This research adopted the conjoint choice model for estimating the
coefficients of attribute levels for two reasons: one, when business startups
or existing operators intended to change existing attractions or
existing types of services, demand in this hypothetical situation would
have to be estimated by using intention as a dependent variable; and,
two, it is well known that intention is a direct predictor of behaviors
(Fishbein and Ajzen 1975). Even though demand in the hypothetical
situation can be estimated by both the conjoint choice model and the
conjoint model, ad hoc assumptions of error distribution should be
carried out for the conjoint model. Due to this drawback of the conjoint
model, this research used the conjoint choice model.