*3.2. Choice Experiment Analysis*

In the present study, the possibility of visitor distribution was examined using CE to determine willingness to pay for night tourism at less preferred times and seasons. Based on marginal utility theory [33], CE estimates the marginal willingness to pay (MWTP) for various attributes that affect the value of goods and services [34], and is one of the most used methods for estimating the value of tourism resources and tourism products [8]. Due to its advantage of being able to estimate the willingness to pay for each attribute that affects the value of tourism resources, CE has been used in various tourism studies [8,34–40].

The willingness to pay for night tours of Suwon Hwaseong Fortress was obtained, and McFadden's conditional logit model (CL) was used to estimate the indirect utility function [33]. MWTP was calculated using the maximum likelihood estimator for each attribute and level. As all variables except for the price variables were binary variables, MWTP values for level 2 and level 3 of each attribute mean additional willingness to pay compared to level 1.

The application of CE requires the setting of attributes and levels that constitute the value of the tourism resource. The attributes and levels used to examine the willingness to pay for Suwon Hwaseong Fortress night tourism were set as shown in Table 1. These settings were in consideration of the specific attributes of Suwon Hwaseong Fortress night tourism as discussed previously in the Section 2.


**Table 1.** Attributes and levels for Choice Experiment.

Considering the types of night tourism in the previous researches [20,21] and types of night programs which currently underway in Suwon Hwaseong, as well as domestic cultural heritage sites, the main attributes were derived. The attributes of Suwon Hwaseong Fortress night tourism include night view, performances, experience activities, rides, and interpretation. The payment vehicle utilized an entrance fee. Each attribute was divided into three levels, with the first level not including any additional content. At level 2 and level 3, additional content is provided.

In the present study, a key question for resolving over-tourism by visitor distribution is "What are the differences in willingness to pay for night tourism at less preferred times and seasons compared with others?" MWTP calculations using CE were used to compare groups that were derived as follows. First, willingness to pay was compared between two groups: Group A that prefers night tourism at the peak time of 6 to 10 p.m., and Group B that prefers night tourism at non-peak times. Second, willingness to pay was also compared between Group C that prefers night tourism during peak seasons (spring,

fall), and Group D that also prefers non-peak seasons. The results allow us to discuss the possibility of distributing tourists to less preferred times and seasons to overcome the problem of over-tourism.
