Dear friends,

I met the collinearity problem in DID model. I selected the sample who did not enter the program in t-1 year. I set a treated dummy for whether some among them entered the program in t+1 year. The model is,

Code:
y = D + Treat*D + Treat + control
Since treat = 1 implies D =1, Treat*D=Treat =1. D is year dummy. D = 1 if it is t+1 year, equals 0 if it is t year.

My code is

Code:
reg y i.treat##i.D
My data is,

Code:
* Example generated by -dataex-. To install: ssc install dataex
clear
input float(y treat D treatD)
     . 0 0 0
-876.6 0 0 0
-877.8 0 0 0
-881.6 0 0 0
-876.4 0 0 0
  -880 0 0 0
-877.4 0 0 0
-876.8 0 0 0
-873.4 0 0 0
-876.8 0 0 0
  -877 0 0 0
-876.6 0 0 0
-873.8 0 0 0
-885.6 0 0 0
-877.6 0 0 0
-876.4 0 0 0
-875.8 0 0 0
-874.8 0 0 0
-876.2 0 0 0
-875.8 0 0 0
-879.2 0 0 0
-876.8 0 0 0
-874.6 0 0 0
-875.2 0 0 0
-878.4 0 0 0
-876.2 0 0 0
-879.2 0 0 0
     . 0 0 0
  -880 0 0 0
  -873 0 0 0
-877.6 0 0 0
-879.8 0 0 0
  -881 0 0 0
-876.2 0 0 0
  -878 0 0 0
-877.6 0 0 0
  -880 0 0 0
-876.6 0 0 0
-877.2 0 0 0
-881.6 0 0 0
-878.4 0 0 0
  -876 0 0 0
-878.8 0 0 0
-881.6 0 0 0
-879.4 0 0 0
-875.6 0 0 0
-867.4 0 0 0
-883.6 0 0 0
-872.8 0 0 0
-879.4 0 0 0
-879.4 0 1 0
-890.6 0 1 0
-875.6 0 1 0
  -884 0 1 0
     . 0 1 0
     . 0 1 0
-882.2 0 1 0
-879.4 0 1 0
-878.6 0 1 0
  -876 0 1 0
-874.8 0 1 0
-878.6 0 1 0
-877.4 0 1 0
  -883 0 1 0
  -877 0 1 0
-878.8 0 1 0
-872.4 0 1 0
-877.8 0 1 0
  -877 0 1 0
-877.6 0 1 0
-873.2 0 1 0
-878.4 0 1 0
-891.8 0 1 0
-882.6 0 1 0
-879.2 0 1 0
  -876 0 1 0
-873.8 0 1 0
-875.2 0 1 0
-879.6 0 1 0
  -876 0 1 0
     . 0 1 0
-873.8 0 1 0
-878.8 0 1 0
  -900 0 1 0
-877.4 0 1 0
  -875 0 1 0
  -878 0 1 0
-875.6 0 1 0
-875.6 0 1 0
-878.6 0 1 0
-891.2 0 1 0
-873.8 0 1 0
-874.2 0 1 0
-874.8 0 1 0
     . 0 1 0
-877.8 0 1 0
-875.6 0 1 0
-874.6 0 1 0
-879.2 0 1 0
-875.2 0 1 0
end
Thanks.