Synopses & Reviews
Do students who work longer and harder learn more in college? Does joining a fraternity with a more academic flavor enhance a student's academic performance? When are the results from an innovation that is tried on one campus applicable to other campuses? How many students and faculty members must participate in a research project before findings are valid? Do students learn best when they study alone or in small groups? These are just some more than fifty examples that Richard Light Judith Singer and John Willett explore in By Design, a lively nontechnical sourcebook for learning about colleges and universities. These authors believe that careful design of research on college effectiveness is the single most important step toward producing useful and valid findings. In that spirit, By Design is a pathbreaking textbook of modern research methods that both practitioners and students will find useful.
Description
Includes bibliographical references (p. [245]-262).
About the Author
Richard J. Light is Professor in the Graduate School of Education and the John F. Kennedy School of Government at Harvard University.Judith D. Singeris an Associate Professor at the Graduate School of Education, Harvard University.John B. Willett is an Associate Professor at the Graduate School of Education, Harvard University.
Harvard Graduate School of Education, John F. Kennedy School of Government
Table of Contents
1. Why Do Research On Higher Education? Many Questions, Many Options
Our Philosophy of Research Design
How This Book is Organized
2. What Are Your Questions? Why Are Research Questions So Important?
Getting Specific
Building on the Work of Others
Correlation versus Causation
The Wheel of Science
3. What Groups Do You Want to Study? Specifying the Target Population
Where Should you Conduct the Study
Selecting Your Sample
More Than One Type of Respondent
Nonresponse Bias
4. What Predictors Do You Want to Study? Types of Predictors
The Important Role of Variation
Other Reasons for Selecting Predictors
The Integrity of Your Treatment
Choosing Which Predictors to Study
5. Compared to What? Why Do You Need a Comparison Group?
Randomized Control Groups: The Best Comparisons
Requiring Informed Consent
Volunteer Bias
Comparison Groups without Random Assignment
Retrospective Case-control Studies
Design Effects Can Swamp Treatment Effects
6. What Are Your Outcomes? Different Kinds of Outcomes
Will You Measure Status or Development
Short-term versus Long-term Effects
Are Your Measures Valid?
7. How Can You Improve Your Measures? What is Measurement Error?
Reliability and Measurement Error
Six Strategies for Improving Measurement Quality
Looking at Measurement Quality
8. How Many People Should You Study? Why Is Sample Size So Important?
What Size Effect Do You Want to Detect?
What Type of Analysis Will You Use?
Instruments Precision and Sample Size
What If Students Drop Out?
9. Should You Try It out on a Small Scale? The Advantages of Pilot Studies
Piloting Instruments
Relational Studies
Informal Small-scale Experiments
Generalizing From a Small Study
10. Where Should You Go From Here? Getting Started
Lessons From Our Seminar
Decisions You Must Make
Planning a Longer-term Research Program
Reference
Index