To paraphrase TS Eliot (maybe), good writers borrow, great writers steal. Let’s
To paraphrase TS Eliot (maybe), good writers borrow, great writers steal. Let’s do the economic history version of that.
Pick a paper on the syllabus. This could be a primary paper you were assigned to read or one of the additional papers, but some of the more recent papers with identification strategies (IV or DD) and attempts to find causality will work best for this assignment. You might want to limit yourself to papers from lectures we have already had (that is, maybe don’t pick a paper in the Crime lecture from late November for the assignment due in September?) but that is your choice.
What is a new outcome (Y) that you could look at with the same identification strategy (X)? Don’t just pick an X and a Y randomly—think about why this would be an interesting question to answer. Hopefully, the new question is really exciting and goes well beyond the original paper.
Please write one page and answer all of the following:
1. Describe the identification strategy in the original paper. Is it instrumental variables or difference in differences or something else? What do the authors assume or use to get causality in their estimates? What is the original result or finding?
2. What new question or outcome can you look at with this old identification strategy? Be specific as to why this is a question worth asking and trying to answer. That is, why should we care about your new question? Pose your question as clearly and succinctly as possible.
3. What kind of data would you need to carry out the research? Presumably the X variables might all come from the original paper but how would you measure the new outcomes you are interested in? If measurement of the Y variables is obvious, state them as clearly as possible. If measurement of the Y variables is tricky or novel or complicated, do you best to describe how you would measure things. Be sure to clarify the level of your analysis (individuals, cities, states, countries, areas, etc), though this will likely match that of the original paper.