cm commands for unlabeled choice experiment with status quo

[Julia Schubert]

Dear community,

I have conducted a DCE (9 choice tasks per respondent, 3 alternatives per choice task) and I am facing several issues with preparing my data for analysis.

1. Can it be supposed that my data is panel data due to the fact that each respondent is answering 9 choice tasks? And therefore, is the only possible model that can be used cmxtmixlogit or can I also cmset my data to panel time choice model data and use cmclogit or cmmixlogit afterwards or probably setting it to cross-sectional data even if one respondent answers several choice tasks but the time variable is not really important as between answers lie within seconds or minutes? And why is it that when cmsetting my data to panel time choice model data and defining that there is not an alternatives variable (as it would be for unlabeled choice experiment) it is not possible to select time unit for time variable again?

2. My choice experiment contains one status quo option (with fixed levels) and two alternatives A and B per choice task. Is it therefore still unlabeled even if the status quo option is always the same and named "status quo"? Or might it be that due to the status quo option it is necessary to use nlogit command? Also I´ve read that the margins command does not work for unlabeled choice experiments. Does anyone have experience with computing marginal willingness to pay for unlabeled choice sets?

Any insights and hints are highly appreciated.
Thank you and best regards,
Julia

[Hadji Jalotjot]

Hi! How did you address this problem? Did you set your data as panel-time because I am going to try it because it looks like there is no other choice.