STATA: How to define the proportional resp. variable in a logistic regression model

#1
Hi everybody,

I'm having problems defining the response variable in a logistic regression model (probit / logit model). The data set is structured as follows:

Observations are defined as a group of people working in a particular company and industry. A group of people in several company industry combinations were asked several questions (which later define the response variable and independent variables). Let's assume, the response variable (answers provided to one particular question asked) can either be A, B or C, whereas A shall be defined as the share of people in a company industry combination answering A, B shall be defined as the share of people in a company industry combination answering B, and C shall be defined as the share of people a company industry combination answering C. The sum of A, B and C for each observation sums up to 1. Likewise, independent variables are defined, whereas the number of possible answers varies.

Let's assume the following for the response variable:
First observation: A: .3, B: .4, and C: .3
Second observation: A: .8, B: .1, and C: .1

Unfortunately, I am not a STATA expert, so I am currently having some issues organizing the data set and generating the response variable appropriately. How do define my response variable for STATA to perform a logistic regression?

Your help would really be appreciated.

Thank you,
TheEconomist