Thank you so much for your help.
But I found it difficult using logit, so for now I think it's best to use probit bivariate (I'm a newbie and since there are many tutorials on the web also).
I'm still struglling cause either logit or probit can't accept missing data though, mine has about 16% missing data. Do you have any suggestion for me?
Anyway I'm really grateful and I'll try to post the progress later. And sorry for the late reply.
Most programs use listwise deletion, so if any variable component is missing, then the whole record gets dumped from the analysis. Multiple imputation is the standard approach to deal with missingness.