Due April 8
a) Produce some initial summary statistics of the data.
b) State the hypotheses that will be tested.
Show me advanced results (analyses, not write-up/paper)
Due April 15
You have to provide me with a report that should be no more than 5-pages long of text, double spaced, 12 point, Times New Roman font, with 1 inch margins. It should include the following short sections:
i) Explain properly the research questions of the original paper you are replicating.
ii) Explain which of those questions you will replicate, and what information they could give you. Also, I expect that you will try other hypotheses.
iii) Description of the data.
iv) Econometric methods used to analyze the data. Here is very important that you justify (explain) why the method used a particular method (why use IV or fixed effects, or run a differences-in-differences regression, etc.).
v) Analysis of results (which should be presented in nice tabular form).
vi) Evaluation of the appropriateness of the utilized econometric method(s). Do we believe the results?
vii) Conclusions and suggestions for improvement of the research.
viii) Tables with results do not count for the 5-page limit. An appendix with the computer code (Stata “do” and “log” files) used in analyzing the data should be included. Also a no more than one-page appendix explaining the construction of any new variables you may have created, could be included.
The objective of this project is to make you follow the process a researcher would normally follow to conduct a research project, of course adjusted by the fact that you have only several weeks to work on it.
So, you are expected to do the following things (this is not an exhaustive list, these are just guidelines to help you through the process):
1) Read the original paper associated to the topic you chose. Think what hypotheses you would be interested in replicating of the original paper, and more importantly, which new ones you could pursue.
3) Analyze the data to understand it properly (the data is in Blackboard). For example, you should analyze outliers etc. or create any additional variables that you need to for your analyses.
4) You should produce a table of summary statistics (and present it) to understand the data. Also, do graphs, etc. (and present the ones you think are worth mentioning) to see any correlations, outliers, etc. that may be causing a problem.
5) Run several regression models building them up and down. (start with a benchmark model, add terms, subtract terms, test model specifications, etc.).
6) Justify why the author (and you) select a particular methodology (why use IV, fixed effects, etc.). Criticize the methodology if you think that there are issues with it.
7) Present your regression results in a few tables. (proper organization of results is very important).
8) Analyze any shortcomings of your methods, data, results, etc.
9) Write clear conclusions.
10) Make suggestions to test your hypothesis with an “ideal” dataset not subject to some of the shortcomings of the data.
The pdf is for the hypothesis and the dta file is for the STATA and other analysis. Could you let me know if you can work with this?
Thank you for your time.
https://www.4shared.com/file/FKa90USz/randomly_assigned_children.html
https://www.4shared.com/office/HKbZ_pAS/How_Large_Are_the_Effects.html
https://www.4shared.com/office/Z4cRZpyP/Codebook_Sacerdote_2007.html