Hi all,
I did a DCE experiment, where 100 people got 3 questions to chose between two therapies: therapy A and therapy B. Then I included three variables, chance for complication X, chance for comlication Y and chance for complication Z. The chances for XYZ are differeing between low and high in therapy A while the chances for XYZ in therapy B are always none.
I coded with dummy variables to perform a conditional logit. However, when performing this conditional logit, however, I cannot include the dummy variables for XYZ for therapy B, as they are collinnear with each other. This raises a question, what is the proper way to investigate the beta coefficients for the complications, X, Y and Z? Should I order my data in another way or should I do something entirely different?
Best, Maarten
I did a DCE experiment, where 100 people got 3 questions to chose between two therapies: therapy A and therapy B. Then I included three variables, chance for complication X, chance for comlication Y and chance for complication Z. The chances for XYZ are differeing between low and high in therapy A while the chances for XYZ in therapy B are always none.
I coded with dummy variables to perform a conditional logit. However, when performing this conditional logit, however, I cannot include the dummy variables for XYZ for therapy B, as they are collinnear with each other. This raises a question, what is the proper way to investigate the beta coefficients for the complications, X, Y and Z? Should I order my data in another way or should I do something entirely different?
Best, Maarten
