Hello everyone,
I would like to ask one question that I am struggling on. I would like to estimate the "Treatment-on-the-Treated effect". There are only two groups - group having the treatment and group having no treatment. (means one treatment group and one control group). Treatment is extension service trainings.
For estimating the TOT effects, I use actual training attendance as the treatment variables. I measure training attendance in tow ways: (1) a dummy variable that takes a value of one if household attended at least one of the training sessions that were offered and zero otherwise, (2) intensity of training attendance, measured by the number of training sessions attended relative to all training sessions offered (takes value from zero to one).
The decision to attend training sessions is endogenous. so to avoid that I am using the instrumental variable (IV) approach. And here comes the struggle. I have never used diff-in-diff approach with IV before.
I wrote a commands:
where group - treatment varialbe, par_a - is (1) method of estimating actual participation, par_b - is (2) method of estimating actual participation.
adoption intensity was how many of "new technology" that household adopted (for example here is how many chickens that households adopted in the follow-up year), takes value from zero to infinity. adoption dummy is whether household adopted or not (takes value of one or zero)
******Treatment on the Treated*******using participation Dummy (par_a)
xtivreg adoption_dummy (group=par_a) i.par_a##i.year, vce (robust)
xtivreg adoption_dummy (group=par_a) i.par_a##i.year $controls, vce (robust)
xtivreg adoption_intensity (group=par_a) i.par_a##i.year, vce (robust)
xtivreg adoption_intensity (group=par_a) i.par_a##i.year $controls, vce (robust)
******Treatment on the Treated*******Intensity
xtivreg adoption_dummy (group=par_b) c.par_b#i.year, fe vce (robust)
xtivreg adoption_dummy (group=par_b) c.par_b#i.year $controls, vce (robust)
xtivreg adoption_intensity (group=par_b) c.par_b#i.year, vce (robust)
xtivreg adoption_intensity (group=par_b) c.par_b#i.year $controls, vce (robust)
And I feel like there are wrong. I would really appreciate the help. Thank you in advance
Related Posts with Difference-in-Difference, with IV
bugs for reclink?Dear all, Please run the code below, it says something was wrong, how to fix this? Code: tempfile …
Super simple syntax questionI have a variable that contains totals of advertising costs for different products from different ye…
Reporting an error with assumptions revision after VECM trend estimation.Greetings. I want to report an issue I've discover when estimating a VECM model with trend option, l…
How to get matching results in combination?Dear all, here is the code: Code: tempfile using * Example generated by -dataex-. To install: ssc …
How do I extract month and year data from combined variables?Hi, I have 51 state name variables(long data) and income data (e.g.starting from inc_JAN_1996, inc_F…
Subscribe to:
Post Comments (Atom)
0 Response to Difference-in-Difference, with IV
Post a Comment