Factor variables in an ordered probit and binary logit

[Cambridge101]

Hi there,

I was wondering if you could help me with something that may seem elementary. I am using an ordered probit and a binary logit with categoric variables. I am using stata 11 and was wondering whether it is necessary to specify factor variables in the following form:

oprobit Y i.X1 i.X2 etc

or whether it is still correct to specify:

oprobit Y X1 X2

Any help would be appreciated!

Thank you!

[Amir Sariaslan]

Cambridge101,
The former example treats the covariates (X1 and X2) as categorical while the latter treats them as continuous.
Amir

[Richard Williams]

If X1 and X2 are 0/1 dichotomies, either will work. If they have more than two categories or the coding is not 0/1, you will get different results. The first coding is better if you want to use post-estimation commands like margins.