EViews.com

Hello,

I am getting my way through probit models. So as for linear regression we have the so called Gauss Markov assumptions that need to be true in order to get the best unbiased linear estimators. My questions is, is there something like this also for probit models?

How do I process with a insignificant constant when I want to get the predicted probability of a certain event? Will I use the value of the constant to calculate the predicted probability as well as all slope values along with their observed (out of sample) x values?