Assignment:
Your final paper will be a short research paper on your chosen subject. It will be written for a general economics audience so you do not need to (and should not) bother defining terms like "endogeneity" or "heteroscedasticity". However, you should assume the reader is not intimately familiar with your particular question and data so you should concisely describe them and any notable peculiarities. Your paper should follow the general economics research form and contain the following sections.
1. Title
2. Introduction and Lit Review
3. Data (including sum stats)
4. Model (regression equation)
5. Results
6. Conclusions
7. References
8. Tables and Figures
Your paper must be clear, informative, and concise. The main text of your paper (1-6 above) should be no longer than 2 pages (1 page front, 1 page back) with 2 columns (single spaced 12-point font).
Specific Requirements:
1. Paper should not exceed 2 pages with 2 columns 12-point single spaced font
2. You must cite at least 2 relevant peer-reviewed journal articles in your literature review.
3. You must also include your .do file, which starts with one of the given raw data sets and cleans the data to fit your needs, then performs the analysis. In other words I should be able to run your do file (changing relevant path names) on the data set chosen and get the same results you present and discuss in your paper.
4. Your model should include: one interaction term; one categorical or dummy variables, and one non-linear term
Breakdown of Paper Sections
1. Title. Give your paper a title. It can be something catchy or something straight to the point. Such as..."The relationship between X and Y" with your key "X" and "Y" filled in
2. Introduction. Briefly (1-2 paragraphs) summarize what you are doing, why it is important, what you find, and what it means. You should cite the relevant literature here as extra motivation for the question you are examining.
3. Data. Briefly describe your data set and the particular parts of it you are using (variables, years, subgroups, etc.). Report relevant summary statistics that you also have attached in a graph or table form and anything about them that is particularly interesting.
4. Model. State your econometric model (the population regression model is necessary) and hypothesis tests. Explain why they are appropriate and what the results of the model will tell us when we get to them in the next section.
5. Results. Report the results of your analysis in tables (and graphs, if appropriate). A common way to do this is to start simple (perhaps a bivariate regression) and then add more complicated aspects (more controls, non-linear effects, etc.) explaining why you include the control variables you do, add non-linear relationships, etc. You should note whether your results reject your hypothesis tests from section 4. You should interpret your findings so that the reader understands what they are saying in English (not just "I find that the estimated β1 is 3.5").
6. Conclusion. Summarize the basic results and why they are interesting. Discuss causality and endogeneity. Is there still potential bias? What might be causing the bias and what direction do you think it pushes your estimates? What are some policy implications if the results are causal, if any? Are there any advanced methods you would use if you had the data available (e.g. diff-in-diff, IV, regression discontinuity, etc.)
7. References should follow standard APA style guidelines for journal articles
8. Results should be presented in "clean" tables similar to those seen in journals. A common way to do this would be to use Excel and paste values into the appropriate cells. Copying and pasting full Stata regression output directly into the paper is not accepted.