Hello
I am new to the forum and I have recently started using Stata. I'm learning something new every day and I needed help related to a project I'm doing.
I am trying to use the "csdid" command in stata which is for treatment with multiple time periods and I have a dataset of about 64 countries of which 38 are treated and 26 are never treated. While I do have treatment year for the "treated group" but nothing for the "never-treated/control group". So, for the treated group I have taken 3 years for each post and pre-treatment year (so for each country it's 6 years in total) , while I have taken data for the control group from 1987-2021. Hence, whenever I run my estimation, I'm getting a very high standard error.
I was hoping to get some help regarding this issue wherein if there is a way to accurately decide which years should be used for the control group so that I run the code correctly. Also, I was thinking if I can create a foreloop in stata wherein I can just test subsamples of 6 years each for control group without having to make separate data sets.
Also, I was thinking of doing a propensity score matching method and then using the "csdid" command but I don't know how helpful that might be or if its even the right way to think about it.
I would be very grateful if someone can help me out with this, I will put a glimpse of how my data looks. In the first picture you will see my treated countries Albania (treatment year = 2009) and Armenia (treatment year = 2006) In the second picture I have my control countries with data from 1987-2021. Both of these are appended data sets.
Thank you
I am new to the forum and I have recently started using Stata. I'm learning something new every day and I needed help related to a project I'm doing.
I am trying to use the "csdid" command in stata which is for treatment with multiple time periods and I have a dataset of about 64 countries of which 38 are treated and 26 are never treated. While I do have treatment year for the "treated group" but nothing for the "never-treated/control group". So, for the treated group I have taken 3 years for each post and pre-treatment year (so for each country it's 6 years in total) , while I have taken data for the control group from 1987-2021. Hence, whenever I run my estimation, I'm getting a very high standard error.
I was hoping to get some help regarding this issue wherein if there is a way to accurately decide which years should be used for the control group so that I run the code correctly. Also, I was thinking if I can create a foreloop in stata wherein I can just test subsamples of 6 years each for control group without having to make separate data sets.
Also, I was thinking of doing a propensity score matching method and then using the "csdid" command but I don't know how helpful that might be or if its even the right way to think about it.
I would be very grateful if someone can help me out with this, I will put a glimpse of how my data looks. In the first picture you will see my treated countries Albania (treatment year = 2009) and Armenia (treatment year = 2006) In the second picture I have my control countries with data from 1987-2021. Both of these are appended data sets.
Thank you
Attached Files
