A Research Paper in Experimental Economics

This assignment serves as the summative learning experience in an upper-level experimental economics course. As such, virtually all of the the steps of doing a research paper. are covered in this project. Specific goals related to the course are for students to: (1) design an original experiment, (2) recruit subjects, (3) perform the experiment, (4) collect and analyze data from the experiment and (5) write a coherent research paper.

This course requires that students have completed both intermediate micro and macro. They typically are second-semester juniors or seniors majoring in economics. The course is typically small, with enrollments between 10-20.

Before proposing their topic, student-teams are required to present a chapter out of the text (Holt 2007) in class. Their presentation includes a classroom demonstration of an experiment discussed in that chapter. Based on the chapter they presented in class, teams will subsequently propose an original experiment related to the content from that chapter. During the first 3 weeks in class, I introduce the concept of experimental economics and teach students to perform ANOVAs. While they have had statistics before, they typically do not know or remember how to do ANOVA. ANOVA is the only statistical tool necessary to analyze the data from their experiments as well-designed experiments need only test treatment effects across groups.

Below is a typical timeline for the project (fall semester):

• Oral Proposals – Oct. 28 & Nov. 2
• Implementing and Analyzing Experiments – Nov. 4 through Dec 2
• Final Presentations – Dec 7 & 9
• Final Paper – Dec 9

Unlike most research papers, experiments introduce another level of logistical complication. First, to help them organize their time, I have them submit an action plan (Acrobat (PDF) 257kB Mar27 12).

The most difficult thing is recruiting subjects. In general, I let the students work it out themselves. They contact other economics instructors to see if they can come into their classes and recruit subjects. Money is not used as incentives, but we ask that instructors provide some extra credit to help increase participation. I typically help them by organizing a couple of sessions ("experiment nights") late in the semester where a number of subjects can come and participate in multiple experiments in a short period of time. We also provide pizza funded by the economics department. Here is an example of the information sheet (Acrobat (PDF) 60kB Mar27 12) my students hand out in economics classes in order to recruit students.

The targeted length of the paper is about 10 to 15 pages, but what I'm really looking for is depth. Assessment is based on the extent to which the paper:

1. has a clear statement of the research hypothesis (i.e., a traditional null/alternative breakdown would be a plus).
2. provides some relevant economic theory and evidence that the student read and understood some professional journal articles (i.e., What are the issues/debates in this arena of economic theory? How might an experiment resolve some of these theoretical issues?, etc.).
3. explains the experimental design (i.e., what are the factors? What are the levels of the factors? Did the student use between-subjects or within-subjects? Did the student have a control group? etc.).
4. discusses procedural regularity in the experiments (i.e., were there any unexpected events arising that caused potential noise and error?).
5. uses both descriptive and inferential statistics (particularly ANOVA) to explain and analyze the experimental data.
6. discusses the real-world applications and/or policy recommendations associated with the experiments.

Text: Markets, Games, & Strategic Behavior (2007), by Charles Holt.