Stata Hint: Collapsing and Merging Large Data Sets

I say I write this blog for my graduate students (current, future, and past), but really, I write it for myself. It's a nice place for me to put things down I want to remember but have a feeling I will forget. Today's post is for my future self who is annoyed because dofiles are taking too long to run. 

Dear Future Delia, 

Why not try ftools and gtools? You can read more about ftools from the slides Sergio Correira presented at the 2017 Stata Conference. Yes, the same Sergio Correira that wrote reghdfe! Mauricio Cáceres Bravo wrote gtools and you can read more about that here. 

Sincerely, 

Current Delia 

Job Talk Season

The European Economics Association (EEA) just posted some very helpful slides on how to give a great job talk. Really, they're useful to read just before giving any seminar--even if you've been giving seminars for more than a decade.

Kjetil Storesletten has some great advice in his talk, "The Ten Commandments for How to Give a Seminar." Very helpful not only for preparing slides but also for writing introductions. The most important thing is to convey the value added of YOUR PAPER. What does your paper bring to the existing literature? What's the magic that makes that happen? 

Eliana La Ferrrara gives very helpful specific examples on how to give a seminar in her talk, "How to Present your Job Market Paper." I really love how she goes into so much detail on the standard structure of these talks. I guess I had never thought much about it, but these talks do have a standard structure. My advice: Sure, you can deviate from the standard if you have a strong reason to do so, but if not, stick to the standard structure--especially now that you know exactly what it is.

More on Clustering..

We all know that, at least, we should be clustering at the level of treatment, but we also know that many times during seminars, we are asked to cluster at different levels. One potential take-away people might get from this is that we should cluster in all of the ways. The paper by THE econometricians said "not so fast." (See my previous post about this issue). 

And now for more on clustering, the Declare Design folks explain when we should cluster above the level of treatment.  I learned of this on David Mackenzie's blog, and there you can see his thoughts on clustering as well.

To my applied micro students (who are not doing RCTs), here is a reminder: These posts are all written using RCT language. You may think none of this applies to you. It does! It is true that we as researchers don't randomly put people into treatment and control groups. But, by using our tricks (RD, differences in differences, IV, etc.), we are assuming that the world is putting people into treatment and control groups for us. This is so important to remember as we think about our own work and evaluate the work of others. Conclusion: (almost?) everything the "randomistas" say about clustering applies to us, too.