loyalty, customer value and satisfaction to find the ways of understanding customer needs and wants (Oh, 1999; Williams, 1999).
3. Methodology
3.1. Population and Sample
    The population of hoteliers was defined as the subjects were recruited from the pool of hotels that the frame was obtained from Tourism Bureau Ministry of Transportation and Communication R.O.C.. A total of 749 hotels are the population in this study. One hundred and eight-one hotels returned the survey and 186 usable responses were eventually available for the data analysis. Thus the response rate was 25.50%.The population of leisure travelers was defined as the people who are Taiwan citizen and more than 12 years old. According to the formula, the 1111 leisure travelers in Taiwan will be selected for the instrument testing at 3% estimated error (e) and .05 confidence level. In the formula,the choice of sample size is governed by the three factors. They are the variance of the population, the margin of acceptable error, and the confidence level (Lewis, 1984;Ryan, 1995; Saunder, Lewis, & Thomhill, 1997; Tull & Hawkins, 1987). The N is the population of leisure travelers defined in this study and it was about 15 million.The subjects were recruited from the hotel guests stayed at the hotel to complete the questionnaire. The research assistants approached the every second guests who entered the lobby during the weekend because there were more leisure travelers(Tourism Bureau, 2005). There were 27 guests refused to answer the questionnaires.Therefore, a total of 1068 leisure travelers were included in this study.
3.2. Measures
    There are some debates of LODGSERV (Ekinci & Riley, 1998;Kivela et al., 1999a; Kivela, Inbakaran, & Reece, 1999b, 1999c) and it failed to confirm the five dimensions when applied the SERVQUAL to restaurants in Taiwan(Huang, 1997). A seven-point Likert scale ranged from 1 for strongly disagree to 7strongly agree to record subjects’ perspectives of hotel service quality. This measure is able to explore the level of subjects’ perspectives with regard to the relationship between service quality and variables.Twenty experts in the field of service quality management and hotelier were invited to serve as panel members in order to ensure the content validity of questionnaire. A pilot study of the questionnaire was conducted prior to collecting the final study data of 50 subjects for hoteliers and 100 subjects for leisure travelers.They were asked to comment on the validity and readability of the questionnaire. A factor analysis with orthogonal rotation was conducted for data reduction.Modifications were made according to the subject suggestions and the results of factor analysis.
3.3. The sampling method
    In this study, the proportionate stratification or uniform sampling fraction was conducted for leisure traveler’s survey sampling. Because a proportionate stratified sample can be more precise than stratified random sampling of the sample size. In addition, the researchers can control the sample sizes in the strata, rather than being randomly determined by the sampling processes (Kalton, 1983). In order to fulfill the requirement of stratification, the population proportions in the strata were calculated and it was possible to draw separate samples from each stratum.
3.4. Data analysis
    Exploratory factor analysis with varimax rotation was conducted to create new variable index and reduce the service quality questionnaires how subjects regard service quality variables to final dimensions. And the principle component method was performed on transforming a given set of variables into a new set of composite variables that are uncorrelated to each other (Hair, Anderson, Tatham, & Black, 1998).Factor loadings of variables were greater than or equal to .50 are chosen to form a factor and the factors with eigenvalues greater than or equal to 1 were significant.For assessing the internal consistency reliability, Cronbach’s alph has been calculated to examine the stability of each variable in each new index. The reliability coefficients for each factor that is greater than or equal to 0.70 values were considered to accept (Nunnally, 1978).In order to examine the service quality perspective of leisure travelers and hoteliers, K-Means cluster analysis was conducted and determined whether they would be grouped into identifiable patterns based on their perspective by using hierarchical agglomerative method. It is imperative to select the clustering variables focusing on the research questions (Jurowski & Reich, 2000). In this study, the purpose is to compare the perspectives of hotel service quality between leisure travelers and hoteliers; the clustering variables would be the perspectives of hotel service quality.Analysis of variance (ANOVA) indicated the significant group differences between dependent variables. It was concerned with different service quality perspectives between the clusters because the analysis is appropriate when the independent variables are categorical in nature and the dependent variables have an interval quality. To examine the variations, post hoc tests were performed. The independent sample t-tests were used to compare the different perspectives with regard to service quality between leisure travelers and hoteliers in Taiwan hotel industry. LODGSERV在客户和酒店经营管理英文文献和中文翻译(2):http://www.youerw.com/fanyi/lunwen_31976.html