Replication Exercise #2
Promotion Policy and Diff-in-diff Estimation
Read Ahn, T, J Niven, and A Veilleux (2021), How long have you been waiting? Explaining the role of irrelevant information in the promotion of US Navy officers,
Reproduce Table 4 in the paper using the Stata dataset “replication_promotion.dta” You are welcome to use R if you wish. I should be able to run your script without errors by only changing the parent directory.
Additional questions to answer:
Who do the authors just run LPM? What is lost or gained by doing this instead of logit/probit?
In the text, the authors use In-Zone group as the control group since the authors claim they are not impacted by the policy change, yet they write that this is
incorrect. Explain their reasoning for why the IZ group is not exactly a control group and why they treat it as a control group anyway.
What other variables or information would be useful to have to answer the research question more fully?
You are welcome to work with a partner – if you do, only one student should submit the answers with both students’ names on the documents.
Submit (1) do-file; (2) replication of Table 4 (*.rtf or *.xls). Include your answers to “Additional questions” in the *.rtf or *.doc file, or as a block comment at the top of the do-file. Filenames should have your name, e.g. “Bacolod_replication.do”
The file “replication_promotion.dta” has the following variables:
mastid: unique index number of anonymized officer
promoted: dummy variable, with officer promoted=1 and otherwise=0
belowzone: dummy variable, with officer currently BZ=1 and otherwise=0
inzone: dummy variable, with officer currently IZ=1 and otherwise=0
abovezone: dummy variable, with officer currently AZ=1 and otherwise=0
female: dummy variable, with female=1 and male=0
nonwhite: dummy variable, with non-white=1 and white=0
married: dummy variable, with married=1 and single/divorced/widowed=0
morethan3kids: dummy variable, with 4+ kids=1 and 3 or fewer kids=0
priorenlisted: dummy variable, with ascension source via enlisted=1 and otherwise=0
md_jag: dummy variable, with doctor/lawyer=1 and otherwise=0
ma: dummy variable, with a MA/MS degree=1 and otherwise=0
LCDR: dummy variable, with current rank at LCDR=1 and otherwise=0
CDR: dummy variable, with current rank at CDR=1 and otherwise=0
totalsuffixcodes: Total # of suffix codes held (proxy for ability) year: calendar year