I have a survey on willingness to pay for health insurance that takes the following form: respondents are first asked if they are willing to pay an Amount A, if they say yes, questioning stops, if they say no, they are offered a lower Amount B. They say either yes or no to Amount B and then questioning stops. So there are three possible outcomes: Yes; No-Yes; No-No.

I would like to run an ordered logit or ordered probit model in Stata where these outcomes make up the independent variable. To complicate things, there are 4 possible starting values of Amount A.. Each respondent is randomly assigned one of the values.

My question is how can I take into account that there are different values of Amount A (and subsequently B)? Is it to do with groups? Or do I need to write a program?

As you can no doubt guess, my experience with this is quite limited, so any information at all would be greatly appreciated.