Estimation of forced choice demand models using discrete choice experiments (DCEs) continue to dominate in health economics. Forcing individuals to choose may misrepresent the demand function, leading to biased policy recommendations. This paper found differences between parameters generated from forced and unforced demand models in a DCE investigating nurse decisions on rural service in Liberia. Given that nurses have the option of remaining in their current situation, results from the unforced model have more face validity. It is important that future DCEs consider inclusion of opt-out options. Failure to do so may bias policy recommendations.