Greetings,
I am working with a modest size sample of survey data (n=990). I'm interested in estimating a probit (or logit model) of one of the response variables, that is coded (0/1). Many of the other reponses are categorial, not continuous variables. For instance, the respondant's household income is recorded as a range, e.g., under $50,000, $50,000 - $74,999, $75,000-$99,999, etc., rather than as the specific income level itself. Can such variables be used as a regressor in a standard logit/probit setting without modification?
Thanks,
Mark
|