We study the choice of dwelling units in a condominium complex, a problem that has not been thoroughly investigated in the housing research literature. We take a revealed preference approach to quantify home buyers’ preference toward attributes associated with a dwelling unit. In particular, we estimate a mixed logit model using the hierarchical Bayes approach based on the Markov Chain Monte Carlo method. We derive the willingness to pay measures for detailed attributes associated with a dwelling unit including its floor level, orientation, location in the complex, location on a floor level, and the type of bathrooms. We also find strong evidence for the preference heterogeneity among home buyers and conduct regression analysis to explain the preference heterogeneity using home buyers’ socioeconomic characteristics. Our results show that home buyers with older ages and higher annual household income tend to focus more on the quality of the dwelling units, while first-time home buyers are more willing to accept dwelling units with less-desirable attributes.