Dear All,
I am struggling a lot with this. I need to calculate the treatment effect among the compliers of my sample, those who responded the Manipulation check question well, in other words, those who 'truly took the treatment'. My professor gave me the following commands to do that:
*Compliance Status
tab treat comply
*CACE via 2SLS
ivreg earnings (comply = treat), ro
ivreg earnings (num_ses = treat), ro
use "/Users/ljk20/Google Drive/PPOL 503/09 - Noncompliance/MayorDebate.dta", replace
* ITT Effect Y
reg ochange watch
reg watchdps watch
di .057/.205
ivreg ochange (watchdps = watch)
However, I do not understand it quite well, and I need to finish the paper by the end of this month. He is away and is unreachable. Basically, I have a variable for my outcome variable, I have another for treatment (control group vs treatment group) and I have a question measuring manipulation check. I would need to calculate the CACE (compliar average causal effect) for two-sided compliance. Can someone help me out with that please?
Thanks a lot!
best
Marie
I am struggling a lot with this. I need to calculate the treatment effect among the compliers of my sample, those who responded the Manipulation check question well, in other words, those who 'truly took the treatment'. My professor gave me the following commands to do that:
*Compliance Status
tab treat comply
*CACE via 2SLS
ivreg earnings (comply = treat), ro
ivreg earnings (num_ses = treat), ro
use "/Users/ljk20/Google Drive/PPOL 503/09 - Noncompliance/MayorDebate.dta", replace
* ITT Effect Y
reg ochange watch
reg watchdps watch
di .057/.205
ivreg ochange (watchdps = watch)
However, I do not understand it quite well, and I need to finish the paper by the end of this month. He is away and is unreachable. Basically, I have a variable for my outcome variable, I have another for treatment (control group vs treatment group) and I have a question measuring manipulation check. I would need to calculate the CACE (compliar average causal effect) for two-sided compliance. Can someone help me out with that please?
Thanks a lot!
best
Marie
