Stata omitting interaction term due to collinearity

[Nikita Shukla]

Hi everyone, I'm having a bit of trouble with a difference-in-difference regression I am trying to run! I have a dummy variable 'time' which is equal to 0 if it's pre-treatment period and 1 if it's post-treatment period. I have another dummy variable 'treated' which is 0 if it's the control group and 1 if it's the treated group. I also created a 'did' interaction term variable which is time*treated. However, when I perform the regression y time treated did, it always omits 'did' due to collinearity. could anyone help out with why this is happening and how to fix this??

Thank you!

[Rich Goldstein]

show the result of:

ta time treated, mi

[Nikita Shukla]

show the result of:

ta time treated, mi

here it is:

[Rich Goldstein]

the zero in the upper right cell is the issue - there is not an "interaction" in the sense you are attempting to model

[Nikita Shukla]

the zero in the upper right cell is the issue - there is not an "interaction" in the sense you are attempting to model

thanks for your answer! sorry for not understanding but what exactly is the issue? and how should I fix it?

[Øyvind Snilsberg]

treated should be 1 for the treated group both when time is 0 and 1

[Nikita Shukla]

treated should be 1 for the treated group both when time is 0 and 1

oh of course! thank you so much!