutoff. For example, in the research literature older
adults are often categorized as “young old” (55–64), “middle old” (65–74), and “old old” (75 and older).

On the Menu Bar, click Transform → Recode into Different Variables

1. Click on the variable

you want to split.
2. Send the variable to
the input output box.
3. Name and label the
new “Output” variable.
4. Click “Change.”

5. Click “Old and New

Values.”

6. Enter a range of
scores.
In this example:
55–64 = 1
65–74 = 2
75 through
highest = 3
7. Click “Continue”
and then “OK.”

How to Create Groups Based on a Median Split

If you do not have a cutoff score that makes sense to use, or if you want approximately equal numbers
in your groups, you can create groups using a median split. The median is the middle score in a sample
distribution. There are two steps to a median split:

1. Find the median of the scale.

 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 187

On the Menu Bar, click Analyze → Descriptive Statistics → Frequencies

1. Send the variable you want to split to 2. Click

the “Variable(s)” box. “Statistics.”

The “Statistics” box will open:

3. Click on “Median”
and then “Continue.”

You will return to

the first Frequencies
box, then click “OK.”

After you click OK, the output window will open.

OUTPUT
Frequencies

Statistics
selfesteem The output reports
N the median.
Valid 25
Missing 0 In this example, the
median = 14.
Median 14.0000

selfesteem
Frequency Percent Valid Percent Cumulative Percent
Valid 6.00 1 4.0 4.0 4.0
8.00 2 8.0 8.0 12.0
9.00 1 4.0 4.0 16.0 The frequency
10.00 4 16.0 16.0 32.0 table tells you how
12.00 4 16.0 16.0 48.0 scores are
distributed.
14.00 1 4.0 4.0 52.0
15.00 4 16.0 16.0 68.0 In this example,
52% of the
17.00 1 4.0 4.0 72.0 scores fall at
18.00 1 4.0 4.0 76.0 or below
19.00 1 4.0 4.0 80.0 the median of 14.

20.00 5 20.0 20.0 100.0

Total 25 100.0 100.0
188   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. Recode the variable just as you would when using a cutoff score.
On the Menu Bar, click Transform → Recode into Different Variables

1. Click on the variable you want to split. 2. Send the variable to the input
output box. 3. Name and label the new “Output” variable. 4. Click “Change.”
5. Click Old and New Values.”

Enter a range of scores.

We can use the
“Range, LOWEST
through value” command
to get the low category.

In this example,
Lowest through 14 = 1.

To get the high

category, use the
“Range, value
through HIGHEST”
command.
In this example,
15 through highest = 2.

After you are finished adding values,

click “Continue” and then “OK.”

Split the Data File, Select Groups, or Select a Random Sample

Split File Command
Use this command if you want to run analyses separately for two or more groups (e.g., based on marital
status). The Split File command will also organize your dataset according to the groups.

On the Menu Bar, click Data → Split File

1. Select “Compare groups.”

2. Select the grouping

variable and send it to
“Groups Based on” box.

3. Click “OK”

(All future commands you

do will separate out the
groups, until you turn off
the split file by clicking
“Analyze all cases…”)
 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 189

Select Cases Command

Use this command if you want to analyze data for only certain participants. For example, you may want
to only examine those participants who are younger than 65. Or, you may want to select a random
sample.

On SPSS Menu Bar, click Data → Select Cases

If you want to select cases based on a

condition, click “if condition is satisfied” and
then “if…” A new screen will open where
you will enter your parameters
(e.g., age < 65).

If you want to select a random sample, click

“Random sample of cases” and then
“Sample…” A new screen will open where
you will enter your parameters.

Be sure you click either “Filter out

unselected cases” or “Copy selected
cases…”
Warning: If you click “delete,” you
could lose those cases!

Remember: When you use the Split Data or Select Cases commands, all the subsequent analyses you do
will follow this command until you return to the command screen and click “All cases.”

PART 1 PRACTICE EXERCISES

Practice Exercise 1
A researcher is interested in how much students value academic honesty. Here is the five-question scale
she used to measure the value of academic honesty:

On a scale of 1 to 5, with 5 indicating strongly agree, how much do you agree that

Academic honesty is essential to a good education.

Talking about academic honesty is a waste of class time.

Cheating is sometimes necessary for academic success.

I would think less of a friend if I found out that he or she cheated on an

assignment.

Honesty and integrity in my academic work are strong values for me.

She collected data from 10 people; here is a summary of the data:

190   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Value of Academic Honesty: Question #s and Answers

ID Age Gender #1 #2 #3 #4 #5
1 19 Male 5 2 1 4 4
2 21 Male 3 4 3 1 3
3 23 Female 3 2 1 3 4
4 18 Male 4 2 1 3 5
5 24 Female 4 1 2 4 3
6 50 Female 4 2 1 4 4
7 20 Female 3 3 4 3 1
8 19 Male 4 3 4 4 1
9 28 Male 3 2 2 3 3
10 26 Female 1 4 3 1 3

1. Enter the data using SPSS®. You can make up whatever variable names make sense to you. All
variables should be numeric, so be sure to code gender as a number.
2. Look over the questionnaire. You will want a higher number to indicate stronger values for
academic honesty. Right now, there are some questions that need to be reverse-coded. Using
SPSS®, recode these into different variables so that a score of 1 = 5, 2 = 4, 3 = 3, 4 = 2, 5 = 1).
3. Check the reliability of the scale by calculating Cronbach’s alpha. Remember to use the
RECODED variables.
a. What is the alpha (α)?
b. What decisions should you make based on these results?
_____________________________________________________________________
_____________________________________________________________________
4. Compute the total score on the value of academic honesty questionnaire. Remember to use
the RECODED variables here, not the original ones.
5. Check your answers and make any corrections to the dataset, if necessary.

Practice Exercise 2
1. Go to https://edge.sagepub.com/adams3e
2. Click on Data Sets.
3. Open the file called Reminiscence Study, and view these files:
a. Description of the Reminiscence datasets (reminiscence description.pdf )
b. Reminiscence data set 1 (reminiscence1.sav)
4. Read over the description of the Reminiscence dataset, paying special attention to how
variables are named and coded. It may be helpful to print this file.
5. Note that there are several measurement scales in the survey.
a. There are two subscales from the Savouring Beliefs Inventory (Bryant, 2003):
i. Savouring the Moment (8 items)
ii. Anticipating the Future (8 items)
b. Satisfaction with Life (5 items; Diener et al., 1985)
 IBM® SPSS® Workbook Part 1 ■ Fundamentals of Data Analysis 191

c. Reminiscence Strategies (9 items created for this survey). Note that you can also
create subscales for the strategies, such as cognitive (3 items) or behavioral strategies
(6 items).
6. Download and open the dataset.
7. Choose one of the measurement scales (or subscales):
a. Look at the individual items for the measurement scale or subscale you chose. Do you
need to recode any of the items for the scale you chose? If so, recode the variables in the
data set using SPSS.
b. Using SPSS, calculate Cronbach’s alpha for the scale you chose.
α=
c. Compute the scale score.
8. If you would like more practice, repeat with another of the measurement scales.

Practice Exercise 3
1. Go to https://edge.sagepub.com/adams3e
2. Click on Datasets
Open the file called “Game of Thrones.” Here you will find two files:
{{ Description of the Game of Thrones dataset (.pdf)
{{ The dataset (.sav)
3. Read over the description of the Game of Thrones dataset, paying special attention to how
variables are named and coded. Note that this is real data based on the first season of the
television series.
a. Use simple random sampling to select a sample of 44 of the 129 characters (which will
give you a 12-point confidence interval with a 95% confidence level).
Use the Select Cases command and click “Copy Selected Cases into a New Dataset.” Click on
“Sample” and select a random sample of 44 from the first 129 cases.
b. Go back to the original dataset with all 129 cases. Suppose you want to have a random
sample with about equal numbers of characters who appeared in a few seasons and who
appeared in many seasons. To do this, you’ll need to follow several steps.
i. First, you’ll need to convert the “totalseasons” variable, that is now ratio, into a
nominal variable with two groups. Because you want about equal numbers, use a
median split (see How to Create Groups Based on a Median Split for directions).
ii. Use the Frequency command to calculate the median.
iii. Use the Recode into Different Variables command to create a new variable, coded
as 1 (scores at or below the median) and 2 (scores above the median). Name it
whatever you like.
Suppose you want a stratified random sample of 44 that represents the distribution in the
population. Look at the frequency table. Notice that 75 scores in the population (58.1%)
fall at or below the median. In a stratified random sample of 44, you will therefore want
about 26 out of 44 (about 58%) to be from the low category and the remaining 18 of the 44
to be from the high category. Next, we will walk you through the steps to randomly select
this sample.
c. Use the Split File command to organize your dataset by your new variable. Cases coded
as 1 (low number of seasons) should all be listed first in the dataset. Then use the Select
192   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Cases command to select a random sample of 26 from the first 75 cases. Click on “Copy
Cases into a New Dataset.” This new dataset will represent your sample from the low
number of seasons.
d. Go back to the original dataset. Use the sort command (Data → Sort Cases) and sort
cases by your new variable. Click “Descending” so that now those coded as 2 will appear
first in your dataset. Use the Select Cases command to select a random sample of 18
from the first 54 cases. Click on “Copy Cases into a New Dataset.” This new dataset will
represent your sample from the high number of seasons.

Extra Practice
The following datasets are available for download. Choose one or more and practice some of the SPSS
skills you have learned.

Dataverse Network: http://thedata.harvard.edu/dvn

This is a large repository of data created by the Institute for Quantitative Social Science at Harvard
University (IQSS). Many datasets are available for immediate download, and others are available with
permission.

General Social Survey (GSS): http://gss.norc.org

Data and codebook available for the General Social Survey that is used to assess demographics and
attitudes of U.S. residents.

Health Information National Trends Survey (HINTS): https://hints.cancer.gov

Data and codebook available for the HINTS program, a study designed to monitor how adults use the
Internet and other communication to learn about health issues.

Monitoring the Future: http://www.icpsr.umich.edu/icpsrweb/ICPSR/studies/20022

Data and codebook available from Monitoring the Future, an ongoing study of the behaviors and
attitudes of American youth.

Pew Research Center: http://www.pewresearch.org/data/

Data available from a variety of polls.
 IBM® SPSS®
WORKBOOK PART 2
ANALYSES TO EXAMINE ONE
VARIABLE AT A TIME

Decision Tree When Analyzing One Variable

Describe Descriptive
Confidence
Statistics
Is the goal to Interval
describe the sample chi-square
OR make inferences Make No goodness
Are you
to the population? inferences nominal of fit
comparing What is the
scores to a scale of
known or Yes measurement
expected for the interval one-sample
value? variable? or ratio t test

DESCRIPTIVE STATISTICS
Nominal Variables
Calculate the frequency to describe nominal data such as gender, marital status, religious affiliation, etc.

On the Menu Bar, click Analyze →Descriptive Statistics → Frequencies

The Frequencies Box will open:

1. Click
on the
variable(s)
you want to
analyze in
the variable
list and 2. Be sure that “Display frequency
send to tables” is checked.
“Variable(s).”

3. Click “OK.”

193
194   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Then you will see the output window:

FREQUENCIES
Statistics
Participant’s Gender
N Valid 10
Missing 0

Participant’s Gender

Valid Frequency Percent Valid Percent Cumulative

Percent
Male 5 50.0 50.0 50.0
Female 5 50.0 50.0 100.0
Total 10 100.0 100.0

This output tells us

• How many total participants there are (N = 10)

• How many of these participants were male (n = 5) and female (n = 5)

• The percentage of the sample in each group or category (50% of the sample was male
and 50% female)

Interval or Ratio Variables

Calculating the central tendency, variability, and skewness of interval or ratio variables Option 1: Use the
Frequency command

On the Menu Bar, click Analyze → Descriptive Statistics → Frequencies

1. Send the variable(s) you want to

analyze to “Variable(s).”

2. Click “Statistics.”
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 195

The “Statistics” box will open:

3. Click on the
Statistics you want
to run.
• Select more options
than you will report,
including the Mean,
Median, SD, Min,
Max, and skewness.
• Use the skewness
statistic to decide
which measure of
central tendency
and variability
to report.
Click “Continue”
when you are done.

You can also request a chart of the distribution:

4. Click “Charts.”

The Charts Box will open:

5. Click
“Histograms” 6. Click “OK”
and “Show when you
normal are finished
curve…” selecting
your statistics
Then click and charts.
“Continue.”

OUTPUT
Frequencies

Statistics
In the output, you will first see the statistics you
participant's age requested.
N Valid 10 Notice here the skewness statistic (G1) is 2.22.
Missing 0 Recall from the textbook that there are two ways to
Mean 29.6000 determine skewness:
Median 28.5000 1. G1 is greater than +/−2.
Std. Deviation 10.89546 2. The absolute value of G1 is greater than twice the
standard error of the skew.
Skewness 2.224
By both these criteria, age is skewed in this sample.
Std. Error of Skewness .687
Therefore, you should not report the mean and
Minimum 19.00 standard deviation. Instead report the median and the
range (or the min and max).
Maximum 58.00
196   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

PARTICIPANT’S AGE
Valid Cumulative
Valid Frequency Percent Percent Percent The frequency table
outlines the frequency
19.00 1 10.0 10.0 10.0 for every age reported.
21.00 1 10.0 10.0 20.0
Looking at the cumu-
23.00 1 10.0 10.0 30.0 lative percent can be
25.00 1 10.0 10.0 40.0 useful with skewed
data:
28.00 1 10.0 10.0 50.0
29.00 1 10.0 10.0 60.0 “The data were posi-
tively skewed due to
30.00 1 10.0 10.0 70.0
one participant who
31.00 1 10.0 10.0 80.0 was 58. The remaining
32.00 1 10.0 10.0 90.0 ninety percent of the
sample were between
58.00 1 10.0 10.0 100.0
the ages of 19 and 32.”
Total 10 100.0 100.0

Note that the histogram can give you an idea of what the data look like, and it might suggest a positive
or negative skew. It can also alert you to a bimodal or uniform distribution that will not be detected by
the skewness statistic (G1).

If the histogram appears skewed, base the final decision about whether or not the distribution is skewed
on the skewness statistic (G1), not the graph.
REMEMBER, there are two ways to assess skewness:

1. If the skewness statistic (G1) is between -2 and 2, the distribution meets the criteria for a
normal distribution. If G1 is greater than +/-2, the distribution is skewed.
OR
2. If the absolute value of the skewness statistic (G1) is more than twice the standard error of the
skew (SES), the distribution is skewed.

If you have a skewed distribution, the sign of G1 (+ or -) will tell you the direction of the skew (positive
or negative).

Calculating the central tendency, variability, and skewness for interval or

ratio variables
Option 2: Use the Descriptives command

On the Menu Bar, click Analyze → Descriptive Statistics → Descriptives

The “Descriptives” box will open:

1. Send the variable(s) you want
to analyze to “variable(s).”

2. Click “Options.”

Note that if
you want to
convert
data to z
scores,
click this
box.
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 197

The Options Box will open:

3. Click on the
statistics you
want to run.
Note you do not
have the option to
calculate the
median.

Click “Continue.”

and then “OK.”

OUTPUT
Descriptive Statistics
N Minimum Maximum Mean Std. Skewness Std. Error
Statistic Statistic Statistic Statistic Deviation Statistic
Statistic
Total RM 10 4.00 11.00 8.1000 2.23358 −.754 .687
attitudes
Valid N 10
(listwise)

Notice that many of the same types of statistics are obtained from the “Descriptives” command
as from the “Frequencies” command.
A key benefit of using the “Descriptives” command is that you can easily convert data to
standardized z scores.
A downside is that it does not provide the median, frequency table, or charts that are
particularly useful if you have skewed data.

Descriptive Statistics Using the Select Cases or Split File Command

Select Cases
Suppose you want to analyze data for only certain participants. For example, recall that in the example
about attitudes toward research methods, age was skewed because of one participant who was 58. You
might want to see what happens when you exclude that participant.

1. Use the Select Cases command.

On Menu Bar, Click Data → Select Cases
1. Select “If condition is satisfied”
2. Click “If”
3. A new screen will open – type in the condition (age < 58 in this example) and click
“Continue”
4. Click “OK”
(See Part 1 in the SPSS Workbook, “Split the Data File, Select Groups, or Select a Random
Sample,” on page 189 for additional information about “Select Cases.”)
198   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

2. Run descriptive statistics for the variable of interest (attitudes toward research methods, in
this example).
a. (See earlier directions for this analysis.)
b. The analyses will exclude anyone who did not meet the criteria. In this example, we have
descriptive statistics for the nine participants who were under 58:

OUTPUT
Descriptive Statistics
Std. Skewness
Minimum Maximum Mean Deviation Std.
N Statistic Statistic Statistic Statistic Statistic Statistic Error
Total RM 9 4.00 11.00 8.2222 2.33333 –.951 .717
attitudes
Valid N 9
(listwise)

3. If you want to examine the entire sample again, you must turn off the “Select Cases”
command:

Click Data → Select Cases, select “All cases” and then “OK.”

Split File
Suppose you want to compare the scores of two or more groups in your sample. For example, you might
want to compare the mean attitudes toward research methods scores for males and females.

1. The first step is to use the Split File command.

On the SPSS Menu Bar, Click Data→Split File
1. Select “Compare Groups”
2. Select the grouping variable from the variable list (gender, in this example) and send it to
the “Groups based on” box
3. Click “OK”
2. Run descriptive statistics for the variable of interest (attitudes toward research methods, in
this example).
Because you are in “Split File” mode, the output will organize the descriptive information by
groups. In this example, the data are split by gender.

OUTPUT
Descriptive Statistics
Std. Skewness
Deviation
Statistic
participant’s N Minimum Maximum Mean Std.
gender Statistic Statistic Statistic Statistic Statistic Error
male Total RM 5 4.00 11.00 8.4000 2.70185 –1.339 .913
attitudes
Valid N 5
(listwise)
female Total RM 5 5.00 10.00 7.8000 1.92354 –.590 .913
attitudes
Valid N 5
(listwise)
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 199

3. If you want to examine the entire sample again, you must turn off the “Split File” command:

Click Data → Split File, select “Analyze all cases” and then “OK.”

Writing up Results
General Guidelines

—— The goal in a Results section is to summarize the statistics that you believe are most
important and useful in understanding your study. You should be selective in the results you
report and how you report them. Do not just provide a laundry list of data that overwhelms
the reader!
—— Good writing skills are just as important in a Results section as in any other writing
assignment. For example, you should use correct grammar and sentence structure, vary your
sentence length and structure, avoid passive voice when possible, avoid awkward statements,
be concise, and so on.
—— In a Results section, you will be using statistical notation that is likely new to you. Use the
examples in this workbook as models but modify the wording according to your study and
data. Think about what your data and results mean and then consider the best way to explain
your results to your reader.
—— When reporting numbers,
{{ Use words for numbers that start a sentence (e.g., Seventy participants took surveys) or are
less than 10 (e.g., There were three groups).
{{ Use digits when the number doesn’t start the sentence, when stating the number of
participants, when reporting numbers 10 and over, or when reporting numbers in a series
in which at least one number is 10 or above. For example, “There were 47 children who
ranged in age from 7 to 15 (M = 10.60, SD = 2.56).”
—— Round most statistics to two decimal places (see section 6.36 of the APA Publication Manual
for more information).
—— Only use a leading zero (a zero before the decimal point) when it is possible for the statistic
to be greater than one but the result is less (e.g., report M = .80 if the highest possible mean
is one or less, but report M = 0.80 if the mean could have been higher than one).
—— If you find that the text is getting too wordy or complicated, consider creating a table or
graph in Word to help report your results. You can explain what the results mean in the text
of your report, and then refer to the table for the specific statistics. (Note that your table or
graph should follow APA format−you should never cut and paste SPSS output into a research
report!)

Writing up Descriptive Statistics

Reporting Frequency Results for Nominal Data
Here are several options:

1. Option 1: A total of 10 (5 male, 5 female) participants completed surveys.

2. Option 2: A total of 10 participants completed surveys. Fifty percent were male and 50%
were female.
3. Option 3: Five males and five females completed questionnaires.
4. Option 4: Our sample was evenly split between males (n = 5) and females (n = 5).
200   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Reporting Descriptive Statistics for Normally Distributed Data

Here are a few examples:
The Research Methods Attitude Scale has a possible range of scores of 3–12. Observed scores ranged
from 4 to 11 and were normally distributed (M = 8.10, SD = 2.23).
OR
The Research Methods Attitude Scale has a possible range of scores of 3–12. In this sample, the mean
was 8.10 (SD = 2.23).
Note that the second example here is acceptable, but it is less common to report means in the text
(usually both the mean and standard deviation are reported within parentheses as shown in the first
example). It is not acceptable to report the standard deviation in the text.

Reporting Descriptive Statistics for Skewed Data

Below is an example:
Ages ranged from 19 to 58, with a median of 28.50. Data were skewed due to one participant who was
58. Ninety percent of the sample was between the ages of 19 and 32.

Putting it All Together for the Method, Results, and Discussion

We have used the example of the attitudes toward research methods study in Part 1 and 2 of the SPSS
Workbook. Let’s put it all together in sections of a research report.
Some important notes about this example:
—— The numbers reported in this example are pulled from this section and previous sections.
Some details were added here for demonstration purposes.
—— This example is not designed to be a template into which you can simply insert your results. Instead,
it is designed to give you one example of how to integrate analyses and make sense of them.
—— At this point, we have not covered statistical significance testing. You will be introduced to
this important concept in Chapter 7 of the textbook.
—— In a full research report, you would have an Introduction that builds a rationale for your
study based on past research. In the Discussion section, you would also explain how your
results fit or don’t fit with past research. This example does not include those components.

Method
Participants
   Ten students (five male and five female) enrolled in a psychology research methods course
agreed to participate in the study. The students ranged in age from 19 to 58, with a median
of 28.50. The inclusion of the 58-year-old student (who was female) skewed the age of the
sample. All the other students were between the ages of 19 and 32.
Procedure
   On the first day of class during the fall 2016 semester, students enrolled in the psychology
research methods course (N = 22) were asked to volunteer to take a brief online questionnaire
about their attitudes toward the course. Students were asked to complete the questionnaire
right after class and to wait to do so until the instructor left the room. The students were
informed that their answers would be anonymous and confidential and that they could with-
draw from the study at any time without penalty. Because the questions were associated with
a course they were taking, students were told that their participation or lack of participation
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 201

would not impact their course grade and that the professor would not know which students
opted to participate and which students opted out. The students had to acknowledge that they
read and understood this before completing the questionnaire.
Materials
  The questionnaire consisted of four items assessing the students’ attitudes toward
research methods (“Research Methods is a good learning experience,” “I am afraid of Research
Methods” [recoded], “I look forward to Research Methods,” and “I love Research Methods”).
The instructor created the items based on feedback and comments from previous classes.
Each item was rated on a 4-point Likert scale with 1 = strongly disagree and 4 = strongly agree.
Cronbach’s alpha was calculated on the four items and results indicated good internal consis-
tency (α = .76). However, the reliability analysis indicated that omitting the last question (“I love
Research Methods”) would raise the alpha to .88. Because this was a substantial increase, and
because the item was relatively vague, it was omitted from analyses. A total scale score was
calculated by summing the three remaining items.
Results
   The total of three attitude questions made up the Research Methods Attitude Scale and
scores could range from 3 to 12. A score of 7.50 designates the middle and would indicate a
neutral attitude. A score over 7.50 indicates a positive attitude. In this sample, scores ranged
from 4 to 11 (M = 8.10, SD = 2.23).
   Because the age in our sample was positively skewed due to one woman who was 58, we
ran analyses excluding this participant. The mean score increased just slightly with this exclu-
sion (M = 8.44, SD = 2.51).
   Analyses were conducted to compare male and female students. Because excluding the
58-year-old female did not seem to have a large impact on the data, we ran these analyses
with all 10 participants. Males in our sample reported more positive attitudes toward Research
Methods (M = 8.80, SD = 2.95) than females in our sample (M = 7.80, SD = 1.92).
Discussion
   Results of this study suggest that the students who took the survey had slightly more posi-
tive than negative views toward the course at the beginning of the semester. This increased
slightly when the 58-year-old participant was excluded from analyses. Male students scored in
the positive range, whereas female students had less positive and closer to a neutral attitude
toward the course.
   It is important to note that statistical significance tests were not run, so these results can
only be used to describe the sample of 10 students. Moreover, although males in the sample
tended to have more positive attitudes than females, we cannot say that this is a statistically
significant difference.
   Less than half of the class opted to take the survey, and it is unclear what motivated some to
participate and others not to and, more important, whether those who participated were very
different from those who did not. Future studies should be done to attempt to obtain a larger
and representative sample of Research Methods students. It would also be interesting to know
how attitudes toward the course change throughout the semester and beyond, and whether
attitudes predict student performance in the class.
202   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Descriptive Statistics Practice Exercises

Practice Exercise 1
Read the example Method, Results, and Discussion sections and answer the following questions to
review and integrate what you have learned.

1. What evidence is presented that supports the reliability of the Research Methods Attitude
Scale?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
2. What evidence is presented that supports the construct validity of the Research Methods
Attitudes Scale?
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________
3. Rewrite one of the paragraphs in the Results section, in your own words.
______________________________________________________________________
______________________________________________________________________
______________________________________________________________________

Practice Exercise 2
Calculating frequencies is also a good way to check for data entry errors.

1. Go to https://edge.sagepub.com/adams3e
2. Click on “Datasets,” open the file called “Reminiscence Study,” and then open the
“reminiscence with errors.sav” data file. This is real data from a study conducted as part of a
class project. However, this file has errors added to the demographic data for the purpose of
this practice exercise.
a. Using SPSS calculate frequencies for the variables named age, gender, and collegestudent.
3. Look at the frequency tables in the output to identify the two errors.
a. What is one of the errors, and how do you know it is an error?
_______________________________________________________________________
b. What is the other error, and how do you know it is an error?
_______________________________________________________________________
4. Now, you might be wondering what to do about errors such as these. If you have access to
the raw data, you could identify the cases (by ID number) and examine the original data to
identify the correct values. It would be a good idea to also do some spot checks to be sure
there were no other data entry errors.
a. In this example, you do not have access to the raw data. All you can do in such a situation is
delete the erroneous value. First sort the data (click on data → sort cases and sort by the variable
with the error), then scan the variable column to find the error, and then delete the error.
b. Delete the two errors you identified and then calculate the appropriate statistics to
describe age, and re-run the frequency of the variable collegestudent.
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 203

c. Write a brief description of the age, gender, and college student distribution in the sample
as you would in a Participants section of a research report. Be sure you only report the
correct information (without errors) and use APA format.
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________

Practice Exercise 3
1. Go to https://edge.sagepub.com/adams3e
2. Click on “Datasets” and open the file called “Academic Honesty Study.” This is real data from
a study conducted as part of a class project.
a. Review the file called “academic honesty survey codesheet.pdf ” to learn about the
variables and codes.
b. Open the data file called “academic honesty.sav.”
3. Using SPSS, calculate the appropriate descriptive statistics to answer these questions:
a. Do most of the participants believe that those who plagiarize should have the opportunity
to redo the assignment? (Hint: Calculate descriptive statistics using the variable called
“responseC” in Data View and “response/redo” in the Frequencies screen)
_______________________________________________________________________
_______________________________________________________________________
b. Do those who report they have never plagiarized believe that those who do should
have the opportunity to redo the assignment? (Hint: Use the “Select if “command to
select only those who reported that they have never plagiarized, and then run the same
descriptive statistics you did for question 3a.)
_______________________________________________________________________
_______________________________________________________________________
c. Do those who identified as undergraduates differ from those who were not currently
undergraduates in their belief that those who plagiarize should have the opportunity
to redo the assignment? (Hint: Turn off the “Select if” command, then split the file to
compare those who are currently in college to those who are not, and then run the same
descriptive statistics as you did above.)
_______________________________________________________________________
_______________________________________________________________________
d. Write up the results of your analyses as you would in a Results section of a research report.
Remember not to refer to SPSS variable names (such as “responseC”) and use APA format.
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
_______________________________________________________________________
204   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

e. For additional practice, run additional descriptive statistics on the practice data set.
Consider what questions you can ask based on the survey questions, and determine what
analyses to run in SPSS to help answer those questions.
f. You might also practice with one of the data sets available for download, listed at the end
of the SPSS Workbook Part 1.

CONFIDENCE INTERVALS
Data Entry
You can use SPSS® to calculate confidence intervals for means by running an analysis called a one-sample
t test. In order to calculate the confidence interval for a mean the variable must be measured on an
interval or ratio scale. In the example below, we have 20 students and their final exam scores (out of
100), which is a ratio variable.
You can then calculate descriptive statistics and verify that the distribution meets the criteria for a
normal curve, using either the Frequency or Descriptives command. The following analysis will repeat
some of that information.

Calculating the Confidence Interval for the Mean

On the Menu Bar, click Analyze → Compare Means → One-sample t test

 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 205

2. To calculate a
confidence interval, leave
1. Send the variable from
the test value as 0.
the sample to “Test
Variable(s).” (Note: the default is a 95%
confidence interval, to
change this, click “Options.”)

3. Click “OK.”

OUTPUT

One-Sample Statistics
Std. Std. Error The first table provides
N Mean Deviation Mean descriptive statistics for the
sample.
finalexam 20 79.8500 10.17880 2.27605

One-Sample Test
Test Value = 0
95% Confidence Interval
of the Difference
t df Sig. (2-tailed) Mean Difference Lower Upper
finalexam 35.083 19 .000 79.85000 75.0862 84.6138

When you are interested in the confidence interval for the mean, focus on the last
two columns of the second table and ignore the rest.

Writing up Results
When reporting a confidence interval for a mean, include the following:
—— The mean and standard deviation (M, SD)
—— What confidence level you are using (e.g., 95%, 99%)
—— The confidence interval, typically but not always, is formatted as CI [lower, upper]

You may also opt to include other descriptive information, such as the sample size or minimum and
maximum scores or other information you gained from conducting descriptive statistics.

Examples:

Twenty students completed a final exam. The mean score was 79.85 (SD = 10.18), 95% CI
[75.09, 84.61].

OR

Student scores on the final exam ranged from 64 to 98 out of 100 possible (M = 79.85, SD =
10.18), 95% CI [75.09, 84.61].

OR

Final exam scores for 20 students indicated an average score of C+/B- (M = 79.85, SD = 10.18).
At a 95% confidence level, we estimate that the population mean falls between 75.09 and 84.61.
206   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Confidence Intervals Practice Exercise

One evening, two waiters record what percentage of the bill their first 10 customers left as tips:

Waiter 1 Waiter 2
.15 .18
.16 .12
.15 .25
.12 .30
.18 .15
.22 .10
.15 .20
.25 .18
.20 .22
.14 .15

1. Review: Enter the data into SPSS.

2. Review: Using SPSS, calculate the appropriate descriptive statistics to describe the tip
percentage for the entire sample. Write the statistics below.
_____________________________________________________________________
3. Review: Using SPSS, calculate the appropriate descriptive statistics to compare the two
waiters’ tip percentages. Write the statistics below.
_____________________________________________________________________
_____________________________________________________________________
4. Practice with confidence intervals: Using SPSS, calculate the confidence interval for the entire
sample mean and for each waiter separately. Write the intervals below.
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
5. Write up your results using correct APA format. Include a brief interpretation of the results
based on the confidence interval.
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
6. What is the practical significance of the results? What other information would be useful in
understanding the practical significance?
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 207

One-Sample t Test
A one-sample t test compares a mean in the sample to a known or expected mean from the population
or another source. The data must be interval or ratio. For example, suppose a professor wants to compare
their current students’ exam scores to the average from past classes, which they know is 75%.

On the Menu Bar, click Analyze → Compare Means → One-sample t test

1.Send the variable

from the sample
to the “Test Variable(s)
box. 2. Enter the known
population mean 3. Be sure that
(mean of the “estimate effect sizes”
comparison group). is checked.

4. Click “OK.”

OUTPUT
T Test

One-Sample Statistics
N Mean Std. Deviation Std. Error Mean The first table provides
descriptive statistics for the
Finalexam 20 79.8500 10.17880 2.27605
sample.

One-Sample Test
Test Value = 75
t df Sig. (2-tailed) Mean Difference 95% Confidence Interval of the Difference
Lower Upper
finalexam 2.131 19 .046 4.85000 .0862 9.6138

One-Sample Effect Sizes

Standardizer Point Estimate 95% Confidence Interval
Lower Upper

finalexam Cohen’s d 10.17880 .476 .008 .934

Hedges’ 10.60390 .457 .007 .897

correction

The second table provides your t statistic, df, and p value. The third provides estimates of the Cohen’s d effect size as well as
confidence interval for d.
In this example, our t is statistically significant at p < .05 and at minimum we would report:
t(19) = 2.13, p = .046, d = 0.48.
208   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Writing up Results

When reporting the results of a one-sample t test, include the following:

—— The mean and standard deviation (M, SD) from your sample
—— The known or expected mean to which you are comparing your sample
—— Whether or not there was a statistically significant difference between your sample and the
comparison
—— The results of the one-sample t test and effect size calculation, formatted as in our example:
t(19) = 2.13, p = .046, d = 0.48.

Include:
{{ the degrees of freedom in parentheses
{{ t value rounded to two decimals, with a leading zero if t is less than 1.
{{ the p value rounded to two or three decimals
—— Do not round to two decimal places when doing so changes the interpretation of
statistically significance (e.g., do not round .046) or changes the p to zero (e.g., do not
round .002).
—— If the p value at three decimal places is zero, report p < .001.
{{ Cohen’s d rounded to two decimals, with a leading zero if d is less than 1.
—— You might also include confidence intervals. Check with your instructor.

One-Sample t Test Practice Exercise

You have the individual percentage of driving trips during which your classmates reported that they
used their cell phone. Their data are below.

.25 .10 .60 .75 .29

.33 .50 .85 .33 .30
.50 .20 .20 .80 .65
.25 .45 .50 .65 .40
.20 .10 .60 .75 .50

1. Enter the data and compute the appropriate test to respond to your belief that your classmates
will use their cell phones significantly less than the national average of 60% reported by the
National Transportation Board .
2. Can you support your hypothesis? Why or why not?
____________________________________________________________________
____________________________________________________________________
3. What is the probability of a Type I error? Type II error?
4. What is the effect size?
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 209

5. Discuss the practical significance of your findings.

____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
6. Write a Results section, including all information and in the format required by APA.
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
7. Write a Discussion section, including all elements required by APA format.
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
210   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

CHI-SQUARE GOODNESS OF FIT

A chi-square goodness of fit compares frequencies in the sample to a known or expected mean from the
population or another source. The data must be nominal.
On the Menu Bar, click

Analyze → Nonparametric Test → Legacy Dialog → Chi-Square

2. Enter your expected

1. Send the variable to (comparison) values. “All
“Test Variable List.” categories equal” is the
default.

If you wish to enter specific

proportions, click on
“Values” and then enter the
comparison percentages in
the same order as your
variable is coded.

3. Click “OK.”

OUTPUT
Chi-Square Test
Frequencies
Cointoss
Observed N Expected N Residual
Heads 18 15.0 3.0 The first table provides the
observed and expected values
Tails 12 15.0 –3.0 and total N.
Total 30

TEST STATISTICS
Cointoss
The second table provides the χ2, df, and p value.
Chi-Square 1.200 a
In this example, the χ2 is not statistically significant and we
Df 1
would report χ2(1, N = 30) = 1.20, p = .27.
Asymp. Sig. .273

Writing up Results
When reporting the results of a chi-square goodness of fit, include the following:
—— The frequency or percentage from your sample and the comparison value.
—— Whether or not results were statistically significant.
—— The results of the chi-square goodness of fit test formatted as in our example: χ2 (1, N = 30) =
1.20, p = .27.
a. Include:
{{ the degrees of freedom and sample size in parentheses
{{ the χ2 value rounded to two decimals, with a leading zero if the value is less than 1
{{ the p value rounded to two or three decimals as needed for interpretation (note in our
example we rounded to two decimal places).
 IBM® SPSS® Workbook Part 2 ■ Analyses to Examine One Variable at a Time 211

Chi-Square Goodness of Fit Practice Exercise

Go to https://edge.sagepub.com/adams3e. Click on “Datasets” and open the file called “Game of
Thrones.” Here you will find two files:
—— Description of the Game of Thrones dataset (.pdf)
—— The dataset (.sav)
Read over the description of the Game of Thrones dataset, paying special attention to how variables are
named and coded.

1. There had been much publicity about the major role of female characters in the television series
Game of Thrones. Using SPSS, calculate chi-square goodness of fit analyses to determine if:
a. the ratio of named male to female characters was greater than what would be expected by
chance.
b. the ratio of named male to female characters was statistically significantly greater than
two to one (e.g., twice as many males than females).
2. There had been much talk about how many characters have been killed on the show. Did
more characters die than stay alive by the end of Season 6 than would be expected by chance
alone? Using SPSS, calculate a chi-square goodness of fit.
3. Write a few sentences summarizing the results as you would in a Results section.
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
____________________________________________________________________
 IBM® SPSS®
WORKBOOK PART 3
ANALYSES TO EXAMINE
RELATIONSHIPS (NO GROUPS)

Decision Tree When Examining Relationships (no groups)

Decision Tree When Examining Relationships (no groups)

What is the ordinal Spearman’s rho

scale of Correlate 2 variables Pearson’s r
measurement for
the variables? interval or ratio
1+ variables predict an outcome Linear Regression

Compare Groups

PEARSON’S r AND SPEARMAN’S RHO

Are the groups independent or dependent?
The following example focuses on Pearson’s r, using two interval variables. The same steps are used to calculate Spearman’s
rho for two ordinal variables. Note that these steps are also used to calculate a point-biserial correlation to compare two
Independent Dependent Both
groups (referenced in Part 4 of this workbook).

What is the scale of What is the scale of What is the scale of

Data Entry
measurement for measurement for the measurement for the
the outcome/DV? outcome/DV? outcome/DV?

2 groups McNemar Test

nominal
Data Entry
chi-square
nominal
test of
Hereindependence
are data that have
3+ groups Cochran Q Test
Mixed
been entered for 10 participants. interval
ANOVA
Note that each participant has a or ratio
pair of data entered—one value
Mann-Whitney
for2 each
groups variable
U Test in the relationship 2 groups Wilcoxon T Test
ordinal (in this example, the variables are
ordinal
frustration tolerance and self
control, bothKruskal-Wallis
measured on an Friedman 2 test
3+ groups H Test 3+ groups
interval scale).

independent- dependent-
2 groups samples t test samples t test
2 groups
(and/or point-biserial r)
interval interval
or ratio 3+ groups or ratio 3+ groups
one -way
one-way within-
between- subjects
subjects ANOVA
2+ ANOVA 2+
factors factors

Create a Scatterplot two-way

two-way
between- within-
It is a good idea to first create a scatterplot to determine whether there are outliers in your data and to determine if as
subjects subjects
ANOVA ANOVA
one variable increases the other increases or decreases in a predictable pattern. This is called a linear relationship when
analyzing interval or ratio data and a monotonic relationship when analyzing ordinal data.

212
 IBM® SPSS® Workbook Part 3 ■ Analyses to Examine Relationships (No Groups)  213

To create a scatterplot, on the menu bar, click

Graphs → Legacy Dialogs → Scatter/Dot → Simple Scatter → Define

Send your Y variable to the “Y axis” (the vertical axis) box and your X variable to the “X axis” (the
horizontal axis) box.
X is the predictor variable (in this example, self-control) and Y is the predicted variable (in this example,
frustration tolerance). When you do not have a clear predictor and outcome, it doesn’t matter how you
identify X and Y.
Click “OK.”

9.00

8.00
A scatterplot will appear in the
output.
Frustration Tolerance

7.00
In the example to the left, we see
that there is a positive and mostly
6.00 linear relationship between
self-control (X ) and frustration
tolerance (Y ). Notice that there are
some data points that don’t fit with
5.00
the overall pattern, but they are
not extreme enough to be
considered outliers.
4.00

25.00 30.00 35.00 40.00 45.00 50.00

Self-Control

Calculate and Interpret a Correlation Coefficient

To calculate a correlation coefficient, on the Menu Bar, click

Analyze → Correlate → Bivariate

1. Send your X and Y variables to

the “Variables” box.

2. Click “OK.”

Notes: Click “Pearson” to calculate either a

Pearson’s r or a point biserial r.

Click “Spearman” to calculate Spearman’s rho.

214   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

OUTPUT
Correlations
Frustration r = .704
Tolerance Self-control
p = .023 (and is statistically
Frustration Pearson Correlation 1 .704*
significant at the p < .05
Tolerance Sig. (2-tailed) .023 criteria)
N 10 10 Notice that this information
Self-control Pearson Correlation .704* 1 is repeated at the top and
bottom of the correlation
Sig. (2-tailed) .023
table.
N 10 10

*. Correlation is significant at the 0.05 level (2-tailed).

You can also have SPSS calculate a confidence interval for the correlation coefficient by clicking on
that option after you send your X and Y variables to the “variable box.” A new box will open where you
can indicate the confidence interval you prefer (e.g., 95%). If your version of SPSS does not include
this option, check with your professor if you would like to learn about how to calculate the confidence
interval for r by hand.
To find the proportion of variance accounted for, simply square r. In this example, r2 = .50, indicating
that 50% of the variance in one variable is accounted for by the other.

Conducting Multiple Correlation Coefficients

Often researchers calculate multiple correlations simultaneously and report them in a table. Below is
an example output with the interval variables frustration tolerance and self-control, and the ratio variable
income.

Correlations
Frustration Annual
Tolerance Self-Control Income
Frustration Tolerance Pearson Correlation 1 .704* .673*
Sig. (2-tailed) .023 .033
N 10 10 10
Self-Control Pearson Correlation .704* 1 .842**
Sig. (2-tailed) .023 .002
N 10 10 10
Annual Income Pearson Correlation .673* .842** 1
Sig. (2-tailed) .033 .002
N 10 10 10
Sig. (2-tailed) .208 .013 .116
N 10 10 10

*. Correlation is significant at the 0.05 level (2-tailed).

**. Correlation is significant at the 0.01 level (2-tailed).
 IBM® SPSS® Workbook Part 3 ■ Analyses to Examine Relationships (No Groups)  215

The output is a matrix that correlates each variable with all the other variables. Notice also that
each of the correlations is listed twice, so you need to pay attention only to the top or bottom half
of the matrix.
For this example, self-control is significantly positively correlated with frustration tolerance (r =
.70, p = .023) and annual income (r = .84, p = .002) and frustration tolerance and annual income are
also significantly positively correlated (r = .67, p = .033).

Writing Up Results
When you write up results from a Pearson’s r, you will include the following information:

—— The statistical analysis used (e.g., Pearson’s correlation)

—— The variables examined (never use SPSS codes such as Q1)
—— The r, rounded to two decimal places
—— The direction of the relationship (positive or negative)
—— The strength of the relationship (weak, moderate, strong, etc.)
—— The results of statistical significance testing with the p value, rounded to two or three decimal
places.
{{ Do not round to two decimal places when doing so changes the interpretation of
statistically significance (e.g., do not round .046) or changes the p to zero (e.g., do not
round .002).
{{ If the p value at three decimal places is zero, report p < .001.
—— Optional: r 2 or the proportion of variance accounted for
You may also report the confidence interval for r. Check with your instructor.

   Use this same format for a Spearman’s rho or point-biserial correlation, but use the appropriate
statistical notation (rs and rpb, respectively).
Examples:
Option 1: Report all the results in text. Note that in this example we reported all of our p values
at three decimal places for consistency. We need to report the result of p = .002 because rounding
would have resulted in p as zero. We could have rounded p = .023 to p = .02 and p = .033 to
p = .03.
Results
   We calculated a series of Pearson’s r correlations to examine the relationship between self-control,
frustration tolerance, and income. Self-control had strong, positive, and statistically significant
correlations with frustration tolerance (r = .70, p = .023) and annual income (r = .84, p = .002).
The correlation between frustration tolerance and annual income was strong, positive, and statistically
significant (r = .67, p = .033).
Option 2: To avoid the redundancy of listing all the correlation coefficients, you might consider using
a table to summarize the results. You would highlight the major findings in the Results section and refer
to the table.
216   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Results
  Pearson’s r correlations indicated that the relationship between self-control, frustration tolerance,
and income were strong and positive. All correlations were also statistically significant at the p < .05
criteria (see Table 1).

(In an APA-style research report, Tables appear after the Reference section.)

Table 1
Correlations Among Study Variables

Self-Control Frustration Tolerance

Self-Control —
Frustration Tolerance .70* —
Income .84** .67*

*p < .05, **p < .01

 IBM® SPSS® Workbook Part 3 ■ Analyses to Examine Relationships (No Groups)  217

LINEAR REGRESSION
Linear regression is used when you want to examine a linear relationship between two or more variables.
Simple linear regression is used when you have two interval/ratio variables, and you wish to predict the
value of one based on the other. Multiple regression is used when you have more than two variables.
To calculate a linear regression, click

Analyze → Regression → Linear

1.Send your predictor (X )

variable(s) to the
“Independent(s)” box and
your criterion (Y ) variable
to the “Dependent” box.

2. Click “OK.”

Note: The example shown is a

simple linear regression. To
conduct a multiple regression,
enter additional predictor
variables in the
“Independent(s)” box.

You can also request a graph of the regression line (line of best fit) for a simple linear regression:

Click Analyze → Regression → Curve Estimation

1. Send your variables to the

Independent (X) and
Dependent(s) (Y) variable boxes.

2. Make sure “Linear” is checked.

3. Click “OK.”

OUTPUT (abbreviated)
Regression

Model Summary
When you only have 2 variables,
Adjusted R Std. Error of
Model R R Square Square the Estimate R = r and R Square = r 2.

1 .704 a .496 .433 1.23186 Std. Error of the estimate is the

average error of Y ’ in predicting Y.
a. Predictors: (Constant), Self-control

ANOVA a
Sum of Mean The ANOVA table tells you if the r is
Model Squares df Square F Sig. statistically significant.
1 Regression 11.960 1 11.960 7.881 .023 b
 Adjusted R Std. Error of
Model R R Square Square the Estimate R = r and R Square = r 2.

1 .704 a .496 .433 Std. Error of the estimate is the

1.23186
average error of Y ’ in predicting Y.
218   Student Study Guide Witha.IBM ®
SPSS
Predictors:
®
Workbook
(Constant), for Research Methods, Statistics, and Applications
Self-control

ANOVA a
Sum of Mean The ANOVA table tells you if the r is
Model Squares df Square F Sig. statistically significant.
1 Regression 11.960 1 11.960 7.881 .023 b
Residual 12.140 8 1.517
Total 24.100 9

a.Dependent Variable: Frustration Tolerance

b.Predictors: (Constant), Self-control

Coefficients a
Unstandardized Coefficients Standardized Coefficients
Model B Std. Error Beta t Sig.
1 (Constant) 1.657 1.838 .901 .394
Self-control .144 .051 .704 2.807 .023

a.Dependent Variable: Frustration Tolerance

If your r is statistically significant, you can calculate the regression equation (Y' = bX + a). You find
the a and b for your regression equation in the B column.

The value labeled “Constant” (1.657) is a (the y intercept) and the value labeled as the x variable
(.144) is b (the slope). These values are used to create the regression equation. In this case,

Y ' = .14X + 1.66

Frustration Tolerance

9.00

Lastly, we have a scatter plot with the

8.00 regression line or line of best fit. Note
that the actual values do not lie on the
line but around it as our prediction of Y
7.00 values from X values is not perfect (we
would need a perfect correlation for
that to happen!).
6.00

5.00
Observed
Linear
4.00
25.00 30.00 35.00 40.00 45.00 50.00
Self-Control

Now that you have the regression equation, Y ′ = .14X + 1.66, you can use it to predict frustration
tolerance (Y ) based on specific self-control (X ) values.
If self-control = 40, then we predict frustration tolerance will be

Y ′ = .14 (40) + 1.66 = 5.60 + 1.66 = 7.26

The Y ′ of 7.26 will lie on the line of best fit, while the actual value for frustration tolerance = 6.00
when self-control = 40. See the scatter plot above. This difference contributes to the standard error of
the estimate (average difference between actual and predicted Y values), which we can see in the above
output is 1.23.
 IBM® SPSS® Workbook Part 3 ■ Analyses to Examine Relationships (No Groups)  219

PART 3 PRACTICE EXERCISES

A marketing consultant for an online dating site wonders whether time spent on the Internet during
leisure hours is related to whether a person believes they can find a romantic partner online. She surveys
20 young adults who are not in a committed relationship about their recent time online (in minutes)
and their belief about the likelihood of finding a romantic partner online (as assessed with a 5-point
scale where 1 = very unlikely and 5 = very likely). The data are below:

Participant Time online Find partner? Participant Time online Find partner?
1 60 4 11 10 1

2 5 1 12 300 4

3 14 2 13 45 2

4 30 3 14 15 1

5 30 4 15 60 4

6 60 2 16 240 5

7 120 2 17 60 3

8 60 3 18 15 1

9 120 5 19 45 3

10 120 4 20 150 4

1. Using SPSS, enter the data and calculate the statistics below.
a. Compute the appropriate descriptive statistics for the two variables. Describe them.


b. What is the appropriate correlational statistic to compute for these data?

c. Compute a scatterplot for the data. What type of relationship does it suggest?


d. State a null and directional alternative hypothesis for the relationship.


e. Compute the correlation coefficient.
f. Is the relationship between time on the Internet and belief in finding a partner on the
Internet significant? How do you know?


g. What proportion of variability is accounted for by the relationship?

h. Is it appropriate to predict belief in finding a partner on the Internet from the time spent
online? Explain your answer.

220   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

i. Compute the regression for the relationship in question f and write the regression
equation. 

j. Comment on the standard error of estimate (small/moderate/large). How do you know
how to interpret this value?


k. If a person spends 90 minutes on the Internet, what is the predicted score on the
likelihood to believe they will find a romantic partner on the Internet? 
l. Describe your findings as you would in a Results section, using APA format.






m. Write a Discussion section for your results. Include all the information that belongs in
this section.






2. Two judges for a poetry contest were asked to rank the 10 finalists’ poems from 1 = best to
10 = least good. Their rankings are below:

Judge 1 1 2 3 4 5 6 7 8 9 10

Judge 2 4 7 3 1 9 2 10 8 5 6

a. What is the appropriate statistic to compute? 

b. Compute the statistic with SPSS and write a few sentences describing your finding and its
implications.


3. For additional practice with the different correlation statistics and regression, pull up the
datasets that are available for this course or datasets your instructor has made available.
a. Identify variables you would like to correlate and their scale of measurement, then decide
on the appropriate correlation to compute.
b. Compute the statistic(s) and interpret your results.
 IBM® SPSS®
WORKBOOK PART 4
ANALYSES TO COMPARE
INDEPENDENT GROUPS

Decision
DecisionTree
TreeWhen
When Comparing Independent
Comparing Independent Groups
Groups or Conditions
or Conditions

chi-square test
nominal
for independence
What is the Mann-Whitney U Test
scale of 2 groups
measurement ordinal
Kruskal-Wallis H Test
for the 3+ groups independent-samples t test
outcome/DV? (and/or point-biserial r)
2 groups

3+ groups one-way between-subjects ANOVA

interval or ratio
2+ factors
two-way between-subjects ANOVA
(or three-way, etc.)

INDEPENDENT-SAMPLES t TEST
You need at least two variables to run an independent-samples t test. One variable is the IV (if you have an experiment)
or grouping variable (predictor if you have a correlational design); the other is the test variable (DV for experiment or
outcome for correlation).

Assumptions of the Independent-Samples t Test

1. Groups are independent.
2. IV (or grouping variable) is nominal and dichotomous.
3. DV (or outcome variable) is interval or ratio.
4. DV is normally distributed (although you can still run a t test if this assumption is violated).
5. The variability (or SD) in both groups are about equal. (A test called the Levene’s test will tell you if this
assumption is violated, and, if it is, you simply need to report a different t, df, and p.)

A Simple Experiment Example

For a simple experiment, we will use a study mentioned based on Kirk and Lewis (2016) in which gesturing was or was
not encouraged as children named novel uses for objects. In this example, we have a variable called “gesturing” that is
coded as 0 = no and 1 = yes. We also have the dependent variable called “creativity” that is the number of novel, feasible
uses for an object generated by each participant.

221
222   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

—— The independent variable: Gesturing (no vs. yes)

—— The dependent variable: Creativity (number of novel, feasible uses for object)
—— Null hypothesis: Gesturing will not affect the number of novel, feasible uses generated
(creativity) by children.
—— Alternative directional hypothesis: The children who are encouraged to gesture will generate
more novel, feasible uses of objects than those not encouraged to gesture

Data Entry
The variable view would look like this:

The data view would look like this:

 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  223

Conducting an Independent-Samples t Test

On the Menu Bar, click

Analyze → Compare Means → Independent Samples t test

1.Send your outcome/DV to

the “Test Variable(s)” box.

2. Send your predictor/IV to

“Grouping Variable.”
4. Click “Define
3. Be sure “Estimate Groups”
effect sizes” is checked.

The “Define Groups” box will open in order for you to identify how you coded your groups or conditions.

5. Indicate how you coded

your predictor/IV.

In this example:
Group 1(no gesturing) = 0

Group 2 (gesturing) = 1

Enter the codes that

match YOUR data.

6. Click “Continue” and then

“OK” to run your t test.

OUTPUT
T Test

Group Statistics
experimental Std. Error
condition N Mean Std. Deviation Mean

Creativity no Gesturing 12 6.2500 1.54479 .44594

The first table provides descriptive
Gesturing 12 8.0833 1.92865 .55675 information for each of your groups.

The t test results are in the second table.

Interpreting the independent-samples t test is a two-step process.
Step 1: Look at Levene’s test to tell you what row to use. The p value of the Levene’s test
compares the variance in each group; it tests the assumption of equal variances, not the
hypothesis.

Step 2: Look at the t test results in the appropriate row. The p value of the t test tells you
whether or not there is a statistically significant difference between the means of the two
groups. In other words, the t test is what tests your hypothesis.
224   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Step 1. First look at the significance level If Levene’s is not statistically significant
of the Levene’s test.* (p ≥ .05), then use the data from the 1st row (equal variances
assumed).

Independent Samples Test

Levene’s
Test for
Equality of
Variances t test for Equality of Means
95% Confidence Interval
of the Difference
Sig. Mean Std. Error
F Sig. t df (2-tailed) Difference Difference Lower Upper
Creativity Equal .180 .675 1st −2.570 22 .017 −1.83333 .71333 −3.31269 −.35398
variances
assumed
Equal 2nd −2.570 20.999 .018 −1.83333 .71333 −3.31269 −.35398
variances
not
assumed

If Levene’s is statistically significant (p < .05), then use the

data from the second row (equal variances not assumed).

In this example, Levene’s is not statistically significant because p = .68 is greater than .05.
Consequently, we can assume equal variances and will use the t test data from the 1st row.

*What Is Levene’s Test?

Levene’s tests the assumption that the groups have equal variances. In other words, that one group
doesn’t have a much higher variance (or SD) than the other. You want Levene’s to be > .05 because it
allows you to use a more powerful test. If Levene’s is significant, you have to use a less powerful test
with less df, which may impact your ability to find a statistically significant difference between the two
groups.
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  225

Step 2. Once you have figured out which row to examine,

determine if the t value meets the criteria for statistical
significance.

Independent Samples Test

Levene’s
Test for
Equality of
Variances t-test for Equality of Means
95% Confidence
Interval of the
Difference
Sig. Mean Std. Error
F Sig. t df (2-tailed) Difference Difference Lower Upper
Creativity Equal .180 .675 −2.570 22 .017 −1.83333 .71333 −3.31269 −.35398
variances
assumed
Equal −2.570 20.999 .018 −1.83333 .71333 −3.31269 −.35398
variances
not
assumed

In this example, we will look at the first row’s data in order to determine if there is a
statistically significant difference in creativity between the gesturing and no gesturing
conditions.
We see that p = .02 (rounded from .017), and this meets the criteria for statistical
significance at the p < .05 level.
We would report: t(22) = -2.57, p = .02.

Effect Size for a Two Independent-Groups Design

SPSS will calculate effect sizes for the independent-samples t test along with confidence intervals. Below
is the output from our example.

Independent Samples Effect Sizes

95% Confidence Interval

Standardizer Point Estimate Lower Upper

creativity Cohen’s d 1.747 −1.049 −1.896 −.181

Hedges’ correction 1.810 −1.013 −1.831 −.175

Glass’s delta 1.929 −.951 −1.825 −.042

226   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

We would report Cohen’s d = 1.05 (rounded and omitting the sign). This indicates that there is a little
over a full standard deviation difference between the two conditions.
Alternatively, you can calculate and report the proportion of variance accounted for, which is rpb2 for
a two independent-groups design. To do this, you would calculate a point-biserial correlation following
the directions in Part 3. Note that the steps and output indicate a Pearson’s r, but you would interpret
it as a rpb.

Correlations
gesturing creativity
Gesturing Pearson Correlation 1 .481*
Square the correlation coefficient to calculate
Sig. (2-tailed) .017
the proportion of variance accounted for.
N 24 24
In this example, rpb = .48 and rpb2 = .23,
Creativity Pearson Correlation .481* 1 indicating that 23% of the variance in creativity
is accounted for by the gesturing condition.
Sig. (2-tailed) .017
N 24 24

In your write-up, describe the effect size (either Cohen’s d or rpb2 ) as weak, moderate, or strong. The
chart below provides guidelines rather than absolute cutoffs for this description. For example, you may
use “moderate to strong” to describe an effect size.

OVERVIEW OF EFFECT SIZE INTERPRETATIONS (COHEN, 1988)

Effect Size

Proportion of Variance
Accounted for Cohen’s d Interpretation

≈ 1% ≈.20 Small/Weak

≈ 9% ≈.50 Medium/Moderate

≈ 25% ≈.80 Large/Strong

A Nonexperimental Example
Do children who speak more than one language (multilingual) come up with more or less ways to use
an object than those who speak one language (unilingual)?

OUTPUT

Group Statistics
language N Mean Std. Deviation Std. Error Mean
number of uses for multilingual 28 6.5714 4.46740 .84426
object
unilingual 29 6.4483 1.76445 .32765
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  227

Independent Samples Test

Levene's Test
for Equality of
Variances t-test for Equality of Means

95% Confidence Interval of

the Difference

Sig. Mean Std. Error

F Sig. t df (2-tailed) Difference Difference Lower Upper
number Equal 5.272 .026 .138 55 .891 .12315 .89387 −1.66821 1.91451
of uses variances
for assumed
object
Equal .136 34.981 .893 .12315 .90561 −1.71537 1.96167
variances
not
assumed

Step 1. In this example, Levene’s Step 2. Looking at the 2nd row, we find that the difference
is statistically significant (p = .03). between means is not statistically significant (p = .89).
Consequently, we cannot assume equal Because we violated the assumption of equal variances,
variances and will use the t test results we had to use a less powerful test with fewer degrees of
from the 2nd row. freedom (in this case, df = 34.98).
We would report: t(34.98) = 0.14, p = .89.

Note that in this example, we rounded all the values to two decimal places. Additionally, we can use
SPSS to calculate Cohen’s d to find d = 0.04.

Writing up Results
When reporting the results of an independent-samples t test, include the following:

—— The type of analysis you conducted and the variables examined.

—— A comparison of the mean and SD for each group
{{ Here are a couple examples from the gesturing experiment:
nn The gesture-encouraged group (M = 8.08, SD = 1.93) generated more novel uses than
the no encouragement group (M = 6.25, SD = 1.54).
nn The children who were encouraged to gesture generated an average of 8.08 novel uses
(SD = 1.93), whereas the children in the control group generated an average of 6.25
(SD = 1.54).
—— The results of the independent-samples t test and effect size calculation
nn Include:
{{ the degrees of freedom in parentheses
{{ t value rounded to two decimals, with a leading zero if t is less than one.
228   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

{{ the p value rounded to two or three decimals

nn Do not round to two decimal places when doing so changes the interpretation of
statistical significance (e.g., do not round .046) or changes the p to zero (e.g., do not
round .002).
nn If the p value at three decimal places is zero, report p < .001.
{{ Cohen’s d rounded to two decimals, with a leading zero if d is less than one OR rpb2
rounded to two decimals or expressed as a percentage
Examples:
nn t(22)= −2.57, p = .02, d = 1.05
nn t(22)= −2.57, p = .02, rpb2 = .23
nn t(22)= −2.57, p = .02. Twenty-three percent of the variance in novel uses was
accounted for by the gesture condition.
—— A statement about whether or not the difference between the two groups was statistically
significant. You can simply add a word or two to one of the above statements.
{{ Examples:
nn The gesture-encouraged group (M = 8.08, SD = 1.93) generated statistically
significantly more novel uses than the no encouragement group (M = 6.25,
SD = 1.54).
nn The difference between the groups was statistically significant, t(22) = −2.57, p = .02,
d = 1.05.
—— If Levene’s test was not significant, you don’t need to mention it at all. If Levene’s was
significant, explain that you had to use a more stringent test because Levene’s was significant
(or because the assumption of equal variances was violated). In the latter case, be sure to cite
the statistics on the second line of the t test output.
—— You might also report confidence intervals. Check with your instructor.

Example Results for a Nonexperimental Study

Results
   We conducted an independent-samples t test to compare children’s ability to generate different
uses for an object between those who are multilingual (M = 6.57, SD = 4.47) and those who are
unilingual (M = 6.45, SD = 1.76). Results were very weak and not statistically significant, t(34.98) =
0.14, p = .89, d = 0.04. The results violated the assumption of equal variance in that the variability
within the multilingual group was much higher than within the unilingual group.

Independent-Samples t Test Practice Exercises

Practice Exercise 1
A researcher is investigating the effect of continuing education on supervisors’ effectiveness. He has
supervisors take either a month-long online course that focuses on mindfulness as an approach to
leadership or one that focuses on general skills of leaders. The supervisors’ effectiveness is rated by a
researcher blind to the conditions of the study, one month after the continuing education courses are
completed. The following data result:
Mindfulness General Leadership

7.00 4.00

6.00 4.00

7.00 7.00
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  229

6.00 5.00

5.00 7.00

5.00 7.00

7.00 5.00

4.00 4.00

6.00 6.00

7.00 4.00

8.00 5.00

7.00 7.00

5.00 3.00

7.00 5.00

5.00 2.00

1. What is the design of this study (be specific)?

2. What is the IV? DV?
3. State a directional alternative hypothesis for the study.
_____________________________________________________________________
_____________________________________________________________________
4. Compute using SPSS all the appropriate statistics to test your hypothesis.
5. Are the findings statistically significant? How do you know?

6. State and interpret the effect size.  

7. Write up the results using APA format.
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
8. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
230   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Practice Exercise 2
A math professor decides to investigate whether stereotype threat is operating in her class for students
of color. She randomly assigns students in her classes to one of two instructions for a practice test.
In one condition, students are told that they are taking a practice test and to do their best. In the
other condition, students are told that they are taking a standardized test that was designed by several
professors at MIT, but to do their best. The test scores for her students of color are shown below.

Regular Instructions Stereotype Threat

90.00 95.00

85.00 65.00

70.00 35.00

75.00 65.00

80.00 70.00

75.00 60.00

75.00 80.00

80.00 50.00

85.00 85.00

85.00 75.00

1. State a directional alternative hypothesis for the study.

_____________________________________________________________________
2. Compute the statistics needed to test your hypothesis.
3. Are your findings statistically significant? Explain your answer.
_____________________________________________________________________
4. What type of error – Type I or Type II – might there be?
5. What potential problem(s) do you identify based on the findings?
_____________________________________________________________________
_____________________________________________________________________
6. Write a few sentences describing the findings as you would in a Results section.
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
7. Interpret the findings as you would in a Discussion section (what are the implications?).
_____________________________________________________________________
_____________________________________________________________________
_____________________________________________________________________
8. What do you suggest for future researchers to improve upon the study?
_____________________________________________________________________
_____________________________________________________________________
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  231

ONE-WAY BETWEEN-SUBJECTS ANOVA

A one-way between-subjects analysis of variance (ANOVA) is used to compare three or more independent
groups when the outcome or DV is measured on an interval or ratio scale. You can use the one-way
ANOVA for experimental and nonexperimental data.
Assumptions of the One-Way Between-Subjects ANOVA

1. Groups are independent

2. IV or grouping variable is nominal (with two or more groups)*
3. DV or test variable is interval or ratio
4. DV or test variable is normally distributed
5. Variability is approximately equal across groups

You can calculate the analysis with just two groups, but the independent-samples t test is the preferred
*

method to compare the means of two independent groups.

Overview of the Analyses for a One-Way Between-Subjects ANOVA

—— The ANOVA analysis will give you an F (instead of a t), although like the t, F = Between-
group differences ÷ Within-group differences (error)
—— When you report an F statistic, you need to report two types of degrees of freedom:
between-groups (treatment) df = number of groups (k) − 1
within-groups (error) df = number of participants (N) − number of groups (k)
—— You also need to report the effect size. Eta (η) is a correlation coefficient used to describe
the magnitude of a relationship containing two or more levels. Eta squared (η2) tells you the
proportion of variance in the DV or outcome accounted for by the IV or predictor.
—— If you find a statistically significant F, you will need to report the results of post hoc tests. Do
not report post hoc tests if the F is not statistically significant.

A Multiple-Group Experiment Example

A researcher hypothesizes that plants will increase perceptions of productivity. She randomly assigns
employees at a company to view one of three pictures: An office with no plants, an office with one
plant, or an office with three plants. The employees then estimate how productive they believe the office
occupant is, on a scale from 1 to 10 with 10 = very productive.

Data Entry
To run a one-way ANOVA, you need at least two variables. One variable is the grouping predictor or
IV, the other is the outcome or DV.
Data entry for the one-way ANOVA is very similar to the independent-samples t test, except that your
IV will have three or more levels.
In the multiple-group experiment example, we have a variable called “plants” that is coded as 0 =
zero plants, 1 = one plant, 2 = three plants and entered in one column. We also have a variable called
“productive” that is scored on an interval scale and entered in a second column.

Conducting a One-Way ANOVA

First, you want to find out whether there is a significant difference between your groups and the
magnitude of this difference.
232   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

On the Menu Bar, click

Analyze → Compare Means → One-Way ANOVA

1. Send your outcome/DV

to the “Dependent list.”
4. Click
“Options”

2. Send your predictor/IV

to “Factor.”

3. Click “Estimate effect size…”

The options box will appear:

5. Click
“Descriptive,”
“Homogeneity of
variance test,”
and “Means plot.”

6.Click
“Continue”
and then
“Post Hoc.”

The Post Hoc box will appear:

7.Click “LSD”
or one of
the other
tests.

8.Click “Continue” and then “OK.”

 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  233

OUTPUT

Descriptive Statistics are in the first output table.

Descriptives

Productive

95% Confidence
Interval for Mean

N Mean Std. Deviation Std. Error Lower Bound Upper Bound Minimum Maximum

Zero 14 6.2857 .82542 .22060 5.8091 6.7623 5.00 8.00

One 14 7.8571 1.02711 .27451 7.2641 8.4502 6.00 9.00

three 14 6.7143 1.20439 .32189 6.0189 7.4097 5.00 9.00

Total 42 6.9524 1.20876 .18652 6.5757 7.3291 5.00 9.00

The next output table depicts the results of the Levene’s test. In this example, Levene’s is not statistically
significant (p = .29), and therefore, we have met the assumption of equal variances. If you find a statistically
significant Levene’s, make sure the data are normally distributed and check for outliers.

Tests of Homogeneity of Variances

Levene Statistic df1 df2 Sig.

productive Based on Mean 1.297 2 39 .285

The ANOVA table will tell you if there is a statistically significant difference between your groups. In this
example, F is statistically significant at p < .001. We would report: F(2, 39) = 8.70, p < .001.

ANOVA

productive

Sum of Squares df Mean Square F Sig.

Between Groups 18.476 2 9.238 8.697 <.001

Within Groups 41.429 39 1.062

Total 59.905 41

In the effect size table, we see that η2 = .31, Which means that 31% of the variance in perceived productivity is
accounted for by the plants in the office. We might also report the confidence interval of the effect size:
CI [0.07, 0.48]

ANOVA Effect Sizes

Point Estimate 95% Confidence Interval

Lower Upper

productive Eta-squared .308 .071 .479

234   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Only report results of post hoc tests if the ANOVA result (F) is statistically significant. Because our F was
statistically significant, we would report the comparisons. In this example, the difference between zero and one
plant (p < .001) and one and three plants (p = .006) was statistically significant, but we do not have a statistically
significant difference between zero and three plants (p = .28).

Post Hoc Tests

Multiple Comparisons

Dependent Variable: productive

LSD

(I) plants (J) plants Mean Difference (I-J) Std. Error Sig. 95% Confidence Interval

Lower Bound Upper Bound

Zero One −1.57143* .38955 <.001 −2.3594 −.7835

Three −.42857 .38955 .278 −1.2165 .3594

One Zero 1.57143* .38955 <.001 .7835 2.3594

Three 1.14286* .38955 .006 .3549 1.9308

Three Zero .42857 .38955 .278 −.3594 1.2165

One −1.14286* .38955 .006 −1.9308 −.3549

* The mean difference is significant at the 0.05 level.

Means Plots

8.00

The Means Plot

7.50 allows you to see
the pattern of
Mean of productive

differences between
your groups.
7.00

6.50

6.00

zero one three

Plants
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  235

Writing up Results
Include the following:

—— The type of analysis(es) conducted and the variable examined.

{{ One-way between-subjects ANOVA
{{ Identify the IV/predictor and DV/outcome
—— The mean and standard deviation for each group or condition.
—— The results of the one-way between-subjects ANOVA and effect size calculation
Include:
{{ the between-groups df and the within-groups df, separated by a comma and in
parentheses
{{ F value rounded to two decimals, with a leading zero if F is less than one.
{{ the p value rounded to two or three decimals as necessary for appropriate interpretation of
statistical significance.
{{ eta squared (η2) rounded to two decimals or expressed as a percentage

Examples:

n F(2, 39) = 8.70, p < .001, η2 = .31

n F(2, 39) = 8.70, p < .001. Thirty-one percent of the variance in the DV was accounted for by
the IV.
{{ A statement about whether or not the difference between the groups was statistically
significant.
{{ If the F was statistically significant, report what post hoc test you ran (e.g., Least
Significant Difference) and the results of that post hoc test. Indicate which groups were
significantly different from each other with p value and the direction of the difference.
{{ In either the Results or Discussion, include an interpretation of the effect size.

Note: An η2 around .01 is weak, around .06 is moderate, and around .16 is strong (Cohen, 1988).

—— You might also report confidence intervals. Check with your instructor.

Example
Results
   Employees estimated the productivity of the occupant of an office with zero plants (M = 6.29,
SD = 0.83), one plant (M = 7.86, SD = 1.03), or three plants (M = 6.71, SD = 1.20). A one-way
between-subjects ANOVA revealed a statistically significant difference between conditions F(2,
39) = 8.70, p < .001, η2 = .31. Post hoc least significant difference (LSD) tests further indicated
statistically significant differences between zero and one plant p < .001) and between one and
three plants (p = .006). However, the difference between zero and three plants was not statistically
significant (p = .28).
236   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

One-Way Between-Subjects ANOVA Practice Exercises

Practice Exercise 1
A consultant who specializes in political campaigns tests the success of three ads in convincing registered
voters to support a candidate. The ads run for 30 seconds and feature issues of the campaign. One ad
presents the candidate’s rationale for supporting issues, a second ad berates the opponent of the candidate
and encourages people to vote for the candidate, and the third ad contrasts the two candidates’ positions
on issues. Registered voters are randomly assigned to view one of the three ads and then rate on their
likelihood (0%–100%) of voting for the candidate. The responses are below:

Rationale Berate opponent Contrast positions

70 25 95

65 30 55

90 55 70

85 60 80

65 35 60

55 40 75

35 65 100

70 15 90

40 75 40

60 30 55

30 35 85

25 45 80

1. What is the design of this study (be specific)?

2. What is the IV? DV?

3. Compute using SPSS all the appropriate statistics to test your hypothesis.
4. Are the findings statistically significant? How do you know?
__________________________________________________________________________
5. State and interpret the effect size.
__________________________________________________________________________
6. Write up the results using APA format.
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  237

7. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________

Practice Exercise 2
Studies suggest that time outside increases one’s sense of well-being, but does more time increase one’s
well-being or is there only an initial gain and then a leveling of well-being? A researcher has students
spend no time (control), 10 minutes, 30 minutes, or 90 minutes outside and then has them complete a
well-being scale (higher scores reflect a greater sense of well-being). The following data result:

Control 10 minutes 30 minutes 90 minutes

18.00 16.00 20.00 21.00

22.00 20.00 25.00 23.00

25.00 23.00 28.00 28.00

15.00 16.00 29.00 18.00

22.00 19.00 30.00 19.00

16.00 25.00 25.00 17.00

22.00 27.00 22.00 19.00

17.00 19.00 16.00 24.00

21.00 16.00 28.00 22.00

15.00 19.00 18.00 15.00

1. What type of design is this?

2. State a directional alternative hypothesis.
3. Compute all the statistical analyses needed to test your hypothesis.
4. Describe your findings.
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
5. Interpret these findings.
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
238   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

TWO-WAY BETWEEN-SUBJECTS ANOVA

(OR THREE-WAY, ETC.)
Data Entry
At minimum, you need three variables: (1) Factor 1; (2) Factor 2; (3) Outcome/DV
Example: Below is fictional data for 20 pregnant women who participated in a 2 × 2 hybrid design in
which they were randomly assigned to have different magazines available in their doctor’s waiting room
(nature vs. family) and whether or not they have been pregnant before (no vs. yes). Fear of delivery is
the DV and measured on an interval scale.

IV conditions:
Nature Magazines Family Magazines
5 2
4 3
Never been pregnant before 5 3
3 4
History: 4 2
3 8
5 6
Has been pregnant 4 7
5 6
3 8

In the example dataset that follows, we have entered these variables in SPSS:

—— “magazine” coded as 1 = nature magazines, 2 = family magazines

—— “history” coded as 0 = never been pregnant, 1 = has been pregnant
—— “fear” scored on an interval scale.
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  239

Conducting a Two-Way Between-Subjects ANOVA

On the Menu Bar, click Analyze → General Linear Model → Univariate

1. Send your outcome/DV to

“Dependent Variable” and
Send your two factors to
“Fixed Factor(s).”

Note: you can add more

factors in order to conduct a
three-way ANOVA,
four-way ANOVA, etc.

2. Click “Options.”
Select Descriptive Statistics, Estimates of Effect Size,
and Homogeneity tests. Then click “Continue.”

3. Click “Plots,” select one factor to send to “horizontal

axis” box and send the other factor to “separate lines” box.
You can have SPSS plot the graph both ways if you like.
Click “Add” and then “Continue.”

4. Click “OK” to run analyses.

OUTPUT (abbreviated)

Descriptive Statistics group

Dependent Variable: fear of delivery means for
cell Factor 1
magazine condition history Mean Std. Deviation N
means
nature magazine never been pregnant 4.2000 .83666 5
has been pregnant 4.0000 1.00000 5
Total 4.1000 .87560 10
family magazine never been pregnant 2.8000 .83666 5 group
means for
has been pregnant 7.0000 1.00000 5
Factor 2
Total 4.9000 2.37814 10
Total never been pregnant 3.5000 1.08012 10
has been pregnant 5.5000 1.84089 10
grand
Total 4.5000 1.79179 20 mean
240   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Levene's Test of Equality of Error Variancesa,b Read the first line of the output
Levene Statistic df1 df2 Sig. box. In this example, Levene’s is
not statistically significant. If it
fear Based on Mean .220 3 16 .881 was, check that the outcome/DV is
Based on Median .267 3 16 .848 normally distributed and check for
outliers. A significant Levene’s may
Based on Median and with .267 3 15.835 .848
occur when there is a statistically
adjusted df
significant interaction effect.
Based on trimmed mean .211 3 16 .887
Tests the null hypothesis that the error variance of the dependent variable is equal
across groups.
a. Dependent variable: fear
b. Design: Intercept + magazine + history + magazine * history

Tests of Between-Subjects Effects

Dependent Variable: fear of delivery
Type III Sum of Mean Partial Eta
Source Squares df Square F Sig. Squared
Corrected Model 47.400a 3 15.800 18.588 .000 .777
Intercept 405.000 1 405.000 476.471 .000 .968
magazine 3.200 1 3.200 3.765 .070 .190
history 20.000 1 20.000 23.529 .000 .595
magazine * history 24.200 1 24.200 28.471 .000 .640
Error 13.600 16 .850
Total 466.000 20
Corrected Total 61.000 19
a. R Squared = .777 (Adjusted R Squared = .735)

Main Effect for Factor 1 Main Effect for Factor 2 Interaction Effect: F(1, 16) =
(magazine condition): (history): F(1, 16) = 23.53, 28.47, p < .001, η2partial = .64.
F(1, 16) = 3.77, p = .07, p < .001, η2partial = .60.
η2partial = .19.

Writing up Results
Include the following:

—— The type of analysis and the variables examined

—— For each main effect: the means, standard deviations, and results from the F test (including
F, df, p, and η 2partial ; see one-way ANOVA write-up directions and example below for format
guidelines)
—— The results of the interaction:
{{ If not statistically significant, simply report that and the F, df, p, and ηpartial
2

{{ If statistically significant, you should report the results as well as an explanation of the
interaction, with the cell means. A table is a good way to report the cell means and
standard deviations. (Tables are placed after the References in an APA-style report.)
—— You might also report confidence intervals. Check with your instructor.
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  241

Example
Results
  A 2 (history of pregnancy) × 2 (magazine type) between-subjects ANOVA was conducted to
examine factors that impact a pregnant women’s fear of delivery. There was no significant difference in
fear levels for women who were given family-oriented magazines vs. nature magazines (M = 4.90, SD
= 2.38 vs. M = 4.10, SD = 0.88), F(1, 16) = 3.77, p = .07, ηpartial
2
= .19. Women who had never been
pregnant (M = 3.50, SD = 1.08) had lower levels of fear than women who had been pregnant (M = 5.50,
SD = 1.84). The relationship between history and fear of delivery was strong and statistically significant,
F(1, 16) = 23.53, p < .001, ηpartial
2
= .60.
   The main effects were qualified by a statistically significant interaction effect, F(1, 16) = 28.47,
p < .001, which accounted for 64% of the variance in fear. Women who had never been pregnant
reported much lower fears about their upcoming delivery when exposed to family magazines than when
exposed to nature magazines. The opposite effect was found for women who had been pregnant in the
past. These women reported much higher fears when exposed to family magazines rather than nature
magazines (see Table 1).

Table 1
Fear as Function of Pregnancy History and Magazine Group

Family Magazines Nature Magazines

M SD M SD

History of Pregnancy

Never been pregnant 2.80 0.84 4.20 0.84

Has been pregnant 7.00 1.00 4.00 1.00

You might also want to include a graph to demonstrate a significant interaction. Don’t cut and paste
from the output; rather, create a graph using MSWord or data processing software such as Excel.

In an APA-style research report, Figures appear after Tables.

Figure 1 

Interaction Between Magazine Condition and Pregnancy History

6
Fear of delivery

2
never pregnant
1 has been pregnant

0
Nature Family
Magazine condition
242   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Two-Way Between-Subjects ANOVA Practice Exercises

Practice Exercise 1
Following are results based on a factorial experiment that examined the effect of posture and forgetfulness
on ratings of independence of an older adult. Use this information to fill in the tables below.
   Posture conditions did not result in a statistically significant difference in independence ratings
(good posture M = 3.46, SD = 0.89 vs. poor posture M = 3.30, SD = 0.88), F(1, 214) = 1.69, p = .36,
ηpartial
2
= .01 . Likewise, the man who forgot his keys (M = 3.32, SD = 0.89) was not rated as significantly
less independent than the man who remembered his keys (M = 3.43, SD = 0.88), F(1, 214) = 0.85,
p = .20, ηpartial
2
= .004 . However, there was a significant posture by forgetting interaction, F(1, 214) =
6.84, p = .01. Results revealed that the man with poor posture who forgot his keys was rated the least
independent (M = 3.09, SD = 0.90), whereas the man with good posture who forgot his keys was rated
the most independent (M = 3.55, SD = 0.83). The man with poor posture who remembered his keys
was rated more positively (M = 3.51, SD = 0.81) than the man with good posture who remembered
his keys (M = 3.35, SD = 0.96). It should be noted that the interaction only accounted for 3% of the
variance in independence.

Descriptive Statistics
Posture condition forget condition Mean Std. Deviation N
1.00 good posture 1.00 no forgetting 51
2.00 forgot keys 56
Total 107
2.00 poor posture 1.00 no forgetting 55
2.00 forgot keys 56
Total 111
Total 1.00 no forgetting 106
2.00 forgot keys 112
Total 3.38 .888 218

Tests of Between-Subjects Effects

Dependent Variable: how independent?
Type III Sum of Mean Partial Eta
Source Squares df Square F Sig. Squared
Corrected Model 7.371a 3 2.457 3.210 .024 .043
Intercept 2481.255 1 2481.255 3241.978 .000 .938
Posture 1.292 1 1.292
Forget .654 1 .654
Posture forget
*
5.237 1 5.237
Error 163.785 214 .765
Total 2656.000 218
Corrected Total 171.156 217

a. R Squared = .043 (Adjusted R Squared = .030)

 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  243

Practice Exercise 2
A teacher has heard that adding a humorous cartoon to an exam might ease tension and even raise
scores. She wonders whether adding an inspirational quote might also have a positive effect and possibly
strengthen the effect of a cartoon. She creates four different versions of an extra-credit quiz and randomly
assigns students to condition. The quiz results for each condition are as follows:

Inspirational Quote

no quote quote

no cartoon 5 6

3 5

4 4
Humorous Cartoon

8 4

6 7

Cartoon 6 4

7 3

9 6

6 6

9 2

1. Using SPSS, enter the data and calculate a between-subjects two-way ANOVA.
2. Write a Results section and put the cell means and standard deviations in a table.
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________

Table 1
The Effect of Humor and Inspirational Quotes on Quiz Scores

No Inspirational Quote Inspirational Quote

M SD M SD

No Humorous Cartoon _______________ _______________ _______________ _______________

Humorous Cartoon _______________ _______________ _______________ _______________

244   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

3. Interpret your results as you would in a Discussion section.

________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
4. Write a few comments about the ethics of this study.
________________________________________________________________________
________________________________________________________________________

Practice Exercise 3
Go to https://edge.sagepub.com/adams3e. Click on “Datasets” and open the file called “Reminiscence
Study”:
—— Read over the description of the datasets (reminiscence description.pdf )
—— Download the compiled dataset (reminiscence compiled dataset.sav)

1. Calculate a between-subjects two-way ANOVA to determine whether college students tend

to positively anticipate the future more than those who are not in college, and whether this
relationship is moderated by gender.
2. Repeat with “savoring the moment” as the outcome variable.
3. Write a Results section that summarizes both these analyses.
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
________________________________________________________________________
 cointoss
The second table provides the 2, df, and p value.
Chi-Square 1.200 a In this example, the 2 is not statistically
significant and we would report 2(1, N= 30)
df 1
= 1.20, p= .27.
IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  245
Asymp. Sig. .273

NONPARAMETRIC STATISTICS
Conducting a Chi-Square Test for Independence
Conducting a Chi-Square Test for Independence

Data Entry

Enter the two variables into SPSS as

independent codes.

In this example, Health status is coded as

1 = healthy, 2 = unhealthy

Perception of one’s aging process

relative to chronological age coded as
1 = younger, 2 = equal, 3 = older.

To calculate chi-square test for independence, click on the Menu Bar

To calculate chi-squareAnalyze
test for independence,
Descriptiveclick on the Menu
Statistics Bar
Crosstabs
Analyze → Descriptive Statistics → Crosstabs
1. Send one of your nominal
variables to “Row(s)” and
the1.other
Send toone nominal variable
“Column(s).”
to “Row(s)” and the other to
“Column(s).
2. Click “Statistics.”
2. ClickChi-Square
Select “Statistics.” &
Phi/Cramer’s
Select Chi-Square
V.
 Chose the appropriate
3. Click effect
“Continue.”
size.*

4. Click
3. Click
“Cells.”
“Continue.”

4. ClickObserved
Select “Cells” &
Expected.
Select Observed &
Expected
5. Click Continue and then
5. Click “Continue” and “OK.”
“OK.”

*Appropriate effect sizes:

Phi = 2 × 2 contingency table
Cramer’s V = uneven contingency table (e.g., 2 × 3)
Contingency coefficient = even contingency table greater than 2 × 2 (e.g., 3 × 3)
246   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

OUTPUT (abbreviated)

Health Orig * Aging Perception Crosstabulation

Aging Perception
younger equal older Total
Health Orig healthy Count 4 8 1 13
Expected Count 2.6 6.5 3.9 13.0
unhealthy Count 0 2 5 7
Expected Count 1.4 3.5 2.1 7.0
Total Count 4 10 6 20
Expected Count 4.0 10.0 6.0 20.0

Chi-Square Tests
Asymptotic We use the Pearson Chi-Square and
Significance would report:
Value df (2-sided) χ2(2, N = 20) = 9.30, p = .01.
Pearson Chi-Square 9.304 a
2 .010
A statistically significant χ2 means
Likelihood Ratio 10.483 2 .005 that our observed values were
Linear-by-Linear Association 7.879 1 .005 significantly different than the
expected values.
N of Valid Cases 20
a
5 cells (83.3%) have expected count less than 5. The minimum expected count
is 1.40.

Symmetric Measures
Approximate Square the phi or V to find the proportion
Value Significance of variance accounted for. In this example,
Nominal by Nominal Phi .682 .010 we square .68 and find V 2 = .46, or 46% of
the variance accounted for.
Cramer’s V .682 .010
N of Valid Cases 20

Writing Up Results

—— Explain the analysis conducted and variables examined.

—— Explain whether there was a statistically significant relationship or effect.
—— Report the results of the chi-square test for independence (e.g., χ2(2, N = 20) = 0.28, p = .60).
—— Report the appropriate effect size statistic or proportion of variance accounted for (e.g.,
ϕ2 = .02).
—— Explain the results by referring to the count or frequency of the groups (in the text or in a
table).

Example
Results
  Results of a chi-square test for independence revealed that reports of health had a statistically
significant relationship with, and accounted for a large amount of the variance in, age perceptions.
Participants view those who are healthy as their correct age (n = 8) or as younger (n = 4) more than
those who are unhealthy (n = 0 and n = 0, respectively). Participants guess older ages for those who are
unhealthy (n = 5) more than for those who are healthy (n = 1), χ2(2, N = 20) = 9.30, p = .01, V 2 = .46.
 IBM® SPSS® Workbook Part 4 ■ Analyses to Compare Independent Groups  247

Chi-Square Test for Independence Practice Exercise

An administrator at University Prestige wants to know whether gender is independent of status for full-
time faculty at the institution. He finds the following data:

Status

Gender Assistant Professor Associate Professor Full Professor

Male 10 20 30

Female 30 15 10

Nonbinary 3 2 0

1. State the alternative hypothesis for the study.

__________________________________________________________________________
2. Compute a chi-square test of independence with effect size. Write a couple of sentences
describing your findings as you would in a Results section.
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________
3. Interpret the results for the administrator.
__________________________________________________________________________
__________________________________________________________________________
__________________________________________________________________________

Mann-Whitney U and Kruskal-Wallis H Tests

Conducting a Mann-Whitney U Test

Analyze → Nonparametric Tests → Legacy Dialogs → 2 independent samples…

In the two-independent samples test box:

—— Send the condition or group variable to the “Grouping Variable “ box
—— Send the ordinal outcome variable or DV to the “Test Variable” box
—— Be sure that “Mann-Whitney U” is checked
—— Click “OK”

Conducting a Kruskal-Wallis H Test

Analyze → Nonparametric Tests → Legacy Dialogs → k independent samples…

In the tests for independent-samples box:

—— Send the condition or group variable to the “Grouping Variable “ box
—— Send the ordinal outcome variable or DV to the “Test Variable” box
—— Be sure that “Kruskal-Wallis H” is checked
—— Click “OK”
 IBM® SPSS®
WORKBOOK PART 5
ANALYSES FOR DEPENDENT
GROUPS AND MIXED DESIGNS

Decision Tree For Dependent Groups and Mixed Designs

Decision Tree For Dependent Groups and Mixed Designs

McNemar Test
2 groups
nominal Cochran Q Test
3+ groups

What is Wilcoxon T Test

the scale of 2 groups
measurement ordinal
3+ groups Friedman χ2 Test
for the
dependent-samples t test
outcome/DV?
2 groups

3+ groups one-way within-subjects ANOVA

interval or ratio

2+ dependent factors two-way within-subjects ANOVA

(or three-way, etc.)

dependent and independent factors

Mixed ANOVA

DEPENDENT-SAMPLES t TEST
You compute a dependent-samples (or matched-pairs) t test when you are examining differences between two related
groups. The IV (if you have an experiment) or grouping variable (predictor, if you have a correlational design) is the
related variable, and the DV for an experiment (or outcome, for correlation) is the measured variable.

Assumptions of the Dependent-Samples t Test

1. Two groups that are related due to matching or repeated measures
2. Because the two groups are related, the n of the groups is always equal
3. IV (or grouping variable) is nominal and dichotomous
4. DV (or outcome variable) is interval or ratio
5. DV is normally distributed

248
 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  249

A Test by Any Other Name...

—— SPSS calls this test a paired-samples t test, but this name should only be used in a report if you
had matched (paired) samples.
—— If you have a repeated-measures design, you could use the term repeated-measures t test.
—— Dependent-samples t test or within-subjects t test are more general terms that can apply to either
matched or repeated-measures designs.

Data Entry for Dependent Designs

Data entry for independent and dependent groups look very different. In dependent designs, each
row represents one participant or one matched group.

Independent Groups vs. Dependent Groups

(20 participants) (10 pairs or 10 participants)

When you enter data for a

dependent groups design,
you must enter the data for
each group in a separate
column.

Repeated Measures Example

A researcher is interested in how the environment impacts problem-solving skills. He has two conditions:
a peaceful environment (quiet with a few people reading quietly) and a stressful environment (with loud
noises and lots of people scurrying about).
Each participant experiences both environments and completes a problem-solving test while in each
environment. The researcher randomly assigned participants to sequence so that some experience the
stressful situation first while others experience the nonstressful environment first.
The variable view would look like this:
250   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

The data view would look like this:

In a repeated-measures
design, each row represents
one participant.

In this example, each

participant has a unique ID
number, a problem-solving
score in the peaceful
environment, and a problem-
solving score in the stressful
environment.

Conducting a Dependent-Samples t Test

On the Menu Bar, Click

Analyze → Compare Means → Paired-samples t test

Use the same steps for matched groups and repeated measures.

1. Send each group you are

comparing to the “Paired
Variables” box.
2. Be sure “Estimate effect
sizes” is checked.
3. Click “OK.”

OUTPUT
Paired Samples Statistics
Std. Error
Mean N Std. Deviation Mean The first table provides
descriptive information for
Pair 1 peaceful 6.9167 12 1.24011 .35799
each of your groups.
stressful 6.4167 12 1.31137 .37856

Paired Samples Correlations

N Correlation Sig.
Pair 1 peaceful & stressful 12 .862 .000
 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  251

Paired Samples Test

Paired Differences
95% Confidence Interval of
Std. Std. Error the Difference Sig.
Mean Deviation Mean Lower Upper t df (2-tailed)
Pair 1 peaceful - .50000 .67420 .19462 .07163 .92837 2.569 11 .026
stressful

The third table shows the results for the paired-samples t test. From the example we would report
t(11) = 2.57, p = .026.

Paired Samples Effect Sizes

Point 95% Confidence Interval

Standardizer Estimate Lower Upper
Pair 1 peaceful - Cohen's d .67420 .742 .085 1.37
stressful

The final table provides the effect size. For our example, we would report d = 0.74

Writing up Results
The write-up is very similar to an independent-samples t test (see Part 4). However, you should be
sure your language is consistent with your dependent design, either matched-groups or repeated-
measures.
Also, you do not need to worry about Levene’s test. Include the following in your report:

—— The type of analysis you conducted, and the variables examined

—— The mean and standard deviation for each group.
—— The t value, df, p value, and effect size. For example: t(11) = 2.57, p = .03, d = 0.74.
{{ Remember to round most statistics two decimal places, except p values can be
rounded to two or three places as necessary for appropriate interpretation of statistical
significance.
{{ Remember to use leading zeros for any numbers that are less than 1 but could be higher
than one (t, d, etc.)
—— A statement about whether or not the difference was statistically significant
—— You might also report confidence intervals. Check with your instructor.

Dependent-Samples t Test Practice Exercises

Practice Exercise 1
A researcher interested in dating violence matches participants on their Attitudes toward Women scale
(AWS). Participants read a scenario describing violence (a slap during an argument) between a dating
couple. Matched pairs are randomly assigned to a scenario in which the perpetrator of the violence
is a male or a female. All participants rate the violence of the act (a slap) after reading the scenario
(10 = extremely violent). The data are shown below:
252   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Male Perpetrator Female Perpetrator

9 8

8 9

8 7

7 6

8 8

7 6

6 5

6 4

5 5

5 4

1. What is the design of this study (be specific)? 

2. What is the IV? DV?
3. State a directional alternative hypothesis for the study.

4. Compute using SPSS all the appropriate statistics to test your hypothesis.
5. Are the findings statistically significant? How do you know?

6. What is the probability of a Type I error? Type II error?
7. State and interpret the effect size. 
8. Describe the Results using APA format.






9. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.






 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  253

Practice Exercise 2
Research has found that physicians often ignore complaints in older patients that they take seriously
in younger patients. In an effort to reduce this prejudice before it becomes too ingrained, a researcher
designs a study to have fourth-year medical students see videos of 10 patients (who are really actors) a
month apart. In one sequence of the 10 videos, an actor is made up to look about 70 and talks about
being retired, and in the other sequence of the 10 videos, the same actor is depicted as his own age (30).
The actor presents similar complaints in each video. Students are randomly assigned to the order of
viewing these two sequences of videos a month apart. After viewing each of the videos, the students list
the questions they would ask the patient, their diagnosis of the complaint(s), and their rating for the
health of the “patient” on a 15-point scale, with higher scores representing better health. The researcher
is interested in comparing the responses to the same complaints for the different “age” patient. The data
for these videos are below:

Older patient Younger patient

9.00 14.00

12.00 15.00

9.00 10.00

11.00 11.00

14.00 15.00

10.00 8.00

9.00 10.00

8.00 8.00

12.00 14.00

11.00 15.00

11.00 8.00

9.00 10.00

5.00 9.00

13.00 12.00

10.00 12.00

1. State a directional alternative hypothesis for the study.


2. Compute the statistics needed to test your hypothesis.
3. Are your findings statistically significant? Explain your answer.

4. What type of error might there be (Type I or Type II)?
254   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

5. What potential problem(s) do you identify in the study design?







6. Write a few sentences describing the findings as you would in a Results section.






7. Interpret the findings (what are the implications?).




8. What do you suggest for future researchers to improve upon the study?



 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  255

ONE-WAY WITHIN-SUBJECTS ANOVA

This analysis is similar to the dependent-samples t test, except it allows you to examine the difference
between three or more dependent groups.

—— SPSS calls the analysis repeated measures ANOVA; you can use this term to describe analyses
for a repeated-measures design.
—— Use the term matched groups ANOVA when you used matching.
—— You can use the term dependent-designs ANOVA or within-subjects ANOVA for either design.

Assumptions for the One-Way Within-Subjects ANOVA

1. Groups are dependent (matched or repeated measures)

2. IV or predictor has three or more levels or groups
3. DV or outcome is interval or ratio scale of measurement
4. DV or outcome is normally distributed
5. Sphericity in variances of the differences between pairs of groups

Overview of the One-Way Within-Samples ANOVA

—— This ANOVA analysis will give you an F like the independent-designs ANOVA, but this
ANOVA is calculated as follows:

F = Between group differences (MSA) ÷ Within group differences minus the variability
associated with individuals (MSAxS)

—— When you report an F statistic, you need to report two types of df:

Between-groups (treatment) dfA = Number of groups (k) – 1

Within-groups (interaction error) dfAxS = df for groups times the number of participants per condition
minus one or (k – 1)(kps – 1)

—— You also need to report the effect size. Eta (h) is a correlation coefficient used to describe the
magnitude of a relationship containing two or more levels. Partial eta 2 ( η 2partial ) is the effect
size used with a dependent ANOVA and reflects the proportion of variance in the test variable
(DV) accounted for by the grouping variable (IV), with the error unique to individuals
removed.
—— If you find a statistically significant F, you will need to report the results of post hoc tests to
explain which groups differ significantly.

Matched-Groups Example
A researcher investigates the impact of Internet use on college students’ impulsivity scores. Impulsivity is
measured by the Continuous Performance Task, which is an individually administered test that requires
the participant to attend to and discriminate among stimuli. Higher scores indicate higher impulsivity.
Students are matched on their ability to identify identical objects from a set of similar objects. All
students agree to spend a day (8 hours) in the research lab. One group is given unlimited Internet access,
another group is allowed access for 3 hours, and the third group is given no access.
256   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Data Entry

In a matched-groups
design, each row
represents one matched
group.
In this example, each
matched group has a
unique ID number, and
then impulsivity scores
for each matched
participant in the
unlimited, three hour,
and no internet access
conditions.

Conducting a One-Way Within-Subjects ANOVA

On the Menu Bar, click

Analyze → General Linear Model → Repeated Measures

Use the same steps for matched groups and repeated measures.

Optional: change “factor1” to the name

of your independent or predictor variable
(e.g., “internet” in this example).
1. Enter the number of levels for the IV
or predictor (in this example, we have
three levels).
2. Click “Add.”
3. Click “Define”

4. Send each of your

variables to the
“Within-Subjects
Variables.”
5. Click “Options”

6. Select “Descriptive
statistics” and
 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  257

6. Select “Descriptive
statistics” and
“Estimates of effect
size.”
7. Click “Continue.”

8. Click on “EM Means”

9. Move your factor to

“Display means for:”

10. Select “Compare

main effects.”

11. Select “Bonferroni”

or another post hoc
choice under
“Confidence interval
adjustment.”

12. Click “Continue” and then

“OK” to run analyses

OUTPUT (abbreviated)
We will only look at the following tables in the output:

—— “Descriptive Statistics”
—— “Mauchly’s Test of Sphericity”
—— “Tests of Within-Subjects Effects”
—— “Pairwise Comparisons”

Descriptive Statistics
Mean Std. Deviation N
The “Descriptives Statistics” table
Unlimited 6.2000 1.37321 15 displays the mean and standard
Threehours 5.8667 1.30201 15 deviation for each level of your
factor (IV or predictor).
nointernet 5.3333 1.04654 15
258   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Mauchly’s Test of Sphericitya

Measure: MEASURE_1
Epsilonb
Within Subjects Mauchly’s Approx. Greenhouse- Huynh- Lower-
Effect W Chi-Square df Sig. Geisser Feldt bound
factor1 .735 4.002 2 .135 .791 .874 .500
Tests the null hypothesis that the error covariance matrix of the orthonormalized transformed dependent variables is
proportional to an identity matrix.
a. Design: Intercept Within Subjects Design: factor1
b. May be used to adjust the degrees of freedom for the averaged tests of significance. Corrected tests are displayed in the
Tests of Within-Subjects Effects table.

Like Levene’s test, Mauchly’s test of sphericity tests an assumption, and we hope that the test is NOT
statistically significant. In this example, Mauchly’s test of sphericity is not statistically significant (p ≥ .05)
and we can assume sphericity.
If Mauchly’s test of sphericity fails (i.e., it is statistically significant at p < .05), you would use an alternative
test such as Greenhouse-Geisser or Huynh-Feldt in the “Tests of Within-Subjects Effects” table.

Tests of Within-Subjects Effects

Measure: MEASURE_1
Source Type III
Sum of Mean Partial Eta
Squares df Square F Sig. Squared
factor1 Sphericity Assumed 5.733 2 2.867 5.375 .011 .277
Greenhouse-Geisser 5.733 1.581 3.626 5.375 .018 .277
Huynh-Feldt 5.733 1.749 3.279 5.375 .014 .277
Lower-bound 5.733 1.000 5.733 5.375 .036 .277
Error Sphericity Assumed 14.933 28 .533
(factor1) Greenhouse-Geisser 14.933 22.135 .675
Huynh-Feldt 14.933 24.480 .610
Lower-bound 14.933 14.000 1.067

Use the “Sphericity Assumed” results when Mauchly’s test is not statistically significant.
In this example, F(2, 28) = 5.38, p = .01, ηpartial
2
= .28
Because we have a statistically significant F, we must report the post hoc test results.
 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  259

Pairwise Comparisons
Measure: MEASURE_1
95% Confidence Interval for
Mean Difference Differenceb
(I) factor1 (J) factor1 (I−J) Std. Error Sig.b Lower Bound Upper Bound
1 2 .333 .187 .288 −.175 .841
3 .867* .291 .030 .077 1.656
2 1 −.333 .187 .288 −.841 .175
3 .533 .307 .311 −.300 1.366
3 1 −.867* .291 .030 −1.656 −.077
2 −.533 .307 .311 −1.366 .300

The “Pairwise Comparisons” table depicts the results of the post hoc tests. We only report these
comparisons if our F was statistically significant.
In this example, we used a Bonferroni correction that accounts for the increased chance of a Type I error
due to computing multiple comparisons. Here we find the corrected p values for all our comparisons:
• (1) unlimited access vs. (2) 3-hour access: p = .29
• (1) unlimited access vs. (3) no access: p = .03
• (2) 3-hour access vs. (3) no access: p = .31

Writing up Results
The write-up is very similar to a one-way between-samples ANOVA (see Part 4). However, you should
be sure your language is consistent with your dependent design, either matched-groups or repeated-
measures.
Include the following:

—— The type of analysis you conducted and the variables examined

—— The mean and standard deviation for each group (or report in a table)
—— The results of the F test and the effect size (e.g., F(2, 28) = 5.38, p = .01, η partial = .28)
2

—— A statement about whether or not the difference was statistically significant

—— If the F was statistically significant, identify the type of post hoc tests you ran (e.g., a
Bonferroni correction). Report the results of all the pairwise comparisons, including the p
values
—— You might also report confidence intervals. Check with your instructor.

One-Way Within-Subjects ANOVA Practice Exercises

Practice Exercise 1
A researcher examines the effects of different types of advertisements on persuasiveness (in terms of
buying a product). She matches a sample of 40 young adults on sense of self as a consumer. One student
in each matched group is randomly assigned to rate 10 advertisements that depict people as sex-objects
(erotic ads), people in traditional gender roles (traditional ads), people in nontraditional gender roles
(non-trad ads), or product-oriented with no humans in the ad (control ads). Participants rate the full-
page color advertisements from several magazines on a variety of characteristics, such as color, layout,
attractiveness, and persuasiveness, to distract the students from the true purpose of the experiment. The
researcher is only interested in the 20-point persuasiveness scale (1 = not at all persuasive to 20 = very
persuasive). The results are presented below:
260   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

Erotic Trad Non-Trad Product

15.00 13.00 17.00 15.00

14.00 14.00 16.00 17.00

12.00 12.00 18.00 15.00

16.00 14.00 19.00 15.00

15.00 15.00 17.00 15.00

17.00 16.00 19.00 13.00

14.00 14.00 18.00 14.00

13.00 13.00 17.00 16.00

14.00 13.00 16.00 17.00

16.00 15.00 18.00 16.00

1. What is the design of this study (be specific)?

2. What is the IV? DV?
3. Compute all the appropriate statistics to test your hypothesis.
4. Are the findings significant? How do you know?

5. State and interpret the effect size.
6. Describe the Results using APA format.






7. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.







Practice Exercise 2
A researcher who was examining the effect of different schedules of verbal reinforcement on
correct answers in a vocabulary test randomly assigns the 4-year-olds to the order of reinforcement
 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  261

conditions and records the number of correct responses for each condition. Scores for each
condition are below:

Reinforce only correct

answers (C) Reinforce all answers (B) No Reinforce (N)

8 7 7

7 5 5

6 5 6

6 5 6

7 4 5

6 6 7

8 5 6

7 6 6

5 5 7

6 6 5

1. How many participants are in this study?

2. Enter the data into SPSS (or another statistical package) and calculate all of the appropriate
statistics.
3. Describe the results using APA format as in a Results section.





4. Interpret the results as you would in a Discussion section. Include one limitation and one way
to address the limitation.






262   Student Study Guide With IBM® SPSS® Workbook for Research Methods, Statistics, and Applications

TWO-WAY WITHIN-SUBJECTS ANOVA AND

MIXED ANOVA
On the Menu Bar, click

Analyze → General Linear Model → Repeated Measures

Follow the directions for the one-way within-subjects ANOVA, except:

—— For a two-way within-subjects ANOVA:

{{ Add two within-subject factors in the first box that opens. Or add three factors for a
three-way, etc. Click “Define.”
{{ Send each of your variables to the “within-subjects variables box,” being sure that the
variables align with the factor numbers.
—— For a mixed ANOVA:
{{ Add the dependent factors in the first box that opens, then click “Define.”
{{ Send each of your within-subject variable(s) to the “within-subjects variables” box and
send your between-subject variable(s) to the “between-subjects factor(s)” box.

NONPARAMETRIC STATISTICS
McNemar and Cochran Q Tests
Conducting a McNemar Test

Analyze → Nonparametric Tests → Legacy Dialogs → 2 related samples…

In the two-related samples test box:

—— Select and send variables to the “Test Pairs” box

—— Check “McNemar” under Test Type
—— Click “OK”

Conducting a Cochran Q Test

Analyze → Nonparametric Tests → Legacy Dialogs → k related samples…

In the tests for several related samples box:

—— Select and send the variables to the “Test Variables” box

—— Check “Cochran’s Q” under Test Type
—— Click “OK”

Wilcoxon T and Friedman χ2 Tests

Conducting a Wilcoxon T Test

Analyze → Nonparametric Tests → Legacy Dialogs → 2 related samples…

 IBM® SPSS® Workbook Part 5 ■ Analyses for Dependent Groups and Mixed Designs  263

In the two-related samples test box:

—— Select and send variables to the “Test Pairs” box
—— Check “Wilcoxon” under Test Type
—— Click “OK”

Conducting a Friedman χ2 Test

Analyze → Nonparametric Tests → Legacy Dialogs → k related samples…

In the tests for several related samples box:

—— Select and send the variables to the “Test Variables” box

—— Check “Friedman” under Test Type
—— Click “OK”