SCALE ANALYSIS

Valid N (listwise) 142

Table 4.5 Descriptive statistics on customer satisfaction

4.2 SCALE ANALYSIS

29

4.2.1 Cronbach’s alpha

Cronbach’s alpha is a statistical test to check the closeness and the correlation between observed variables. It is involved in the two aspects including the correlation of variables themselves and the correlation of the score of every single observed variable with the total score of variables of each respondent.

This method allows analyst to eliminate the inappropriate variables and minimize the amount of unwanted variables in the model. The reason is that, if not doing that, we cannot know variability and deviation of variables. According to the method, only variables with Corrected Item-Total Correction > 0.3 and the Alpha coefficient > 0.6 are acceptable and suitable to be analyzed in the next step (Nunnally and BernStein, 1994). According to many researchers, if a scale has Cronbach’s alpha >= 0.8, it is a good scale and its level of correlation will be higher. Looking at the Table 4.6, we can know the result of reliability analysis as follows:

As for the factor of CONVENIENCE, all 3 observed variables have the Corrected Item- Total Correction > 0.3 so they are chosen. Meanwhile the factor SERVICE CATALOGUE does not satisfy the condition, so it is eliminated. However, when combining two factors CONVENIENCEand SERVICE CATALOGUE, the 5 observed variables STT01, STT02, DMDV01, DMDV02 have the suitable corrected item-total correction > 0.3 and their Alpha coefficient = 0.7611, which is high (while Alpha coefficient of every single variables STT01, STT02, STT03 is only 0.6779). Thus, they are eligible to be included in factor analysis. In conclusion, the factor of CONVENIENCE is the combination of measured variables of two smaller factors that are CONVENIENCEand SERVICE CATALOGUE.

The reason is that SERVICE CATALOGUE is a component of CONVENIENCE.

As for the factor of TANGIBLES, all the observed variables have the Corrected Item- Total Correction > 0.3 and Alpha coefficient > 0.6 (0.7121),so they meet the requirement of reliability to be analyzed in factor analysis.

About the factor of SERVICE STYLE, the variable PCPV04 is not eligible because Corrected Item-Total Correction of 0.1410 < 0.3, so it is eliminated. Other variables have corrected item-total correction > 0.3 and Alpha coefficient = 0.8604, which is qualified to be included in factor analysis.

As for the factor of APPROACHING THE CUSTOMERS, 4 variables are qualified to be analyzed in exploratory factor analysis because they have Corrected Item-Total Correction > 0.3 and Alpha coefficient = 0.6384.

About the factor of CREDIBIITY, 3 observed variables have Corrected Item-Total Correction > 0.3 and Alpha coefficient = 0.7274 so they are qualified to be included in

30

exploratory factor analysis.

As for the factor of PRICE COMPETITIVENESS, all the variables qualify the condition of reliability testing (Corrected Item-Total Correction > 0.3 and Alpha coefficient = 0.6533) so they can be included in exploratory factor analysis.

As for the factor of CORPORATE IMAGE, the observed variable of HADN04 has unsuitable corrected item-total correction 0.0551<0.3, so it is eliminated. The other 3 variables including HADN01, HADN02, HADN03 have corrected item-total correction >

0.3 and Alpha coefficient = 0.7119, so they are qualified to be included in exploratory factor analysis.

About the factor of CUSTOMER SACTISFACTION, 03 observed variables have corrected item-total correction > 0.3 and Alpha coefficient= 0.7653 so they are included in exploratory factor analysis.

In conclusion, in total, there are 26 variables (Table 4.6) of 7 scales compared with 28 observed variables of 8 initial scales (PCPV04, HADN04 are eliminated). In addition, the three variables measuring the customer satisfactionis also included in exploratory factor analysis

RELIABILITY ANALYSIS

Item-total Statistics

Scale Scale Corrected

Alpha

Mean Variance Item-

if Item if Item Total if Item

Deleted Deleted Correlation Deleted

STT01 13,5745 9,4319 ,5696 ,7349

STT02 13,6667 8,9524 ,6289 ,7145

STT03 13,7163 9,4618 ,4812 ,7634

DMDV01 13,7305 9,1268 ,5693 ,7339

DMDV02 13,9362 8,8602 ,5332 ,7478

Alpha = ,7798

SHH01 11,4296 5,6936 ,3398 ,7309

SHH02 12,0141 4,2126 ,4661 ,6802

SHH03 11,7042 4,1956 ,5945 ,5877

SHH04 11,5845 4,2729 ,6237 ,5725

Alpha = ,7121

PCPV01 11,5141 4,7197 ,7928 ,7842

PCPV02 11,6690 4,8187 ,7180 ,8194

PCPV03 11,5000 5,2872 ,6957 ,8264

31

PCPV05 11,2254 5,9489 ,6387 ,8507 Alpha = ,8604

TXKH01 12,7042 3,1601 ,4181 ,5694

TXKH02 12,8944 2,9320 ,4101 ,5801

TXKH03 12,5211 2,9038 ,4406 ,5543

TXKH04 12,1620 3,6970 ,4434 ,5738

Alpha = ,6384

STN01 7,2042 2,9013 ,6050 ,5724

STN02 7,3239 3,3837 ,5083 ,6885

STN03 7,3028 2,7516 ,5450 ,6521

Alpha = ,7274

TCTG01 8,0493 2,4302 ,4756 ,5411

TCTG02 8,0634 2,4144 ,4698 ,5504

TCTG03 7,6056 2,9356 ,4561 ,5749

Alpha = ,6533

HADN01 7,5634 2,6023 ,5099 ,6568

HADN02 7,2113 3,2316 ,4770 ,6856

HADN03 7,3803 2,6061 ,6189 ,5091

Alpha = ,7119

SHL01 9,4155 ,9254 ,5278 ,7625

SHL02 9,4366 ,9428 ,6011 ,6843

SHL03 9,4296 ,7858 ,6731 ,5944

Alpha = ,7653

Table 4.6: Result of analyzingCronbach’s Alpha

4.2.2 Exploratory factor analysis

Exploratory factor analysis is an analyzing technique to summarize and reduce the data, which is necessary to identify the set of variables for the research. The correlation between groups of variables is considered as the number of basic factors. Each observed variable will be calculated a ratio called factor loading. Each factor loading will tell the researcher each measured variable belongs to which factors.

In exploratory factor analysis, a necessary condition is the indicator KMO (Kaiser-Meyer –Olkin (KMO) must has big value (0.5<KMO<1), which shows that the exploratory factor analysis is suitable. If the KMO < 0.5, the exploratory factor analysis can be unsuitable for data. According to Gerbing & Anderson, 1988, additionally, for satisfying the condition of exploratory factor analysis, factor loading of each variable must be greater than 0.45, the stop point at Eigenvalue (representing the variation explained by each factor) > 1 (default in the SPSS program) and the total variance must be > 50%. When conducting exploratory

32

factor analysis, the author uses extraction method, which includes Principal Axis factoring and Promax rotation and the Regression.

The process of exploratory factor analysis is presented as follows:

Step 1:

The collection of observed variables satisfies the condition of reliabilityand included in the exploratory factor analysis (26 measured variables of factors affecting customer satisfaction and 03 observed variables measuring the level of customer satisfaction). This process is called the first exploratory factor analysis (APPENDIX 2) with the following results:

About the factors affecting customer satisfaction: KMO = 0.776 and there are two variables are eliminated that are TXKH03 and TCTG03 (factor loading < 0.45). The rest of other variables will be analyzed in the second exploratory factor analysis.

About the level of customer satisfaction: KMO = 0.665, Eigenvalue > 1, total variance >

50% (54.058%), which is eligible to conduct exploratory factor analysis. In conclusion, the result of exploratory factor analysis of the level of customer satisfaction (APPENDIX 5) shows that 3 observed variables SHL01, SHL02, and SHL03 all have factor loading > 0.45 and it is appropriate to use to explained the scale of the level of customer satisfaction.

Step 2:

Qualified observed variables in the first exploratory factor analysis (24 variables) included in the second exploratory factor analysis (APPENDIX 3) show that the KMO is slightly reduced to 0.765 and another observed variable is eliminated (SHH01).

Step 3:

The third exploratory factor analysis (APPENDIX 4) includes the rest of 23 observed variables and has the following results:

KMO: 0.765

EIGENVALUE: 1.221 Total variance: 51.49%

The number of factors: 6

PCPV: including 4 variables of PCPV and 1 variable of TXKH STT: including 3 variables of PCPV and 2 variables of DMDV STN : including 3 variables of STN and 2 variables of TXKH SHH: including 3 variables of SHH

HADN: including 3 variables of HADN TCTG : including 2 variables of TCTG

33

Pattern Matrix(a)

Extraction Method: Principal Axis Factoring. Rotation Method: Promax with Kaiser Normalization. a Rotation converged in 6 iterations.

Table 4.7 The result of explanatory factor analysis

4.3 GENERAL RESEARCH MODEL