I am using instrumental variable estimation, and my instrument is a binary variable (1=yes, 0=no). My dependent variable is binary too so I am using -ivprobit-.

My question is whether I should use factor notation for the IV in the -ivprobit- command or not?

For example, should I use

Code:
 ivprobit y x1 (x2=i.z)
or

Code:
ivrprobit y x1 (x2=z)
(y is the dependent variable, x2 is the endogenous variable, z is the instrument and x1 is an exogenous regressor)

Please advise! Thanks in advance!