 Shopping Bag ( 0 items )

All (8) from $91.72

New (2) from $177.92

Used (6) from $91.72
More About This Textbook
Overview
Continuing the Chain of Discovery and Understanding
Beginning Behavioral Research introduces students to the broad base and conceptual underpinnings of basic principles of ethical research with human participants from (a) the development of ideas and testable hypotheses, to (b) the selection of appropriate methods of measurement and data collection, to (c) the design and implementation of empirical research, to (d) the statistical analysis and interpretation of results, and finally, to (e) the reporting of findings and conclusions. The authors emphasize good research, empirical reasoning, and the continuity of psychological science in the continuing cycle of discovery and understanding.
The beginning in the title of this text is intended to have a double meaning. It not only describes the level of the material but also conveys the idea of a journey. For some students, the journey will be completed at the end of the course in which this text is used. For others, the journey will have just begun. In either case, students will have a deeper understanding of the applicability and limits of the scientific method. They will also have learned how to frame questions about the scientific results they read or hear about in the media in ways that will allow them to reach beyond other people’s conclusions or assertions and decide for themselves what is true.
Additional Learning Goals
Upon completing this book, students who are expected to conduct a research study should be able to:
Note: MySearchLab with eText does not come automatically packaged with this text. To purchase MySearchLab with eText, please visit www.mysearchlab.com or you can purchase a ValuePack of the text + MySearchLab with eText (at no additional cost):
ValuePack ISBN10: 0205871895 / ValuePack ISBN13: 9780205871896
The book contains predominantly blackandwhite illustrations, with some color illustrations.
Editorial Reviews
Booknews
A text for undergraduates with no experience in collegelevel statistics, showing how to plan an empirical study, interpret data, and report findings. Features summaries, discussion boxes, key terms, and review questions and answers. Treatment of statistics includes examples, basic computations on a pocket calculator, and coverage of dataanalytic procedures not typically found in statistics software packages. Contains substantial appendices on reports, statistical tables, and metaanalysis. Annotation c. Book News, Inc., Portland, OR (booknews.com)Product Details
Meet the Author
Ralph L. Rosnow is now Thaddeus Bolton Professor Emeritus at Temple University in Philadelphia, PA, where he taught courses in research methods and statistics for many years and directed the Ph.D. program in social and organizational psychology. He also taught research methods at Boston University in a master’s degree program in communication research and at Harvard University as a visiting professor in the psychology department.
http://astro.temple.edu/~rosnow
Robert Rosenthal is a Distinguished Professor at the University of California at Riverside and Edgar Pierce Professor of Psychology, Emeritus, Harvard University. In the realm of statistical data analysis, his special interests are in experimental design and analysis, contrast analysis, and metaanalysis. He served as cochair of the Task Force on Statistical Inference of the American Psychological Association.
http://psych.ucr.edu/faculty/rosenthal
Table of Contents
In this Section:
1. Brief Table of Contents
2. Full Table of Contents
1. BRIEF TABLE OF CONTENTS
PART I GETTING STARTED
Chapter 1 Behavioral Research and the Scientific Method
Chapter 2 From Hunches to Testable Hypotheses
Chapter 3 Ethical Considerations and Guidelines
PART II OBSERVATION AND MEASUREMENT
Chapter 4 Methods of Systematic Observation
Chapter 5 Methods for Looking Within Ourselves
Chapter 6 Reliability and Validity in Measurement and Research
PART III DESIGN AND IMPLEMENTATION
Chapter 7 Randomized Experiments and Causal Inference
Chapter 8 Nonrandomized Research and Causal Reasoning
Chapter 9 Survey Research and Subject Recruitment
PART IV DESCRIBING DATA AND DRAWING INFERENCES
Chapter 10 Summarizing the Data
Chapter 11 Correlating Variables
Chapter 12 Understanding p Values and Effect Size Indicators
PART V STATISTICAL TESTS
Chapter 13 The Comparison of Two Conditions
Chapter 14 Comparisons of More Than Two Conditions
Chapter 15 The Analysis of Frequency Tables
Appendices
Appendix A Reporting Your Research Results
Appendix B Statistical Tables
Appendix C Introduction to MetaAnalysis
2. FULL TABLE OF CONTENTS
Chapter 1: Behavioral Research and the Scientific Method
Preview Questions
Why Study Research Methods and Data Analysis?
What Alternatives Are There to the Scientific Method?
How Do Scientists Use Empirical Reasoning?
How Is Empirical Reasoning Used in Behavioral Research?
How Do Extraempirical Factors Come Into Play?
What Does Behavioral Science Cover?
How Does Research Go From Descriptive to Relational to Experimental?
What Are the Characteristics of Good Researchers?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 2: From Hunches to Testable Hypotheses
Preview Questions
What Is Meant by a Cycle of Discovery and Justification?
What Are HypothesisGenerating Heuristics?
What Is the Potential Role of Serendipity
How Can I Do a LiteratureSearch?
How Should I Go About Defining Variables?
What Identifies “Good” Theories and Working Hypotheses?
What Is the Distinction Between an Independent Variable and Dependent Variable?
What Belongs in My Research Proposal?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 3: Ethical Considerations and Guidelines
Preview Questions
How Do Ethical Guidelines in Research Function?
What Is Informed Consent, and When Is It Used?
How Are Ethics Reviews Done and Acted On?
What Are Obstacles to the Rendering of “Full Justice”?
How Can a “Relationship of Trust” Be Established?
How Do Scientific Quality and Ethical Quality Intertwine?
Is Deception in Research Ever Justified?
What Is the Purpose of Debriefing, and How Is It Done?
How Is Animal Research Governed by Ethical Rules?
What Ethical Responsibilities Are There When Writing Up Research?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 4: Methods of Systematic Observation
Preview Questions
What Is Meant by Systematic Observation?
How Do Researchers Simultaneously Participate and Observe?
What Can Be Learned from Quantifying Observations?
How Are Judgment Studies Done?
How Does Content Analysis Work?
How Are Situations Simulated in Controlled Settings?
What Are Plausible Rival Hypotheses and the ThirdVariable Problem?
What Is the Distinction Between Reactive and Nonreactive Observation?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 5: Methods for Looking Within Ourselves
Preview Questions
What Are the Uses and Limitations of SelfReport Measures?
What Are OpenEnded and FixedChoice Items?
How Are Personality and Projective Tests Used?
What Is Meant By Measuring Implicit Attitudes?
What Are Numerical, ForcedChoice, and Graphic Ratings?
What Are Rating Errors, and How Are They Controlled?
What Is the Semantic Differential Method?
What Are Likert Scales and Thurstone Scales?
How Are Items Prepared for a Questionnaire or an Interview?
How Are FacetoFace and Telephone Interviews Done?
How Are Behavioral Diaries Used in Research?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 6: Reliability and Validity in Measurement and Research
Preview Questions
What Is the Difference Between Validity and Reliability?
What Are Random and Systematic Errors?
What Is the Purpose of Retest and AlternateForm Reliability?
What Is InternalConsistency Reliability, and How Is It Increased?
What Is Acceptable TestRetest and InternalConsistency Reliability?
How Is the Reliability of Judges Measured?
How Is Reliability Related to Replication and External Validity?
How Are Content and Criterion Validity Defined?
How Is Construct Validity Assessed in Test Development?
How Is Construct Validity Relevant to Experimental Design?
What Is the Importance of StatisticalConclusion Validity and Internal Validity?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 7: Randomized Experiments and Causal Inference
Preview Questions
What Is the Purpose of Randomized Experiments?
How Is Random Assignment Accomplished?
What Are BetweenSubjects Designs?
What Is the Formative Logic of Experimental Control
What Are WithinSubjects Designs?
What Are Factorial Designs?
What Is Meant by Counterbalancing the Conditions?
Why Is Causality Said To Be “Shrouded in Mystery”?
How Do Scientists Logically Puzzle Out Efficient Causality?
What Conditions Pose a Threat to Internal Validity?
What Are Artifacts in Research?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 8: Nonrandomized Research and Causal Reasoning
Preview Questions
How Is Causal Reasoning Attempted in the Absence of Randomization?
How Is the ThirdVariable Problem Relevant?
What Is Meant By Subclassification on Propensity Scores?
What Are TimeSeries Designs and “Found Experiments”?
What WithinSubjects Designs Are Used in SingleCase Experiments?
How Are Correlations Interpreted in CrossLagged Panel Designs?
What Is the Purpose of Longitudinal Research Using Cohorts?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 9: Survey Research and Subject Recruitment
Preview Questions
What Are Opportunity and Probability Samples?
What Is Meant by Bias and Instability in Survey Research?
Why Do We Not Know “For Sure” the Bias in Sampling?
How Is Simple Random Sampling Done?
What Are Stratified Random Sampling and Area Probability Sampling?
What Did the Literary Digest Case Teach Pollsters?
What Are Point Estimates and Interval Estimates?
What Are the Benefits of Stratification?
How Is Nonresponse Bias Handled in Survey Research?
What Are the Typical Characteristics of Volunteer Subjects?
How Is Volunteer Bias in Opportunity Samples Managed?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 10: Summarizing the Data
Preview Questions
How Is Visual Integrity Ensured When Results Are Graphed?
How Are Frequencies Displayed in Tables, Bar Graphs, and Line Graphs?
How Do StemandLeaf Charts Work?
How Are Percentiles Used to Summarize Part of a Batch?
How Is an Exploratory Data Analysis Done?
How Does Asymmetry Affect Measures of Central Tendency?
How Do I Measure How “Spread Out” a Set of Scores Is?
What Are Descriptive and Inferential Measures?
How Do I Estimate a Confidence Interval Around a Population Mean?
What Is Distinctive About the Normal Distribution?
Why Are z Scores Called Standard Scores, and How Are They Used?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 11: Correlating Variables
Preview Questions
What Are Different Forms of Correlations?
How Are Correlations Visualized in Scatter Plots?
How Is a ProductMoment Correlation Calculated?
How Is Dummy Coding Used in Correlation?
When Is the Phi Coefficient Used?
How Is a Correlation Calculated on Ranks?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 12: Understanding p Values and Effect Size Indicators
Preview Questions
Why Is It Important to Focus Not Just on p Values?
What Is the Reasoning Behind Null Hypothesis Significance Testing?
What Is the Distinction Between Type I and Type II Error?
What Are OneTailed and TwoTailed p Values?
What Is the Counternull Statistic?
What Is the Purpose of Doing a Power Analysis?
How Do I Estimate a Confidence Interval for an Effect Size Correlation?
What Can Effect Sizes Tell Us of Practical Importance?
What Does Killeen’s p _{rep} Tell Me?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 13: The Comparison of Two Conditions
Preview Questions
What Do SignaltoNoise Ratios Have to Do With t Tests?
How Do I Compute an IndependentSample t Test?
What Can a Table of p Values for t Teach Me?
What Is an Effect Size Index for an IndependentSample t?
How Do I Interpret Cohen’s d for Independent Groups?
How Do I Compute Interval Estimates for Cohen’s d?
How Can I Maximize the IndependentSample t?
How Does a Paired t Test Differ From an IndependentSample t Test?
What Is an Effect Size Index for a Paired t?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 14: Comparisons of More Than Two Conditions
Preview Questions
What Is Analysis of Variance (ANOVA), and How Are F and t Related?
How Is Variability Apportioned in a OneWay ANOVA?
How Are ANOVA Summary Tables Set Up and Interpreted?
How Can I Test for Simple Effects After an Omnibus F?
How Is Variability Apportioned in a TwoWay ANOVA?
How Do I Interpret Main and Interaction Effects?
How Is a TwoWay ANOVA Computed and a Summary Table Set Up?
What Are Contrasts, and How Do I Compute Them On More Than Two Groups?
What Do r _{effect} _{size r} _{alerting} and r _{contrast} Tell Me?
How Are Contrasts on Multiple Repeated Measures Computed?
How Are Latin Square Designs Analyzed?
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Chapter 15: The Analysis of Frequency Tables
Preview Questions
What Is the Purpose of ChiSquare (X ^{2})?
How Do I Compute 1df ChiSquares?
How Do I Obtain the p Value, Effect Size, and Confidence Interval?
What Is the Relationship Between 1df X ^{2} and Phi?
How Do I Deal With Tables Larger Than 2X2?
How Is Standardizing the Margins Done, and What Can It Tell Me?
What Is a Binomial EffectSize Display Used For?
A Journey Begun
Summary of Ideas
Key Terms
MultipleChoice Questions for Review
Discussion Questions for Review
Answers to Review Questions
Appendix A Reporting Your Research Results
Research Reports in APA Style
Getting Started
Title Page
Abstract
Introduction
Method
Results
Discussion
References
Footnotes
Tables and Figures
Appendix
Writing and Revising
Appendix B Statistical Tables
B.1. z Values and Their Associated OneTailed p Values
B.2. t Values and Their Associated OneTailed and TwoTailed p Values
B.3. F Values and Their Associated p Values
B.4. r^{2} Values and Their Associated p Values
B.5. r Values and Their Associated p Values
B.6. Transformations of r to Fisher zr
B.7. Transformations of Fisher zr to r
Appendix C Introduction to MetaAnalysis
The Purpose of MetaAnalysis
Some Pro and Con Arguments
Comparing Two Effect Sizes
Combining Two Effect Sizes
Obtaining an Overall Significance Level
DetectiveLike Probing of Reported Data
The File Drawer Problem
Glossary of Terms
References
Name Index
Subject Index
Preface
Welcome to the fourth edition of Beginning Behavioral Research. This book was originally conceived of as an undergraduate text for students who, as part of an introductory course in research methods, are required to plan an empirical study, to analyze and interpret the data, and to present their findings and conclusions in a written report. It is also intended to encourage students to be analytical and critical not only in interpreting their own research findings, but also in seeing what is behind the claims and conclusions in newspaper, TV, and Internet reports of scientific and pseudoscientific results and claims. Boyce Rensberger (2000), a prominent science journalist and author of several popular science books, commented recently on how relatively undiscriminating the general public is about science and pseudoscience:
difference between science based on real data and something that resembles science
—at least in their eyes—but is based on uncontrolled experiments, anecdotal
evidence, and passionate assertions. They like it all. (p. 61)
Our hope is that this book will teach students to see that what defines science is, as Rensberger went on to say, that "evidence has to meet certain standards" (p. 61).
Thus, although the primary emphasis here is on behavioral research, we have tried to connect this approach with the empirical reasoning used in other fields in order to underscore the unity of scientific thinking. The examples we have chosen give a sense not only of traditional ways of doing, analyzing, andthinking about research, but also of some recent developments that may not be as well known. For example, we introduce students to statistical methods that, although enormously useful, do not yet generally appear in undergraduate methods texts: metaanalysis, contrast analysis, interval estimates of effect sizes and their practical interpretation, and so on. The emphasis of these discussions is intended to resonate with the spirit and substance of the guidelines recommended by the American Psychological Association's Task Force on Statistical Inference (Wilkinson et al., 1999).
Although this book was conceived of as an undergraduate text for students assigned to do a research project, we have been pleasantly surprised to learn that it has also been successfully used in ways that go far beyond its original purpose. For example, it has been used in undergraduate courses in which the production of a research project was not a major goal, as well as by master's and doctoral students to slip into our advanced text, Essentials of Behavioral Research (Rosenthal & Rosnow, 1991). Beginning Behavioral Research has also been used to teach research methods and data analysis to several thousand students in distance learning programs. We are gratified that the book has been found useful by so many.
Organization
As in earlier editions, the chapters in this edition are presented in a linear sequence corresponding to the steps involved in conducting an empirical research study and analyzing and reporting the results. The beginning researcher is led step by step through the following process:
Understanding empirical reasoning, the scientific method, levels of empirical investigation, and the scientific outlook (Chapter 1); creating, shaping, and polishing a research idea, and conducting a search of the relevant literature for the research proposal and the project itself (Chapter 2); weighing and balancing ethical considerations, and preparing for an ethics review (Chapter 3)
Knowing what methods are available for watching and recording behavior in laboratory and field research, using archival data and outside observers, and preparing a research proposal (Chapter 4); collecting data in which the participants describe their own behavior or state of mind (Chapter 5); assessing the reliability and validity of measuring instruments and research de signs (Chapter 6)
Designing a randomized experiment while controlling for artifacts arid other threats to validity (Chapter 7); using timeseries, singlecase, longitudinal, and correlational designs (Chapter 8); surveying opinions and behavior, controlling for selfselection bias, and testing the methods and instruments (Chapter 9)
Using graphics and statistical summary procedures to develop an overall picture of the results (Chapter 10); identifying relationships (Chapter 11);testing hypotheses, estimating effect size, creating a confidence interval around the obtained effect, using the BESD to interpret practical importance, and doing a power analysis (Chapter 12)
Using t to compare two independent or two correlated conditions (Chapter 13); computing F in oneway and twoway designs. examining the simple effects, and interpreting an obtained interaction (Chapter 14); analyzing smaller and larger tables of counts by the chisquare and other procedures (Chapter 15)
Writing up the findings and conclusions in the style recommended in the Publication Manual of the American Psychological Association (4th edition) and developing a poster (Appendix A)
Our Approach
In our long experience of teaching research methods and statistical data analysis, we have noted the questions and uncertainties of undergraduate students engaged in empirical research for the first time. The vast majority of students have not planned to pursue a career in research, but most of them have recognized the vitality and ubiquitousness of scientific research in their daily lives. So we have tried to anticipate and confront questions and uncertainties from their perspective not as potential professional producers of research, but as intelligent consumers of scientific results. For example, we describe procedures that can also be used to analyze the results reported in published articles (including, in some cases, the meager ingredients in news stories), and thus possibly to reach beyond the original researcher's (or the journalist's) published conclusion or interpretation. It is also essential for educated consumers to understand the limitations of particular research methods, and therefore, at the same time that we explain the utility of various methods, we also mention some of their boundaries. We are not wedded to any single scientific method, theory, or unit of analysis, and indeed, the mantra of our approach in this book is methodological pluralism and theoretical ecumenism.
Instructors who know our earlier work will recognize that this book—as well as our more advanced text (Essentials of Behavioral Research)—grew out of a 117page paperback book that we wrote many years ago: Primer of Methods for the Behavioral Sciences (Rosenthal & Rosnow, 1975a). Over the intervening period, we have developed and refined that material. Most of our undergraduate students have been psychology majors required to take a research methods course as part of their concentration, but a substantial number have majored in fields as diverse as communications, computer science, physical education, mathematics, statistics, accounting, nursing, biology, education, sociology, marketing, and even English, art, and theology. Whether they took this course as part of their major or as an elective, many dreaded the thought of having to wrestle again with statistics. On the assumption that few readers have total recall of statistics or will have come away from a statistics course with an intuitive understanding of what was taught, we review basic aspects of data analysis procedures, purposely avoiding the use of any mathematics beyond the high school level. We focus primarily on the most popular procedures (t, F, and chisquare), but in a way that can also be used outside a research methods course to examine the practical importance of a set of results.
Most students with no college training in statistics should find that they can master basic dataanalytic skills by reading the chapters and repeating the exercises in the order in which they are presented. In this age of the computer, the speediest method of doing complex calculations is with the aid of a computer; yet, as statistician John W Tukey (1977) noted, much can be learned by simply changing our point of view and examining the data in different ways (e.g., exploring for moderator variables by using the stemandleaf procedure). Our philosophy of data analysis is to treat statistics (in Chapters 1015 and Appendix C) by showing, through intuitive reasoning and simple examples, what the results tell us. Instructors who plan to teach students to perform their main calculations on a computer will find that our emphasis on the concrete and arithmetical aspects of data analysis will complement any statistics package they choose. As we implied, we also describe useful dataanalytic procedures that may not yet be available in basic computer packages (e.g., the effect size correlation and the confidence interval around the effect, the method of standardizing the margins in chisquare tables, the isolation of interaction residuals, the detectivelike probing of reported data for an unreported effect size, new measures of effect sizes using the correlation coefficient, and the filedrawer method of figuring out the robustness of an overall p value in metaanalysis), but that can be easily computed on a calculator or looked up in one of the tables in this text.
Instructors familiar with Essentials of Behavioral Research will recognize that Beginning Behavioral Research can be used for students up to, but just below, the level of Essentials, and that the conceptual and philosophical treatment of methods and data analysis is similar in both texts. We again emphasize the utility of the Pearson r as a general effect size measure that can be conveniently interpreted as one indicator of practical importance. We also introduce students to statistical power analysis in a way that many should be able to apply in their individual projects, and that will also give them a sense of why it is important not to confuse statistical significance with practical importance. In Appendix C, we give students a flavor of the use of contrasts to address focused questions in betweengroup designs with more than two groups or conditions, and we also introduce them to a family of new correlational effectsize indices for use with such procedures (Rosenthal, Rosnow, & Rubin, 2000; Rosnow, Rosenthal, & Rubin, 2000). Chapter 3, on ethics, draws on recently published guidelines (Sales & Folkman, 2000) and also raises a number of questions that are intended to get the student thinking about ethical issues that project well beyond this book. Students who are interested in pursuing any of these topics will find more detailed discussions in books and articles that are also mentioned. Throughout this book, we have also sought to communicate the richness, diversity, and excitement of research that we ourselves find so challenging and stimulating.
Other Features and Additions
In an effort to make this book more useful and userfriendly to a wide variety of students, we have incorporated various pedagogical devices. Each chapter begins with a set of preview questions, which students can refer to as they progress. Box discussions highlight and enliven concepts with practical examples and illustrations. Each chapter ends with a summary of the main ideas, followed by a list of key terms pegged to particular pages, and finally a number of review questions to stimulate thought and discussion, with the answers located at the end of the chapter. A glossary at the end of the book lists and defines all the key terms in boldface in the chapters and appendixes and notes the primary location where each term is discussed. There is also an Instructor's Manual, prepared by David B. Strohmetz of Monmouth University, and a set of diskettes, which are available to instructors from your Prentice Hall representative. New to this edition is a Web site connected to Prentice Hall's Web site, which provides students with study aids for each chapter and convenient links to other useful Internet resources.
There are also a number of other features that are new to this edition. For example, the sample report in Appendix A has been rewritten to reflect the most recent APA guidelines for the reporting of research, and the data have been reanalyzed to introduce students to the assumptions of statistical tests (corresponding to the brief discussion in the chapter on t tests). Some material has been shifted; for example, the (rewritten) sample proposal has been moved to Chapter 2, in order to give a preview of what is expected. Appendix A now also contains a sample poster that is based on the research project. The chapters intentionally refer back and forth to ideas, so that connections are emphasized and built upon. In Chapter 2, the discussion of how to do a literature review has been rewritten to reflect the evolution of PsycINFO, PsycLIT, and other recent developments (Rosnow & Rosnow, 2001).
The APA's publication manual (American Psychological Association, 1994) encourages authors of research reports to provide effect size information, a recommendation underscored in the report of the APA's Task Force on Statistical Inference (Wilkinson et a1.,1999).We describe how to compute and interpret the effect size correlation between an independent and a dependent variable, symbolized as r_{effect size} (introduced in Chapter 12). In reporting r_{effect size}, the convention is to represent it as positive when the observed effect is in the predicted or hypothesized direction and as negative when it is in the opposite direction. The effect size is reported only in the case of focused tests of significance (i.e., any t test, F tests with numerator df= 1, and X^{2} with 1 df), since meaningful effect sizes cannot be computed for omnibus tests of significance (e.g., F tests with numerator df > 1 or X^{2} with df > 1). Focused statistical tests can, however, be performed on more than two groups, and in Appendix C, we illustrate how to compute the t_{contrast} and F_{contrast} and their associated effect size correlations.
Acknowledgments
We have benefited once again from working with David Strohmetz, who prepared the Instructor's Manual; Margaret Ritchie, who did the copy editing; and Mary Lu Rosenthal, who prepared the indexes. We are grateful for their creative, elegant, and helpful assistance. We thank Bruce Rind for again allowing us to include an edited version of his work in Appendix A. We thank Robert E. Lana for permission to borrow or adapt ideas from Introduction to Contemporary Psychology (Lava & Rosnow, 1972). We thank a long line of excellent teaching assistants and students at Temple University, the University of CaliforniaRiverside, and Harvard University for their valuable comments on and criticisms of the lectures, handouts, drafts, and earlier editions on which this fourth edition was based. We thank the following reviewers of one or more editions of this book for their constructive feedback: Bernard C. Beins, Ithaca College; B. LaConyea Butler, Spelman College; Patricia R. DeLucia, Texas Tech University; Paul W. Foos, University of North Carolina at Charlotte; Allan J. Kimmel, Groupe tcole Superieure de Commerce de Paris; Wilson McDermut, William Paterson College; Anthony Uzwiak, Rutgers University; John W Webster, Towson State University; Paul J. Wellman, Texas A & M University; and Jon L. Williams, Kenyon College. We thank Jayme Heffler and April Dawn Klemm at Prentice Hall for their editorial support. We also thank our Senior Production Editor, Shelly Kupperman, for her invaluable contributions. And finally, we thank Mimi Rosnow and Mary Lu Rosenthal for counseling us in ways too numerous to mention.
Certain tables, figures, and passages (specifically noted in the text) have by permission been reproduced in part or in their entirety, for which we thank the following authors, representatives, and publishers: E. Earl Baughman; Leonard Berkowitz; Jacob Cohen; Mihaly Csikszentmihalyi; J. A. Hagenaars; R. Vance Hall; Howard Kahane; Paul Slovic; Alan Sockloff; Laurence Steinberg; Robert Weisberg; Academic Press; American Association for the Advancement of Science; American Psychological Association; American Sociological Association; American Statistical Association; Biometrika Trustees of the Imperial College of Science, Technology & Medicine; Brooks/ Cole Publishing Company; Cambridge University Press; Elsevier Science Publishers; HarperCollins Publishers; Helen Dwight Reid Educational Foundation and Heldref Publications; Holt, Rinehart & Winston; Houghton Mifflin Company; Iowa State University Press; Journal of Applied Behavior Analysis; Lawrence Erlbaum Associates, Inc.; McGrawHill, Inc.; W W Norton & Company, Inc.; Oxford University Press; Pergamon Press; The Rand Corporation; Sussex Publishers, Inc., and Psychology Today Magazine; The University of Chicago Press; Wadsworth Publishing Company; and John Wiley & Sons, Inc. We are also grateful to the Longman Group UK Ltd., on behalf of the Literary Executor of the late Sir Ronald Fisher, ER.S., and Dr. Frank Yates, ER.S., for permission to reprint Table V from Statistical Tables for Biological, Agricultural, and Medical Research (6th ed., 1974).
Ralph L. Rosnow
Robert Rosenthal
Introduction
difference between science based on real data and something that resembles science
—at least in their eyes—but is based on uncontrolled experiments, anecdotal
evidence, and passionate assertions. They like it all. (p. 61)
Our hope is that this book will teach students to see that what defines science is, as Rensberger went on to say, that "evidence has to meet certain standards" (p. 61).
Thus, although the primary emphasis here is on behavioral research, we have tried to connect this approach with the empirical reasoning used in other fields in order to underscore the unity of scientific thinking. The examples we have chosen give a sense not only of traditional ways of doing,analyzing, and thinking about research, but also of some recent developments that may not be as well known. For example, we introduce students to statistical methods that, although enormously useful, do not yet generally appear in undergraduate methods texts: metaanalysis, contrast analysis, interval estimates of effect sizes and their practical interpretation, and so on. The emphasis of these discussions is intended to resonate with the spirit and substance of the guidelines recommended by the American Psychological Association's Task Force on Statistical Inference (Wilkinson et al., 1999).
Although this book was conceived of as an undergraduate text for students assigned to do a research project, we have been pleasantly surprised to learn that it has also been successfully used in ways that go far beyond its original purpose. For example, it has been used in undergraduate courses in which the production of a research project was not a major goal, as well as by master's and doctoral students to slip into our advanced text, Essentials of Behavioral Research (Rosenthal & Rosnow, 1991). Beginning Behavioral Research has also been used to teach research methods and data analysis to several thousand students in distance learning programs. We are gratified that the book has been found useful by so many.
Organization
As in earlier editions, the chapters in this edition are presented in a linear sequence corresponding to the steps involved in conducting an empirical research study and analyzing and reporting the results. The beginning researcher is led step by step through the following process:
Understanding empirical reasoning, the scientific method, levels of empirical investigation, and the scientific outlook (Chapter 1); creating, shaping, and polishing a research idea, and conducting a search of the relevant literature for the research proposal and the project itself (Chapter 2); weighing and balancing ethical considerations, and preparing for an ethics review (Chapter 3)
Knowing what methods are available for watching and recording behavior in laboratory and field research, using archival data and outside observers, and preparing a research proposal (Chapter 4); collecting data in which the participants describe their own behavior or state of mind (Chapter 5); assessing the reliability and validity of measuring instruments and research de signs (Chapter 6)
Designing a randomized experiment while controlling for artifacts arid other threats to validity (Chapter 7); using timeseries, singlecase, longitudinal, and correlational designs (Chapter 8); surveying opinions and behavior, controlling for selfselection bias, and testing the methods and instruments (Chapter 9)
Using graphics and statistical summary procedures to develop an overall picture of the results (Chapter 10); identifying relationships (Chapter 11);testing hypotheses, estimating effect size, creating a confidence interval around the obtained effect, using the BESD to interpret practical importance, and doing a power analysis (Chapter 12)
Using t to compare two independent or two correlated conditions (Chapter 13); computing F in oneway and twoway designs. examining the simple effects, and interpreting an obtained interaction (Chapter 14); analyzing smaller and larger tables of counts by the chisquare and other procedures (Chapter 15)
Writing up the findings and conclusions in the style recommended in the Publication Manual of the American Psychological Association (4th edition) and developing a poster (Appendix A)
Our Approach
In our long experience of teaching research methods and statistical data analysis, we have noted the questions and uncertainties of undergraduate students engaged in empirical research for the first time. The vast majority of students have not planned to pursue a career in research, but most of them have recognized the vitality and ubiquitousness of scientific research in their daily lives. So we have tried to anticipate and confront questions and uncertainties from their perspective not as potential professional producers of research, but as intelligent consumers of scientific results. For example, we describe procedures that can also be used to analyze the results reported in published articles (including, in some cases, the meager ingredients in news stories), and thus possibly to reach beyond the original researcher's (or the journalist's) published conclusion or interpretation. It is also essential for educated consumers to understand the limitations of particular research methods, and therefore, at the same time that we explain the utility of various methods, we also mention some of their boundaries. We are not wedded to any single scientific method, theory, or unit of analysis, and indeed, the mantra of our approach in this book is methodological pluralism and theoretical ecumenism.
Instructors who know our earlier work will recognize that this book—as well as our more advanced text (Essentials of Behavioral Research)—grew out of a 117page paperback book that we wrote many years ago: Primer of Methods for the Behavioral Sciences (Rosenthal & Rosnow, 1975a). Over the intervening period, we have developed and refined that material. Most of our undergraduate students have been psychology majors required to take a research methods course as part of their concentration, but a substantial number have majored in fields as diverse as communications, computer science, physical education, mathematics, statistics, accounting, nursing, biology, education, sociology, marketing, and even English, art, and theology. Whether they took this course as part of their major or as an elective, many dreaded the thought of having to wrestle again with statistics. On the assumption that few readers have total recall of statistics or will have come away from a statistics course with an intuitive understanding of what was taught, we review basic aspects of data analysis procedures, purposely avoiding the use of any mathematics beyond the high school level. We focus primarily on the most popular procedures (t, F, and chisquare), but in a way that can also be used outside a research methods course to examine the practical importance of a set of results.
Most students with no college training in statistics should find that they can master basic dataanalytic skills by reading the chapters and repeating the exercises in the order in which they are presented. In this age of the computer, the speediest method of doing complex calculations is with the aid of a computer; yet, as statistician John W Tukey (1977) noted, much can be learned by simply changing our point of view and examining the data in different ways (e.g., exploring for moderator variables by using the stemandleaf procedure). Our philosophy of data analysis is to treat statistics (in Chapters 1015 and Appendix C) by showing, through intuitive reasoning and simple examples, what the results tell us. Instructors who plan to teach students to perform their main calculations on a computer will find that our emphasis on the concrete and arithmetical aspects of data analysis will complement any statistics package they choose. As we implied, we also describe useful dataanalytic procedures that may not yet be available in basic computer packages (e.g., the effect size correlation and the confidence interval around the effect, the method of standardizing the margins in chisquare tables, the isolation of interaction residuals, the detectivelike probing of reported data for an unreported effect size, new measures of effect sizes using the correlation coefficient, and the filedrawer method of figuring out the robustness of an overall p value in metaanalysis), but that can be easily computed on a calculator or looked up in one of the tables in this text.
Instructors familiar with Essentials of Behavioral Research will recognize that Beginning Behavioral Research can be used for students up to, but just below, the level of Essentials, and that the conceptual and philosophical treatment of methods and data analysis is similar in both texts. We again emphasize the utility of the Pearson r as a general effect size measure that can be conveniently interpreted as one indicator of practical importance. We also introduce students to statistical power analysis in a way that many should be able to apply in their individual projects, and that will also give them a sense of why it is important not to confuse statistical significance with practical importance. In Appendix C, we give students a flavor of the use of contrasts to address focused questions in betweengroup designs with more than two groups or conditions, and we also introduce them to a family of new correlational effectsize indices for use with such procedures (Rosenthal, Rosnow, & Rubin, 2000; Rosnow, Rosenthal, & Rubin, 2000). Chapter 3, on ethics, draws on recently published guidelines (Sales & Folkman, 2000) and also raises a number of questions that are intended to get the student thinking about ethical issues that project well beyond this book. Students who are interested in pursuing any of these topics will find more detailed discussions in books and articles that are also mentioned. Throughout this book, we have also sought to communicate the richness, diversity, and excitement of research that we ourselves find so challenging and stimulating.
Other Features and Additions
In an effort to make this book more useful and userfriendly to a wide variety of students, we have incorporated various pedagogical devices. Each chapter begins with a set of preview questions, which students can refer to as they progress. Box discussions highlight and enliven concepts with practical examples and illustrations. Each chapter ends with a summary of the main ideas, followed by a list of key terms pegged to particular pages, and finally a number of review questions to stimulate thought and discussion, with the answers located at the end of the chapter. A glossary at the end of the book lists and defines all the key terms in boldface in the chapters and appendixes and notes the primary location where each term is discussed. There is also an Instructor's Manual, prepared by David B. Strohmetz of Monmouth University, and a set of diskettes, which are available to instructors from your Prentice Hall representative. New to this edition is a Web site connected to Prentice Hall's Web site (www.prenhall.com/rosnow), which provides students with study aids for each chapter and convenient links to other useful Internet resources.
There are also a number of other features that are new to this edition. For example, the sample report in Appendix A has been rewritten to reflect the most recent APA guidelines for the reporting of research, and the data have been reanalyzed to introduce students to the assumptions of statistical tests (corresponding to the brief discussion in the chapter on t tests). Some material has been shifted; for example, the (rewritten) sample proposal has been moved to Chapter 2, in order to give a preview of what is expected. Appendix A now also contains a sample poster that is based on the research project. The chapters intentionally refer back and forth to ideas, so that connections are emphasized and built upon. In Chapter 2, the discussion of how to do a literature review has been rewritten to reflect the evolution of PsycINFO, PsycLIT, and other recent developments (Rosnow & Rosnow, 2001).
The APA's publication manual (American Psychological Association, 1994) encourages authors of research reports to provide effect size information, a recommendation underscored in the report of the APA's Task Force on Statistical Inference (Wilkinson et a1.,1999).We describe how to compute and interpret the effect size correlation between an independent and a dependent variable, symbolized as r_{effect size} (introduced in Chapter 12). In reporting r_{effect size}, the convention is to represent it as positive when the observed effect is in the predicted or hypothesized direction and as negative when it is in the opposite direction. The effect size is reported only in the case of focused tests of significance (i.e., any t test, F tests with numerator df= 1, and X^{2} with 1 df), since meaningful effect sizes cannot be computed for omnibus tests of significance (e.g., F tests with numerator df > 1 or X^{2} with df > 1). Focused statistical tests can, however, be performed on more than two groups, and in Appendix C, we illustrate how to compute the t_{contrast} and F_{contrast} and their associated effect size correlations.
Acknowledgments
We have benefited once again from working with David Strohmetz, who prepared the Instructor's Manual; Margaret Ritchie, who did the copy editing; and Mary Lu Rosenthal, who prepared the indexes. We are grateful for their creative, elegant, and helpful assistance. We thank Bruce Rind for again allowing us to include an edited version of his work in Appendix A. We thank Robert E. Lana for permission to borrow or adapt ideas from Introduction to Contemporary Psychology (Lava & Rosnow, 1972). We thank a long line of excellent teaching assistants and students at Temple University, the University of CaliforniaRiverside, and Harvard University for their valuable comments on and criticisms of the lectures, handouts, drafts, and earlier editions on which this fourth edition was based. We thank the following reviewers of one or more editions of this book for their constructive feedback: Bernard C. Beins, Ithaca College; B. LaConyea Butler, Spelman College; Patricia R. DeLucia, Texas Tech University; Paul W. Foos, University of North Carolina at Charlotte; Allan J. Kimmel, Groupe tcole Superieure de Commerce de Paris; Wilson McDermut, William Paterson College; Anthony Uzwiak, Rutgers University; John W Webster, Towson State University; Paul J. Wellman, Texas A & M University; and Jon L. Williams, Kenyon College. We thank Jayme Heffler and April Dawn Klemm at Prentice Hall for their editorial support. We also thank our Senior Production Editor, Shelly Kupperman, for her invaluable contributions. And finally, we thank Mimi Rosnow and Mary Lu Rosenthal for counseling us in ways too numerous to mention.
Certain tables, figures, and passages (specifically noted in the text) have by permission been reproduced in part or in their entirety, for which we thank the following authors, representatives, and publishers: E. Earl Baughman; Leonard Berkowitz; Jacob Cohen; Mihaly Csikszentmihalyi; J. A. Hagenaars; R. Vance Hall; Howard Kahane; Paul Slovic; Alan Sockloff; Laurence Steinberg; Robert Weisberg; Academic Press; American Association for the Advancement of Science; American Psychological Association; American Sociological Association; American Statistical Association; Biometrika Trustees of the Imperial College of Science, Technology & Medicine; Brooks/ Cole Publishing Company; Cambridge University Press; Elsevier Science Publishers; HarperCollins Publishers; Helen Dwight Reid Educational Foundation and Heldref Publications; Holt, Rinehart & Winston; Houghton Mifflin Company; Iowa State University Press; Journal of Applied Behavior Analysis; Lawrence Erlbaum Associates, Inc.; McGrawHill, Inc.; W W Norton & Company, Inc.; Oxford University Press; Pergamon Press; The Rand Corporation; Sussex Publishers, Inc., and Psychology Today Magazine; The University of Chicago Press; Wadsworth Publishing Company; and John Wiley & Sons, Inc. We are also grateful to the Longman Group UK Ltd., on behalf of the Literary Executor of the late Sir Ronald Fisher, ER.S., and Dr. Frank Yates, ER.S., for permission to reprint Table V from Statistical Tables for Biological, Agricultural, and Medical Research (6th ed., 1974).
The Web sites listed for each chapter were carefully selected by David Strohmetz, author of the Instructor's Manual prepared to accompany this edition of Beginning Behavioral Research. Any questions about these Web sites should be addressed to him at dstrohme@monmouth.edu.
This is our 14th book together in a collaboration that began over 35 years ago, and the beat goes on!
Ralph L. Rosnow
Robert Rosenthal