*4.2. Visitors' WTP for Less Preferred Times*

Groups were divided according to their preferred times for night tourism, with Group A's preference for peak times of 6 –10 p.m., and Group B's preference for non-peak times of 7–9 a.m. and 10 p.m. to 12 a.m. (Table 2). A comparison of the frequency of visits did not show a large difference in that 48.2% of Group A visitors who prefer peak times visited Suwon Hwaseong Fortress more than three times in the past year, whereas 49.2% of Group A visitors who prefer non-peak times visited more than three times.

Table 3 showed the conditional logit model estimation results of the selection experiments for each group. First, the coefficient estimates of payment amount are negative and statistically significant at the p < 0.01 significance level. Thus, when other conditions are held constant, the larger the amount of payment, the lower the probability of selection, indicating rational decision-making has occurred in the experiment. Again, the coefficient estimates by attribute and level were positive and statistically significant (except for interpretation). That is, when other conditions are constant, the probability of choosing another alternative (level 2 or level 3) increases compared to the baseline (level 1).


**Table 2.** Conditional Logit Model Estimation Results by Preferred Time Group.

<sup>1</sup> Parentheses indicate standard errors, \*\* *p* < 0.05, \*\*\* *p* < 0.01.



<sup>1</sup> Willingness to pay is a payment amount that one is willing to pay as content is added compared to level 1. Square brackets indicate 90% confidence interval using the Krinsky–Robb [43] parametric bootstrapping. <sup>2</sup> \*\* The different in payment amount is statistically significant at the 0.10 level.

MWTP estimation results for each group are summarized in Table 4. Overall, MWTP increases from level 2 to level 3, but not at a statistically significant level. Group B that prefers non-peak times has a relatively lower MWTP, except for rides. However, as the differences are small, Group B still demonstrates the willingness to pay, and the dispersion of visitors to non-peak times would be possible through discounts and product differentiation.


**Table 4.** Conditional logit model estimation by preferred seasons group.

<sup>1</sup> Parentheses indicate standard errors, \*\*\* *p* < 0.01.

The order of preference for attributes differs between the two groups with MWTP for level 3 in Group A that prefers peak times in the order of Performance > Experience activities > Paid rides > Night views > Interpretation. In contrast, Group B's order of preference is Tourism vehicles > Performance > Experience activities > Night view > Interpretation. Thus, willingness to pay for performances and rides varies the most depending on the preferred time of visits. Group B visitors who prefer non-peak times have a relatively lower willingness to pay. However, considering the margin of error, the differences in payment amount in willingness to pay are not statistically significant at the 0.10 level. Group B also indicated a higher willingness to pay for Tourism vehicles, and this difference was statistically significant.
