Prepare a detailed research proposal for an empirical research project that aims to identify a causal effect of interest using one of the research designs in the course.
A good proposal consists of:
1 A question
2 The contribution
3 Research design
4 Data
Here are guidelines for the proposal. Feel free to consult with me if you want to modify these guidelines for your project.
This part might be the most difficult! Why? It requires one to have a good understanding of the current literature(s) and how this new proposal fits.
How do you get started? Once you have a question, do the following:
Use Google Scholar. Typically the studies with the highest citations will appear at the top of the list. Focus on papers geared towards economists. Papers published in economics journals (“economics” is in the title of the journal) or NBER/IZA working papers are good clues. You should also customize the time range to see if there are any recent papers on the topic (i.e. since 2017). Do not panic if you see a study that is similar to your proposal. There are enough permutations available to make your study a contribution. A common used permutation is to change the outcome of interest. Please discuss with me if you need additional ideas.
Just read the introductions to the papers. An well written econ paper does the following: a) summarize the main findings, b) discuss the research design, and c) convince the reader why the study is an important contribution.
Use a spreadsheet (or reference app like Zotero) to keep track of the studies. I attempt to summarize key findings from each study. Here’s an example. Modify it as you see best.
Draft a proposed contribution section for your proposal. Use the “contribution” sections of the introductions that you have read as a template. See if you can craft a compelling reason for why your proposal is novel and builds on our understanding of the specific topic.
Chat with me for feedback.
Is it an experiment (RCT) or natural experiment (RD, IV, Diff-in-Diff)? (Or, can you make a strong claim that neither is necessary?)
Write down the econometric model you will estimate, describe and define it. Be specific and include all subscripts.
Discuss in detail the identifying assumptions and why you believe they are plausible.
Discuss the ways the identifying assumptions may fail and how that would affect your estimates.
Places to look for data:
The proposal should be approximately 5-7 pages (12-pt Times New Roman Font, 1.5 spacing, 1 inch margins) and should cite major related papers and include a bibliography. (The bibliography is not included in the page count.)
There are three tiers for the research proposal.
Basic: Submit a proposal that addresses the four main topics (question + contribution + research design + data). The research question is clear and illustrated with a DAG. A presentation is prepared.
Complete: Basic + Research Design and Data sections are extensive and well developed. The identifying assumptions have strong rationale and there is a thoughtful discussion on how they may be violated and the bias it would generate. The data section details specifics about the variables that will be used. The data section convinces the reader that obtaining the data is feasible within a month. Note: be aware that certain variables (geographic identifiers at a granular level) may not be easily accessible.
Extensive: Complete + fully developed contribution section that signals to the reader that the author has extensive knowledge of the literature and the contribution(s) the proposal will make. A well written contribution section should read well and be similar to the introduction sections of published econ papers.
I can review your work and give guidance on the likely tier of your proposal. I can also offer suggestions on what can be done to reach the next tier. Please allow up to 5 days for a review.
Note: These guidelines were inspired by Professor Myers.