EDP 380E  |   EDP 482K: Experimental Design and Statistical Inference   |  EDP 380P
Preface
Table of Contents
General Information
Class Schedule
Computer Laboratory
Student Assistance
Homework Solutions

Fundamental Experimental Designs

Gary D. Borich

Department of Educational Psychology

The University of Texas at Austin

______________

*Available from: Able's Reprints, 715 D West 23rd in Towers Court. 472-5353

PREFACE

This text is intended as an introductory handbook of experimental design. Its main purpose is to provide practitioner-oriented readers with an overview and organization of the field of experimental design. In one sense this text illustrates the many and varied forms of the technique known as the analysis of variance, but, in another more practical sense, it acquaints the reader with the most frequently used designs for developing research proposals and analyzing research data in the social and behavioral sciences.

The techniques presented in this text stand apart from others in the field of statistics in that they presume the investigator begins with some theory or rationale that can explain and/or account for the results obtained. These techniques are separate from those that might be used simply to "explore" and describe "happenings" revealed by the data apart from previously stated theory and hypotheses. The relationship between theory, hypotheses and data is an important one for purposes of this text. Theory or some semblance of it in the form of logical arguments, cogent rationale, or observations grounded in practice are used to form hypotheses about the relationships among variables, and these hypotheses, in turn, are used to design a study in which behavioral data are collected and analyzed for the purpose of accepting or rejecting the hypotheses. This orderly chain of events, beginning with theory development, extending to hypothesis generation and, finally, to the collection and analysis of data, has proven to be a practical and efficient method for studying and improving many aspects of our social and behavorial environment.

Although it is natural to refer to this volume as a text, it is neither as comprehensive nor as detailed as other books covering this topic. This is intentional in as much as the author has found some texts on experimental design of limited use to practitioner-oriented students who (1) have poor backgrounds in math and the extensive symbolization so often attendant to it, (2) aspire to become intelligent consumers of the research literature more than researchers, and (3) have career paths and professional programs of study that simply do not leave time to learn more than what is essential to becoming effective practitioners in their chosen fields. This text is accommodating to these concerns in that it avoids some of the detail of other texts in favor of a menu of the most frequently used designs and most practical concepts for understanding experimental design. Not coincidentally, it also serves as an advance organizer for exploring the more comprehensive content that may be found in other more advanced texts and coursework. This has resulted in the exclusion of some content which the professional researcher and quantitative specialist would no doubt find useful. However, these more in-depth topics often are more than what practitioner-oriented readers will need or can master in the time available. The rationale for this text is that it may well be better to study the basics than to know sparingly the details, especially if more advanced applications can be learned "on the job" or from examples and references relevant to a particular problem. The purpose of this text, therefore, is to acquaint the reader with the many options available for designing an experiment in the social and behavioral sciences and to provide an awareness of how these options actually are used to make practical decisions about the effectiveness of an intervention, program or treatment. It is not intended as a substitute for the many excellent and more comprehensive texts on experimental design, some of which are cited at the end of each chapter in this text, or as a substitute for the more advanced study of experimental design.

This text is accompanied by a laboratory manual that instructs the reader how to use computer programs in the Statistical Package for the Social Sciences. In that hand calculation and even the use of the electronic calculator have become impractical for analyzing the complexity of the data suggested by the designs in this text, it is to the student's advantage to learn how to direct the computer to solve data analysis problems, using one of the widely available computer program packages. To this end, I have included in a companion volume instructions and illustrations for using the SPSS "package" to analyze the data resulting from each of the designs presented in this text. These instructions and illustrations provide examples and models for extending the contents of this text to many additional designs which fit the general form of those presented, thereby allowing the reader to tailor the contents of this volume to one's own research interests.

This, then, is the direction of my text. Above all, it has been written to quickly bring readers to a practical, introductory understanding of the field of experimental design and to help them function as applied researchers and intelligent consumers of the research literature.

GDB

Austin, Texas

Back to Top


Table of Contents

1. Review of Introductory Concepts 1

2. Introduction to Experimental Design 29

3. One-Way Analysis of Variance 49

4. Planned Comparisons 73

5. Post Hoc Comparisons 99

6. Factorial Designs 115

7. Higher Order Factorial Designs 145

8. Regression Analysis 171

9. Randomized Blocks Designs 209

10. Repeated Measures Designs 229

11. Nested Designs 249

12. Split-Plot Designs 267

13. Analysis of Covariance 287

14. Survey of Multivariate Techniques 315

Appendix A Table of F 333

Appendix B Table of Studentized Ranges 337

Appendix C Table of Proper Error Mean Squares Random and Mixed Factorial Designs 341

Appendix D Table of Generalized Student Ranges 351

Appendix E Determining Strength of Association in Fixed, Random and Mixed Factorial Designs with Omega Square and the Intraclass Coefficient 355

Back to Top

General Information


Lecture Friday 9:00 - 11:30:

Texts:

Required:

Borich, G. (2006). Fundamental Experimental Designs (Abel's Copies, 715 D, West 23rd, in Towers Court)

Borich, G. (2006). Laboratory Manual for Fundamental Experimental Designs (Abel's Copies, 715 D, West 23rd)

Optional:

Williams, F (2000). Reasoning with Statistics: 5th Edition. New York: Holt, Rinehart & Winston

References:

Kirk, R. E. (1992). Experimental Design: Procedures for the Behavioral Sciences. Belmont. CA: Brooks/Cole

Back to Top


Class Schedule

Session Number, Topic and Text Page/Lab Manual Page

1. Review of Introductory Concepts 1/3

2. Introduction to Experimental Design 29/13

3. One-Way Analysis of Variance 49/29

4. Planned Comparisons 73/33

5. Post Hoc Comparisons 99/37

1st Quiz [10%]

6. Factorial Designs 115/41

7. Higher Order Factorial Designs 145/51

8. Regression Analysis 171/55

[Spring Break]

9. Review

10. Midterm Exam [30%]

11. Randomized Blocks Designs 209/59

12. Repeated Measures Designs 229/63

13. Nested Designs 249/73

14. Split-Plot Designs 267/79

2nd Quiz [10%]

15. Analysis of Covariance 287/85

16. Overview of Some Multivariate Techniques 315/91

17. Review

18. Comprehensive Final Exam [50%]

(The final exam must be taken at the regularly scheduled time.)

Back to Top


Computer Laboratory

Group Laboratory times will be used only for an optional introductory session to the College of Education's Computer Laboratory and two Lab quizzes. Three alternative times can be chosen for each of these group sessions. All other Labs will be at your convenience or completed on your computer at home, provided you have the required software. SPSS (Basic, Advanced, and Professional) software will be used for class illustrations and lectures. Other software that cover the full spectrum of Analysis of Variance Designs, Analysis of Covariance, and Multivariate Analysis of Variance may be used. SPSS software for the Mac and PC is available at the COOP Bookstore.

For those who do not have or chose not to purchase the required software, SPSS is available at the College of Education Computer Lab (SZB 439). Hours are: M-Th. 8:30AM-10:45PM, Fri. 8:30AM-6:45PM, Sat. 9:00AM-6:45PM, Sun. 10:00AM-5:00PM. Statistical data for the lecture and homework problems can be downloaded from Blackboard or obtained from the Lab manual.

Back to Top


Student Assistance

A teaching assistant and the instructor will be available to assist with general questions and the computer analysis of homework problems in EDB 506N or EDB 506M at specified times and by appointment.

Back to Top


Homework Solutions (PDF files)

Solutions to the homework problems can be accessed from the following links:

Back to Top
Last Updated: 1/26/2006