The samples were collected from the database of Lifewin (www.lifewin.com.tw). Lifewin is an online survey network consisting of 50,000 members in Taiwan. This research employed online, stratified sampling and sent the questionnaire to 1200 Lifewin members including 600 in the north, 300 in the middle, and 300 in the south of Taiwan. A total of 390 members (who had reserved a room from a hotel website in the last 12 months) replied and the response rate was 32.5%. After deleting samples with questionable responses (e.g., participants answered “Strongly Disagree” or “Strongly Agree” to all questions appeared in the second section of the questionnaire), the numbers of valid observations in the north, middle, and south of Taiwan were 173 (47.3%), 91 (24.9%), and 102 (27.9%), respectively, for a total sample size of 366. To measure the representativeness of the observations in each region, a goodness-of-fit test was used and the result (p-value = 0.747) is not significant, showing that there is no significant difference between the population percentage and the sample percentage in the three regions. Stratified sampling is a probability sampling approach that is distinguished by the following two steps. First, the population is divided into two or more mutually exclusive and exhaustive subsets. Second, simple random sampling is employed to choose samples from each subset. In our research, the total samples were first divided into sub-samples based on different regions (north, middle, and south) in Taiwan. Then the stratified samples were randomly selected from the north, middle, and south of Taiwan.