Panel data : assigning random treatment to same no of observation yearwise

[Aditi Roy]

Hi, Stata people.

I have got a 20 years panel data 1990- 2010. For the falsification test. I drop all the treated variables and I randomly want to assign treatment to same no of observation as it was in year x .

So far I am manually doing the work. The toy story is :
[generate random= uniform() // generating random values
sort year random // sorting the random value by year as I have to assign same no treatment as it was in year X
gen treat =0 now I am generating the treatment variable treat
replace treat =1 in 1/5 // as the years are sorted. I am manually looking into the year and in the year 1990 (which runs from cell 1 to 25), only 5 was treated
replace treat =1 in 26/30 // again in 1991 5 was treated, 1991 year cell runs from 26 to 45
][/CODE]


Can anyone help me how to go about it rather than looking at the cell for the year range and then assigning the n observation.

Any help is appreciated.

[Aditi Roy]

Dear stata pros,

To be more clear on what i am trying to do. I am randomly generating variables after dropping all the treated variables. is there any way out on how I can assign treatment to same no of provinces ( say 24 as it was originally treated) in a particular year.

Eg: I want to assign treatment to 24 provinces in 1990
45 provinces in 1991 and so on.

so far I am doing this manually by generating random , sorting it by year as the total provinces assigning changes by year. So I am doing it by looking at the cell no’s.

Can anyone help me on how I can do it via coding ?