In the last few years, a new method of behavioral research has become available, thanks to the World Wide Web. Those who have explored this new medium of research are enthusiastic about its advantages over traditional methods. Virtually anything that can be done in the laboratory using paper and pencil, a slide projector, or a computer can be done via the Web. The advantages of Web research over paper and pencil research in the lab include:
Freedom from the constraints of testing people at a particular time and place, Automatic coding and construction of data files, Opportunity to obtain large and heterogeneous samples, Possibility to conduct cross-cultural research without the expense of traveling, Opportunity to study specific populations of rare conditions, Reduced costs of experimental assistants, and Standardization of experimenter effects.
At the same time, there are new issues with this new method: experimental control and sampling of participants.
Research that compares Web and lab results shows a good convergence between the two methods (Krantz & Dalal, 2000). In fact, some comparisons show that Web data are of higher quality than lab data (Birnbaum, 1999c). Between 1998 and. 1999, the number of studies online doubled.
Even for those investigators who plan to conduct research in the laboratory with college freshmen and sophomores, the Web is a convenient network for the collection of data and also for the publication to other scientists of the exact method by which an experiment was conducted. The advantages in convenience of data processing and for open communication among scientists will cause more and more behavioral research to be done via the Web.
This book presents an introduction to methods and techniques basic to conducting research on the Internet. It also covers problems unique to this form of research and various ways to handle these potential problems. The book also teaches content and methods of data analysis. Each main idea is illustrated by examples included on the companion CD. PREREQUISITES AND ORGANIZATION OF THE BOOK
It is assumed that you know how to use a browser such as Netscape Navigator or Internet Explorer to "surf the net" (explore files on the Web) and use a search engine, such as Yahoo yahoo. It is also assumed that the reader knows how to use a word processor program (such as Microsoft Word or WordPerfect) and a text editor (such as Notepad, SimpleText, or BBEdit). Everything covered in this book has been tested on both a PC and Mac. Everything covered runs under Netscape Navigator (3.0 or above) and almost everything also runs under Internet Explorer (4.5 and above).
The ideas have been presented in a sequence designed to allow a college sophomore enrolled in a course in research methods, a graduate student planning a first-year project, and a professor of psychology experienced in research (but wanting to learn these methods) to profit from the same book.
To get the most out of this book, you should have taken courses in introductory psychology, research methods, and statistics. However, the book is designed to be self-contained and it does not presume this background. Psychological content (logical reasoning in Chapter 7, decision making in Chapter 8, contextual effects in judgment in Chapter 9, personality testing in Chapter 10, impression formation in Chapter 12, social balance theory in Chapter 13, psychophysics in Chapter 14, scaling in Chapter 15, or Bayesian inference in Chapter 16) is presented without the assumption that the reader is knowledgeable in these topic areas. A glossary of terms used heavily in this area of research is also included at the end of the book.
When studying this book, let your computer and the Web be your companions. The CD that accompanies this book contains example HTML, programs, and sets of data. You can follow along with the book by doing the analyses described in the chapters and working with the examples. Many useful links to resources on the Web are also included in the list of examples on the CD. The best way to get started with the examples is to load the file examples.htm in your browser. This file provides links to the chapter examples.
Examples are designed to illustrate both computer techniques and principles of psychological research at the same time. Chapters 1 through 4 teach basic HTML (HyperText Markup Language), use of FTP (File Transfer Protocol) to transfer files, links in HTML, and methods of formatting text and paragraphs in HTML. A novice can study each of these chapters in about an hour each and should spend about an hour practicing the techniques of each chapter on a computer. A college instructor could present each of these chapters in a single 50-minute class or lab session and cover the first four chapters in 2 weeks, assuming that the students will work with the examples and exercises in the lab or as homework. The reader who already knows HTML and FTP can skim Chapters 2 through 4 and proceed to Chapter 5.
Chapter 5 illustrates the use of HTML forms to collect data from people via the Web. A simple experiment is developed that collects a single response from the reader of the page, who can type an answer in a box and push a button to send the datum. This first example illustrates how data can be sent to the experimenter by email. The next example shows how to use a script to append data to a file on the Web server. Three input methods are featured: the text box allows the participant to type a number or short answer, radio buttons allow the reader to click from a multiple choice set, and pull-down selection lists allow the user select one response from a list. An experiment on the classic St. Petersburg Paradox is presented to illustrate a simple method for assigning participants to different conditions and also to illustrate how the choices in a selection list might bias the responses. Chapter 5 requires two 50-minute lectures, and it could be covered in class in 1 week, with students working 2 hours in the lab.
Chapter 6 gives a basic introduction to both Excel and SPSS. Data from the studies in Chapter 5 are analyzed to illustrate the use of Pivot Table Reports in Excel and crosstabs in SPSS. The techniques required to import, filter, and work with data in these programs are described. Procedures for calculating a t-test are also described. An instructor could spend either one or two 20-minute lectures on this chapter, depending on whether the plan was to cover only one or both of these programs. Students would require 2 or 3 hours in the lab or at home to work through the examples in this chapter with their computers.
Chapter 7 explains how to incorporate images and define image maps n HTML files. These ideas are illustrated with an experiment and a debriefing page based on the classic Wason (1960) logic problem. Experimental materials and sample data are included on the CD. The data can be analyzed by either SPSS or Excel using methods described in Chapter 6. This brief chapter would require about 20 to 30 minutes of class time and about the same in the lab.
Chapter 8 presents basic ideas of decision-making research and data analysis. The chapter is intended to make this historical area of psychological research accessible. The chapter should prepare an undergraduate to read current work in the field (Birnbaum, 1999b) and design his or her own Web studies in the field. Both the experimental materials and sample data are included on the CD to facilitate this process. These chapters build on and consolidate ideas presented in Chapters 1 through 6, showing how those ideas alone are sufficient to conduct interesting research. An instructor may choose to present only one of Chapters 7 and 8, depending on the student level. Chapter 8 is the more difficult in terms of psychological content, requiring at least one 50minute lecture. Two lectures are needed if theories beyond the Expected Utility theory will be covered.
Chapter 9 explains how to use surveyWiz (software included on the CD) to make simple surveys and questionnaires. With this software, a person with minimal background can quickly make a simple survey, questionnaire, or personality test that will run on the Web. The chapter also illustrates an example of a between-subjects study of judgments of the "size" of numbers to show how between-subjects designs can lead to the conclusion that 9 is a "bigger" number than 221 (Birnbaum, 1999a). The chapter describes contextual effects in judgment that are important to interpreting results of between-subjects designs.
Chapter 10 presents another example of the use of SurveyWiz showing how to construct and evaluate a personality test. The chapter explains how to make the HTML for the personality test (also included on CD) and how to analyze data from a personality test. Data included on the CD are analyzed by correlational analyses, including factor analysis, to examine the factor structure of the test and the internal consistency of the scales measured.
Chapter 11 explains how to use FactorWiz, additional software included on the