Probit: Assumption e~N(0,1), why?

RBla

New Member
#1
Good evening,

I have got a question according the Probit-Model. In this model we always assume that the error term is standard normally distributed. I understand, that it's a reasonable assumption that the error term has expectation zero - if we did not forget important variables - and that it is normally distributed. I do not understand, why we assume that it is standard normally distributed... why not N(0,4) or N(0,9), or whatever?

Thanks a lot for your help in advance!