OBJECTIVES:   This study, using DCE method, figures out the characteristics of the decision-making for cancer treatment and investigates the attributes affecting the respondents’ choice.  Also it ascertains marginal willingness to pay and relative preferences for cancer treatment among the general population.  METHODS:   In the survey, the respondents are asked to choose more than one discrete choice option, resulting in multiple observations for each individual. For each pair-wise comparison of choice set, the respondent will make a choice among three alternatives; A, B, or opt out. Thus, the nested-logit model using full maximum likelihood allows us to empirically estimate multi-levels of dependent variables. For the robustness check of our empirical results, we try considering the nature of distribution of error terms in the utility function in several ways. The key assumption of logit models is the independence of irrelevant alternatives (IIA), which results from the assumption that other alternatives are independent. The validity of IIA assumption can be simply conducted by a Hausman-McFadden test. We also consider an alternative specification where error terms are independently, but not identically distributed. And finally stand errors are calculated by using bootstrapping and compared to the previous results. RESULTS:   In this DCE study, evidence of reliability was found at both input and output level. The estimates of MWTP between survival rate and monthly insurance premium and MTWP between total treatment costs and monthly insurance premium, by employing “Hybrid Conditional Fixed Effects Logit Model” to figure out the existence of heterogeneity of any observed and unobserved components, are reflecting reasonable range of 817 KRW~1,324KRW, and 23,690 KRW~38,139 KRW, respectively. CONCLUSIONS:   Compared to female counterparts, male respondents have higher MWTP of monthly insurance premium for two related attributes. Besides, currently married respondents, with higher income, and higher educational attainments have more MWTP compared to their respective counterparts.