How do I find a topic?
Start by thinking of subjects you’re interested in, perhaps based on classes or an internship you really liked. Maybe there’s some recent public policies or economic stories in the news that interest you. The topic should lend itself to a formally stated research question that can be addressed empirically using archival data or theoretically using a mathematical model in some novel way. For empirically based theses, you will be examining the relation between at least one outcome and explanatory mechanism (i.e. individual or entity-level attribute, specific event), controlling for other factors that could influence the relation. Importantly, the outcome, explanatory mechanism, and controlling factors must be quantifiably measurable and there must be some economic theory supporting a hypothesized relation between the explanatory and outcome variables.
How do I know if my research question is novel?
You will want to review the academic literature for related papers. Google Scholar is an efficient way to identify research that exists in the area you’re interested in. Read through the abstracts of related papers to get a feel for what prior research has already found. For closely related papers, you’ll want to read at least the introduction more thoroughly, check out the papers referenced, and identify research that cites the related papers. You may need to access the library’s website for the full version of some articles. The Social Science Research Network ("SSRN") also includes unpublished working papers you can scan for related research. This legwork provides the backbone for the literature review component of your thesis.
It is ok to examine a question that has been previously researched, but you will want to explain what distinguishes your approach. For instance, perhaps you will be employing a distinct sample, time series, variable specification, or regression model. In such cases, be sure to articulate how your unique approach could affect inferences from prior research.
When should I reach out to possible readers?
Preferably once you’ve settled on a research question in the semester prior to writing your thesis. Reaching out in advance can help you gauge your reader’s availability and allow you to get some preliminary feedback on the suitability of your research question. Faculty have constraints as to how many theses they can read in a given semester, so if you wait until the semester you’re writing your thesis to approach readers, the readers you prefer may not be available. It can also give you a head start in collecting data or reviewing related literature to alleviate time pressure during the semester you’ll be writing your thesis.
Should I take ECON 180?
It is recommended. This seminar is required for students pursuing the Financial Economics Sequence, BA/MA, and Honors in Economics, and is highly recommended for others. Your thesis is treated as a one-unit course with a letter grade given by your reader. ECON 180, meanwhile, is a 0.5 unit course graded on a P/NC basis and typically has weekly meetings that can help keep you on track during your thesis semester by setting project milestones and providing additional feedback. Econ 180 must be taken in the semester you submit your thesis.
Where do I find data?
You can manually hand collect data from public or private sources or extract it in machine readable form. For instance, the Wharton Research Data Services ("WRDS") hosts a number of third-party subscriptions containing firm-level financial information. WRDS, along with a number of other databases, can be accessed by contacting the Business Librarian at the Honnold Library. Moreover, government agencies such as US Census Bureau, Bureau of Labor and Statistics, or the Bureau of Economic Analysis all have unique data related to characteristics of the U.S. economy. Additionally, Google dataset also has search tools that allow you to discover some novel data sources. As researchers become more savvy in extracting and quantifying information, the sources for available data should continue to grow. Feel free to be creative in your data sources for this project. For instance, perhaps a summer internship employer will allow you to exploit some unique dataset it maintains. What’s important is that you can objectively measure the economic constructs you’re examining and articulate the variable specification in your thesis.
How do I get started?
Every research project should start with a roughly single-spaced page proposal that serves as a road map for your thesis. It should identify the following:
- What is the research question? You may also want to include some background in this paragraph to set up your question.
- Why is it important? That is, why would anyone be interested in reading your thesis? Oftentimes, this is the most difficult paragraph to write. It’s often helpful to motivate your thesis by explaining how your findings might impact the decision-making of policymakers, consumers, managers, or investors, or the direction of future academic research.
- What is the theory behind the hypothesized relation? Summarize in a couple sentences the economic theory (i.e. market efficiency, loss aversion, price theory) that predicts a certain relation between the two constructs you are examining.
- How will you test this question? Where will the data come from? What regression model will you be estimating and how will you be specifying the key variables?
What should a completed thesis look like?
The thesis should contain at least five sections:
- Introduction (3-5 pages double spaced). This will be an extension of your proposal that additionally summarizes the results and conclusions. Paradoxically, it is usually written after all other sections are finalized.
- Literature Review and hypotheses development (4-8 pages). In this section, provide background on what has already been done and how your question fits with the existing literature. You will want to draw on prior literature to develop your theory that predicts a relation between your explanatory mechanism and outcome. Each hypothesis can be explicitly stated in a single sentence that identifies the predicted directional (i.e. positive or negative) association between your explanatory and outcome variables. If there are competing theories that predict opposite associations, discuss each. You can state your hypothesis in the null (i.e. no association) or use a competing theory to explain why you may not find results consistent with a directional hypothesis.
- Research methods (3-5). For empirically based theses, this section will typically have two subcomponents:
- Sample Selection: discuss the data sources for your thesis. How many observations exist? Where did you have to make cuts in the sample and why? How many observations were lost through these cuts?
- Design: Specify the tests you conduct to address the research question. For instance, what is the regression model? How do you specify the variables? Be explicit as though writing an instruction manual for anyone interested in replicating your work.
- Results (as long as necessary). This section usually starts with discussion of the summary statistics (i.e. mean, median, standard deviation) of your variables. If you have a treatment and control group, you may want to compare differences between them. After discussing the univariate data, you will want to discuss the results of your regression, if applicable. Specifically, you will want to comment how the coefficient on your variable(s) of interest fits with your hypotheses and what inferences can be derived from these results. Consult with your reader how to format tables and whether to insert them within the body of the paper or at the end. Each table should be self-contained and include a title and footnotes at the bottom with explainers (i.e. ***, **, and * indicate significance at the 1%, 5%, and 10% level respectively using two-tailed p-values).
- Conclusion (1-2 pages). In this section, summarize your results, re-emphasize how those results fit with prior work, and explain why the reader should are. You should also discuss the limitations/caveats/weaknesses of your study and discuss possible future extensions. That is, how can future research build on what you have done?
Many theses will also contain an institutional background section if it is necessary to familiarize the reader with unique features of the research setting. Some theses will also include an additional analysis section that explores the robustness of results to research design choices or that shows results of additional tests that were not explicitly hypothesized in the theory development.