Friday, January 31, 2020

Handy Dandy Checklist for Those Doing RD

I just found this handy dandy checklist from the Department of Labor for those of your working with regression discontinuity designs. Be sure to check the list before submitting your third year papers or sending a paper to a journal. Actually, I think it's a handy list for referees at journals to use, too! Especially referees who themselves do not have a lot of experience with RD design.

Thursday, January 30, 2020

Stata Tip: How to Make Your Graphs Look Nice

Tal Gross tweeted some really nice, easy tips on how to make your Stata graphs look just a little bit nicer than the default graphs.

Tip #1: Use labels instead of legends. They look so much more professional!
Tip #2: Make all text horizontal (I guess I don't feel so strongly about this one).
Tip #3: Get rid of unnecessary ink. Yes, this does make the graph look like you took just a little bit of time to make the graph your own.

The best part: Stata code is included in the thread. Go ahead and copy-paste his code into your code.

Stata Tip: Create a String Variable from a Numerical Variable's Labels

Maybe I should have known about this command. Maybe I did know it at some point in my life. Either way I re-discovered it recently so I thought I'd share it. 

Here's the situation. Imagine you're using one data set with a variable called, let's say "country of birth," with numeric codes for the different countries and then labels with the  names of the actual countries. Now let's say that you want to merge GDP data for these countries but that other data set lists countries by name, not numerical code. How do you quickly create a new variable in the original data set with the name of the country? Easy! Just use the decode command in Stata! You're welcome. 

Monday, January 13, 2020

Mastering Mostly Harmless Econometrics: 2020 AEA Continuing Education Webcasts

Dear PhD students who have finished the core courses and are about to get started on field courses, 

Please watch these videos very carefully. Pay attention. Take good notes. If you are considering a field in applied micro, then be sure to read this book and this one, too. If you took good courses as an undergrad on modern empirical techniques, then maybe you can watch the videos at 2x speed--which is actually kind of funny. But watch the videos. 

Sincerely,
Delia

P.S.
This is especially true for any students who are considering working with me in the future!


How to be a Good Seminar Participant

Berk Özler has some great advice on how to behave at seminars. My favorite is probably this one:

 Allow presenters time at the beginning to frame their talk without interruption. 

I don't think anyone has to be looking at their watches to see if it is exactly 10 minutes into the seminar, but let the presenter get through the introductory slides before asking an introductory question. And yes, I know you have some concerns about the identification strategy and you are eager to share them, but please, please, let the presenter explain the question and the identification strategy before voicing these concerns. Yes, you know this literature and maybe you read the paper, but it confuses everyone in attendance if we're discussing problems with the technique before knowing what the technique is. Just wait a few minutes. You are encouraged to ask those questions right after the baseline results. 

There is only piece of advice that I would tweak a bit. The post mentions a concern that men ask a lot more questions than women, and a solution to this is for men to "think twice before asking a question (“does it really need to be asked?” “is the answer coming up in a few slides?” “is it better to ask this when I meet the speaker later in the afternoon?”)." I would say that seminars would run more smoothly if everyone asked themselves these questions before asking a question. 

All of that said, I want to stress that I really enjoy lively seminars--both as a participant and as a presenter. I do not think the world would be a better place if only one or two perfectly though out questions were asked in seminars. I am just afraid that we sometimes spend too much time asking unnecessary questions at the beginning--perhaps throwing off the speaker--and we don't have enough time to ask the really juicy questions at the end. And that is a pity. 

From PhDcomics: