辅导 Econ 2560, Fall 2024, Paper Assignment

Paper Assignment

Econ 2560, Fall 2024

Over the course of the semester, you will conduct independent statistical analysis on a topic of your choosing, which will culminate in a write-up to be handed in on the last day of class, 4/16. The objective of this project is for you to see how the concepts discussed in lecture can be applied to a real dataset of interest to you. This analysis must be novel and cannot borrow any work you have done for another class or project.

Think of the end product as being the type of long-form. analysis that could be posted as an article on a quantitative blog. It will involve rigorous econometrics and thoughtful analysis, but it is less substantial than an academic paper.

Schedule:

This project will be developed over the course of the semester according to the schedule below. I will give you some feedback as the schedule unfolds, but please reach out to me if you would like to discuss what you are doing in more depth. This is especially true in the last couple weeks of the semester, when you will be pulling the paper together.

1. 10/4 (as part of Problem Set 2): Propose 3 datasets that you think might be interesting to study. A long list of publicly available datasets are available at the end of this document; this can be a very helpful place to look.

2. 10/18 (as part of Problem Set 3): Choose a dataset and propose a research question that the variables in the dataset could help you to answer.

3. 11/5 (as part of Problem Set 4): Provide summary statistics (number of observations, mean, variance, median, minimum, maximum) of at least 3 variables in the dataset.

4. 11/19 (as part of Problem Set 5): Show results from at least 2 regressions in your dataset, including both coefficients and standard errors.

5. 12/3: Final write-up due.

Final Write-up:

(Steps 1-4 above will be discussed in greater detail in their corresponding Problem Sets.)

The final write-up (step #5 due on 12/3) should be roughly 1,500-2,000 words, excluding tables and figures. It should include the following components:

1. A discussion of a research question or hypothesis. For instance, “Do the returns to education (in terms of salary) differ for people with different IQs?” You should discuss why this issue is interesting and/or important, and how knowing the answer could guide how individuals, organizations, or policymakers might make decisions. (~250 words)

2. A description of the dataset. This should include the time period covered, the number of observations,the variables included, how the data was collected, and anything else noteworthy about the dataset. (~125 words)

3. Summary statistics describing all variables to be used in the analysis. These should be “publication quality” tables showing (at least) number of observations, mean, variance, median, minimum, and maximum. If there are two variables of primary interest, it may be helpful to show a scatterplot. (~125 words)

4. An explanation of the empirical methodology. What will be the lefthand side variable in your regressions, and what will be the key righthand side variable(s)? Do you have to do any data transformations (e.g. take logs of variables)? Why do the regressions you propose help to answer the research question you proposed above? How will you interpret the coefficients that Stata spits out? What key hypothesis test(s) will you conduct? (~250-500 words)

5. Discussion of results from at least three regressions. (~500 words)

a. A “baseline” regression ofy on x1 (and potentially other regressors x2, etc.)

b. A “flexible” regression that allows x1 to enter non-linearly (e.g. quadratic, non-parametric)

c. A version that adds an additional regressor to the baseline (you can also add it to the flexible version ifyou think it makes sense)

You should discuss how the implications ofyour regression results for your hypothesis, including sign (positive, negative), magnitude, and statistical significance. If you feel you got a clear answer to your question, what was it? If the answer is still unclear, why is that? Be sure to discuss how adding the regressor (in regression c) affected the coefficient and the standard error on your main x1. Also discuss whether you think the baseline or the flexible version is more useful. This is the main section of the report, and it will be different for each person; the questions above are not an exhaustive list. Think hard about what your results say (and don’t say) and discuss that here.

6. Literature review; described further below. (~250-500 words)

7. Conclusion. What was your empirical analysis able to teach you? What are its

shortcomings? What are some new questions it brings up? Ifyou had more time and/or better data, how would you try to improve the analysis? (~250-500 words)

Evaluation:

Your paper will be evaluated along the following dimensions:

● Did you employ the tools from the course in a proper and sensible way? Concretely, do the regressions that you ran on your dataset constitute a reasonable approach to answering your research question?

● Did you interpret the results of your regression analysis in a reasonable way?

● Did you explain the results in a clear, transparent way? I will not explicitly grade you on the quality ofyour English prose, but papers will receive higher grades if they are able to explain the ideas in a way that is easy for a reader to follow. The writing should be formal (e.g. avoid contractions like “won’t”) and should include economic/econometric terms whenever relevant.

○ Note: Think of the “target audience” as your Econ 2560 classmates. You do not need to try to explain your ideas in a way that the “(wo)man on the street” could understand – you can assume an understanding of econometrics from your readers. However, you should write as an independent project that is separate from the components you handed in during the semester.

Literature Review:

You must provide a literature review of prior work relevant to your research question. This literature review will help you conceptualize and operationalize your research question. This literature review must be comprised of at least three sources. For most projects, these citations should come from academic journals. However, if your project is about a topic that is not closely studied by academics (e.g. some sports topics), then we can discuss suitable alternatives.

These three (or more) sources can be searched via academic journals’ electronic portals such as, for example, “JStor” (http://www.jstor.org/), “EconLit” (http://www.aeaweb.org/econlit/), as well as other electronic resources accessible via the NU Libraries (http://library.northeastern.edu/find). Other useful academic sources are, among others, the “Journal of Economic Perspectives” (http://www.aeaweb.org/jep/), the “Journal of Economic Literature” (http://www.aeaweb.org/jel/index.php), the “Handbook of Economics” series (http://www.elsevier.com/books/book-series/handbooks-in-economics), and the “New Palgrave Dictionary of Economics” (http://www.dictionaryofeconomics.com/dictionary). Each of these resources is available to you, free of charge, via the Northeastern library website (http://library.northeastern.edu/find). If you are in doubt on how to use any of these resources to locate academic articles relevant to your research topic, I highly recommend you consult a librarian at the Northeastern library.

For each of the formal sources you select for your literature review, you should provide a brief and to-the-point summary of the article. Then, you must discuss how your findings relate to the literature you are citing to. Do your findings contradict anything the other papers found? If so, why might this be (e.g. different setting? different econometric approach)? Do your findings extend the conclusions of the prior literature (e.g. “the prior literature finds a positive effect of education on earnings, but I find this is mostly driven by men”)? Exactly what this section should contain is case-specific, but the above questions are good starting points for how you should think about what to include.