SCALE EVALUATION

The scales are evaluated by two main tools: (1) EFA (Exploratory Factor Analysis) and (2) Cronbach Alpha.

Exploratory factor analysis is a statistical method used to shrink and summarize data (technique to reduce data). This is an interdependence technique, in which all the interdependence relation will be researched. Exploratory factor analysis is commonly used in the process of creating a scale to measure different aspects of the research concept, and to check the individual aspect status of the scale.

Testing by using Cronbach Alpha in order to verify this scale with many variables to have a good scale for each factor being researched (measuring whether a set of question sections drawn from each factor has connection or not). Cronbach

22

Alpha is a statistical test in terms of the level of closeness that questions in the scale interrelated with each other. In other words, it shows us that if a certain scale is good in a certain aspect or not.

To avoid omitting variables, the sequence to evaluate scale is as follows: firstly, all five components of service quality scale: Reliability, Responsiveness, Assurance, Empathy and Tangibles with 19 variables put into exploratory factor analysis to identify the factors. After performing the EFA with 19 variables, each factor is identified and evaluated in terms of the reliability by using Cronbach Alpha.

Like the service quality scale, the credibility scale with four question sections and the satisfaction scale with three question sections are also performed EFA and evaluated the reliability by using Cronbach Alpha.

4.2.1 Exploratory factor analysis

Exploratory factor analysis is performed with 19 variables of the service quality scale. Exploratory factor analysis is considered appropriate when the value of KMO (Kaiser–Meyer–Olkin) is greater than 0.5. Factor loading less than 0.5 will be removed, cut off pointwhen eigenvalue is greater than 1 and total variance explained is greater than 50% (Hoang Trong and Chu Nguyen Mong Ngoc, 2005). The chosen analytical method is Principal components with varimax rotation. They are performed as follows:

Analyzing the set of 19 observed variables, the obtained result is as follows (Appendix 1): KMO = 0.857, sig.= 0.000 of Bartlett’s test. Therefore, the variables are closely correlated with one or more factors and the overall correlative matrix is a rejected unit matrix according to the results of Bartlett’s test, satisfying the conditions of exploratory factor analysis.

Five factors were extracted with the variance explained of 60.656% (explained the variability of 60.656%). Two variables RES03 and EMP07 that have factor loading less than 0.5 will be deleted.

Continuing to reanalyze the factors with 17 variables (RES03 and EMP07), the result is (Appendix 2) as follows: KMO = 0.841, sig.= 0.000 of Bartlett’s test. This satisfies the conditions of exploratory factor analysis. Four factors are drawn with the

23

variance explained of 57.479%.

The variables with factor loading are located in two different factors like EMP05 of factor 1, 2 and TAN11 of factor 2, factor 3, it shows that the correlation of these variables and those two factors are not definitive. After performing the factor rotation, variable which has large coefficients of a factor proves that it correlates primarily with the same factor. Thus, it belongs to that factor (Hoang Trong and Chu Nguyen Mong Ngoc, 2005). Variable EMP05 belongs to factor 1 and variable TAN11 belongs to factor 2. After performing exploratory factor analysis with 17 variables above, four factors have been drawn:

+ Factor 1 includes 5 following variables:

You were warmly welcomed at the first step of the hotel (RES01)

Employees are always willing to help you (RES02)

Employees provides you with friendliness (EMP05) Employees promptly respond to your requests for information (RES04) You receive prompt services from our employees (REL17)

+ Factor 2 includes 5 following variables:

You feel safe in the hotel (ASS10)

Our employees are polite (ASS09)

The lounge appears spacious, clean and comfortable (TAN15) Our employees are well dressed and appear neat (TAN11) Provided food and drinks are hygienic (REL19)

+ Factor 3 includes 3 following variables:

Physical facilities keep up with the “star” rating (TAN12) The layout of hotel facilities makes you comfortable(TAN13)

Our services keep up with the “star” rating(REL16) + Factor 4 includes 4 following variables:

Our employees are qualified, thus responding well to your requests (ASS08) The appropriate quietness is provided to you (TAN14) Our services keep up with the “star” rating (EMP06)

24

Our employees answer your questions clearly and exactly (REL18)

Like the service quality, the credibility with 4 variables is also analyzed, the result is (Appendix 3) as follows: KMO = 0.657, sig.= 0.000 of Bartlett’s test. This satisfies the conditions of exploratory factor analysis. A factor is drawn with the variance explained of 51.342%, in which variable TRU22 which has factor loading less than 0.5 will be deleted.

Continuing to reanalyze the factor of credibility with 3 variables (TRU22), the result is (Appendix 4) as follows: KMO = 0,.643, sig.= 0. of Bartlett’s test, with variance explained of 64.721%. This satisfies the conditions of exploratory factor analysis.

The factor of credibility (TRU) includes 3 following variables:

Our services is better than you expected (TRU21)

You trust the quality of our services (TRU23)

Our price is reasonable (TRU20)

In addition, the satisfaction is also be analyzed, the result is (Appendix 5) as follows: KMO = 0.751, sig.= 0.000 of Bartlett’s test. This satisfies the conditions of exploratory factor analysis. A factor is drawn with the variance explained of 82.470%.

Therefore, the factor satisfaction (SAT) incudes 3 variables:

How satisfied are you with our services (SAT01)

How is our response to your requests (SAT02)

How is the manner of services of the hotel (SAT03)

4.2.2 Testing the factors by using Cronbach Alpha

After performing the exploratory factor analysis, the factors have been drawn.

Testing the factors by using Cronbach Alpha for each factor in order to measure it a set of questions drawn from each factor has connection or not. Many researchers agreed that when the Cronbach Alpha has value of 0.6 or higher, then it can be used. The value of corrected Item – Total correlation must be greater than 0.3 (Hoang Trong and Chu Nguyen Mong Ngoc, 2005).

25

The Cronbach Alpha of factor 1 (Appendix 6) has value of 0.801; all 5 variables have the value of Corrected Item-Total Correlation greater than 0.3, satisfying the conditions of the test.

The Cronbach Alpha of remaining factors is as follows: 0.731 for factor 2 with 5 variables; 0.724 for factor 3 with 3 variables; 0.648 for factor 4 with 4 variables. All the variables of factor above have the value of corrected Item – Total correlation greater than 0.3, satisfying the conditions of the test. After performing the exploratory factor analysis for the service quality scale, four factors have been drawn and each was tested by Cronbach Alpha, satisfying the conditions for further analysis.

The credibility is also tested by Cronbach Alpha. The value of Cronbach Alpha is 0.726. All 3 variables in the factor of credibility have the value of corrected Item – Total correlation greater than 0.3, satisfying the conditions for further analysis.

The satisfaction with 3 variables is tested by Cronbach Alpha. The value of Cronbach Alpha is 0.894, the value of corrected Item – Total correlation is greater than 0.3, satisfying the conditions for further analysis.