The results of the paper were fully reproducible with the data & code provided. I used Stata 17 for the replication, the results where the same when using version 17 or setting it back to 14 (as in the original analysis). There were no edits to the code necessary (except for setting the working directory).
I am quite familiar with Stata, but have little knowledge of the applied methods. Therefore, I cannot judge whether these make sense.
None
Stata 17 (14)
Assigning the results of the Stata output to the edited table in the paper.
The graphs where 100% reproducible with the code provided. Even the results mentioned in a footnote where reproducible.
It would be nice to be also able to see also data coding and editing that led to the 'clean' data provided for the replication. But obviously that is a lot of work and requires a very different level of transparency, as this code tends to be very messy.
The comments in the code are very helpful to assign the numbers in the Stata output to the results in the paper. However, it still takes quite some time to do this.
Very helpful comments in the code, full replication of the graphs, everything works well.
Well, I just run the code, but that does not mean I am familiar with the methods now.
Thanks for uploading one of the few social science papers on reprohack and the great code! I learned a lot about doing nice graphs in stata.