What do Students Learn from a Classroom Experiment: Not much, Unless they Write a Report on it

Abstract

The authors ask whether writing a report on a classroom experiment increases a student's performance in an end-of-course test. To answer this question, the authors analyzed data from a first-year undergraduate course based on classroom experiments and found that writing a report has a large positive benefit. They conclude, therefore, that it is important to constructively integrate classroom experiments with some form of assessment or homework in order to realize the maximum benefit from them.

Keywords:

• assessment
• classroom experiments
• experiment report

JEL codes:

• A2
• C9

Acknowledgments

The authors thank two anonymous referees and the associate editor for their useful suggestions and comments.

Notes

1. The results of Cardell et al. (1996) are harder to explain but may reflect the format of assessment used.

2. For a discussion on the related topic of essay assessment, see Walstad (2005).

3. Students received feedback on their first report before having to hand in the second report.

4. The test paper and full guidelines that students were given for writing a report are available on request from the authors. This is a condensed version of the guidelines: “Following each seminar/experiment, we will put the data from the experiment onto Moodle [the course Web page]. You should use this data in writing your report. For each experiment, you will find below the questions that we would like you to answer in writing your report with the percentage weight that we give to each question. The questions cover the theory behind the experiment, the data, and real-world applications. Note that we are relatively flexible on how you structure your report and what you write about if you demonstrate understanding of the theory, data, and real-world applications relevant to this experiment.” As an illustration of the questions for each experiment, we herewith provide the questions for week 3. “Theory: What do the demand curve for tokens and the supply curve for tokens show? Draw both curves for each of the markets in the seminar. Explain what is market equilibrium and calculate the market equilibrium price and quantity for each of the markets in the seminar. (40%); “Data: Find the average price and quantity sold for each market in the seminar. Was the experimental data consistent with the equilibrium prediction? (40%); Real-world applications: Thinking of the goods that you buy (and sell), discuss some of the different ways that there are for buyers and sellers to meet each other and ‘agree’ on a price. How do these ways differ from what happened in this experiment? (20%).”

5. The attendance data for week 6 was unreliable because the university closed at late notice due to snow. Some seminars did, however, run according to schedule, and so there was sufficient data to write reports.

6. Most previous studies have been based on multiple-choice questions.

7. The coefficients of reports 1 and 2, in tables 3 and 4, remain unchanged if attendance is omitted.

8. Recall that in our course, students who did not attend a seminar could access all the relevant data and material online. A student who does not attend a seminar, therefore, only misses participation in the experiment and the discussion that follows. They do still have the opportunity to look at and work with the data. They also still have the opportunity to attend the lecture where we discuss the data.

9. For example, contrary to the second interpretation, Frank (1997) found that students who watched an experiment did less well than those who participated in the experiment.