Dear Statalists,

I create a variable = 1 in case of treatment and = 0 in case of no treatment. Then, I run the following probability model:

Code:
xi: probit f.treatment  ly ldef ltreatment
where ly and ldef are continuous variables and ltreatment is the lag of the variable treatment.

Stata gives the following error message:
outcome does not vary; remember:
0 = negative outcome,
all other nonmissing values = positive outcome
r(2000);
Find below the report on the variable treatment, indicating that actually the variation is present (however small).
Code:
tab treatment

  treatment |      Freq.     Percent        Cum.
------------+-----------------------------------
          0 |         51        3.12        3.12
          1 |      1,581       96.88      100.00
------------+-----------------------------------
      Total |      1,632      100.00
What's happening? and what do you suggest I should do?

Thanks,
Alessandro