Seminar Quantitative Economics Winter 2021/22

Online teaching

The seminar will be offered online. Each week we will have a regular seminar meeting where participants present and discuss.

For this course online teaching seems to worked rather well. Online presentations and online discussions of your research are becoming more and more important. In this seminar students will learn to carry out independent research. Students will also practice their online presentation skills.

Learning aims:

During the seminar participants will learn to

• formulate their own research question,
• develop a laboratory experiment designed to answer this question,
• conduct this experiment with other seminar participants as a pilot,
• analyze the collected data,
• present their results as a talk, as a poster, and as a written essay.

Language:

The language for this seminar will be English.

Prerequisites:

Basic knowledge in empirical work and statistics; interest in scientific work. The Lecture BW24.2 might be helpful but is not a formal requirement.

Assignments:

• Four weeks before the first session (i.e. until calendar week 38 all participants send me the completed registration form with their research question. The purpose of the form is to help you finding a good research question.
• During the first weeks of the seminar students present their own question.
• Later during the seminar students will use different formats to present their progress.
• During the entire seminar students will discuss formally and informally presentations of the other students.
• At the end students will also provide a brief essay, summarising their project.

Research question:

Your research question is usually one sentence, and it is actually a question, i.e. it ends with a question mark. Keep in mind: “The efficiency of mechanisms” is not a question. “Is the ABC mechanism more efficient than the DEF mechanism?” is a question. Ask yourselves: why is an answer to this question relevant for economics (i.e. why does an answer eventually help increase efficiency, reduce inequality, improve stability, etc.).

A main focus of the chair is experimental economics. The research question will be an original question relating to the experimental method. Your question should have the potential to be answered with the help of a laboratory experiment (see question 2 below). Other research methods, e.g. a survey, may be interesting but do not fall into the scope of the seminar.

Your question should also be interesting (see point 5 below).

It is very hard to develop a good research question out of nowhere. Almost always it is easier to start from the existing literature. It also helps if the existing literature is recent (the last five years) and if it is very well published. Let me briefly explain:

• When working on your project, it is important to focus on a specific aspect of the world that you can explain thoroughly. If your research question is not closely related to existing research, you may need to cover a wide range of topics in order to gather enough information. However, by looking at research that is already available, you may be able to find a more focused topic that you can research effectively within a given time frame.
• Starting from the recent literature (last five years) makes it more likely that people find the topic interesting and that you can cover some new ground.
• Starting from a well published paper makes it again more likely that people find the topic interesting.

Read, e.g., current issues of Econometrica, American Economic Review, Review of Economic Studies, etc. and find out what researchers are currently discussing (not all articles in these journals deal with experiments. Do not forget that your question should relate to a laboratory experiment.) If you have the feeling that the current discussion leaves questions aside and that these questions can be answered with an experiment then you might have a starting point.

Alternatively you can look at current experiments. Read the journal Experimental Economics and look for improvements in the experiments that are presented there. You can also take a look at the websites of the GfeW or esa who organize conferences on experimental economics. There you will find abstracts of current research. More details can be found on the websites of the respective authors. Use these opportunities to inspire your research. It is often possible to combine two research ideas to a new research question.

Please make sure that your research question clearly relates to the existing experimental literature (see also point 2 below). Many students start from too ambitious, too complicated problems. We can't do this in the available time.

Once you have a candidate for a research question, the following tests help you to assess the potential of your question:

1. What do we know about answers to your question? How have similar questions already been discussed in the literature? Which gap do you want to close with your study?

   Visit the library to answer this question, use your favourite search engine, have a look at related papers you find at the IDEAS/RepEC database. If your question has already been answered, please find a new question.

   Include, if possible, approaches from other fields of economics (e.g. field data, theory).

   You should find at least one publication from a well ranked journal in economics which raises a similar question. Keep in mind: In your project you have to cover the entire distance between what science knows today and the issue you want to investigate. If the distance between the literature you can find and your research question is too large, you will have too much work to do. If you find references only in other (non-economic) journals, chances are that your topic is not really relevant for economics (the topic might still be relevant for other fields, but your project should be in economics).

2. What alternative possibilities (experimental, theoretical, field data), different from your approach, does one have to find an answer to your question?

   Include approaches from all fields of economics (field data, theory, experiments). Your summary should not be longer than 100 words. Present what you know in a structured way. “While Smith makes a theoretical argument, Miller presents some field data. Additional insight could be gained from a laboratory experiment which studies the influence of GHI on ABC and DEF as follows. If participants with similar GHI levels play both mechanisms in random order...”

3. What are the details of your approach?

   Think about the cornerstones of your approach. “In contrast to the approach of Frank (2004), who just randomizes over different GHI levels, I plan to manipulate GHI with the help of a pre-experimental stimulus as follows...”

4. What are advantages and disadvantages of your approach (compared to theory, field data, experiments)?

   Explain why your approach is more suitable. Again, your summary should not be longer than 100 words. If you find that your approach has more disadvantages than advantages, it might be better to change your approach or move to a different question.

5. Is it possible that your approach yields a surprising answer?

   Here briefly explain why and how different outcomes are possible. “Smith (2001) comes to the conclusion that GHI should affect positively the performance of ABC. If this is the case, ABC should have a clear advantage over DEF. The evidence of Miller (2002) suggests that GHI can also have a negative impact. Then DEF has a chance to outperform ABC...”.

   If you conclude that your question can't have a surprising answer, please find a different question.

6. Is this the simplest possible way to answer your question?

   Please go through all aspects of your approach, e.g. the number of treatments in an experiment, the number of players, etc., and explain why a smaller number of treatments, a smaller number of player types, a simpler game, etc., does not allow to answer your question. If your approach is not the most simple approach, make it more simple.

Schedule

Tue., 8-10 (at least for the first three sessions)

The seminar needs a minimal number of participants for discussion and for experiments. We will, hence, try to merge groups if fewer than 4 participants remain in a group.

The following is an example of how the seminar might develop. We might spend more or less time on the individual items.

• CW 42-43: Participants present and discuss research questions.

  Each presentation should last for about 15 minutes. You may use slides to support your presentation.

• CW 44-45: Participants present revised research questions and idea for experimental design.

  During this time I will present some remarks on presentation (here is a handout with my notes).

• CW 46-47: Experiments are being conducted (this step depends on the number of participants of the seminar).
• CW 48-49: First presentation and discussion of results as a talk.

  In your presentations and in your essay you should cover the following points:

  1. What is the state of the art in your research field? Summarise the relevant literature.
  2. What is your research question and why is it interesting?
  3. What economic experiment do you propose to answer this question? Why did you choose your design and not a different design? Which hypotheses do you want to test with this experiment? How do these hypotheses relate to the literature? Explain the details of your experimental protocol.
  4. Provide a plan for the statistical analysis. Include a power analysis.
  5. Which results did you observe / do you expect to observe in your experiment?
  6. Which conclusions do you draw? In which direction should further research go?
• CW 50-1: Presentation and discussion of results as a poster (A3 format) (here are some comments regarding posters.).
• CW 2-3: Formal discussion of the work of another participant as a talk.
• Until CW 6: Submission of an essay of about 1500 words (+/- 10%, formulae, tables and bibliography do not count). For this essay, keep in mind some remarks that might help you in the preparation of your essay and your presentation
• All assignments count for the grade (This includes presentation of own research as well as discussion of the research of the other students).

Credits:

To pass the seminar, students participate (actively and in a constructive way) in at least 70% of the sessions. Students who miss a session due to health reasons (or any other reason) can present a certificate from their GP within three days, otherwise this session will count as a “fail” (5).

Literature:

• Charles A. Holt; Markets, Games & Strategic Behavior; Pearson, 2007.
• J. H. Kagel and A. E. Roth; The Handbook of Experimental Economics; Princeton University Press, 1995.
• Daniel Friedman, Shyam Sunder; Experimental Methods: A Primer for Economists; Cambridge Univ Press, 1994
• Douglas D. Davis and Charles A. Holt; Experimental Economics; Princeton University Press, 1993.