Chapter I. THE PROBLEM AND ITS BACKGROUND
IV. ANALYSIS AND FINDINGS OF DATA
The descriptive statistics Identifying customer needs
Mean
Accurate and on-time delivery 2,11
Price competitiveness 2,12
Full ranges of services 2,14
Available business networks 2,27
Staff competence 2,30
Simple procedure 2,66
Willingness to help customers 2,88
Ensuring commitments 3,01
Clear and easily-understood documents 3,18 Modern equipment and technology 3,20 Valid N (listwise)
Table 3. Descriptive statistics of customer expectations
The Table 3 shows for customers needs factors that they expect most in the following order are (1): Accurate and on-time delivery; (2) Price competitiveness; (3) Full ranges of services. When processing data, the choice are arranged in the order of the most expectations in scale of 1 to 3 and the other opinion is marked 4 to calculate.
More of that, the Table 3 shows that 5 scales with the lowest MEANs show the possible level of customers’ agreement a little different in the observed variables.
Those are POR01 – the bank provides a full range of services (Mean: 3.42); CON03 - The bank has convenience branch/locations (Mean: 3.43); CON02 - The bank has simple procedure for doing business (Mean: 3.49); POR02 - The bank is always a pioneer in providing new services (Mean: 3.54); TAN02 - The bank has well- displayed posters, brochure and handouts (Mean: 3:56)
Factors influencing to the customer satisfaction Factor of quality of service
Minimum Maximum Mean Std. Deviation
TAN01 3 5 4,15 ,743
COU05 3 5 4,08 ,744
INT01 2 5 4,06 ,815
INT03 1 5 4,24 ,891
INT04 2 5 4,60 ,584
Valid N (listwise)
Table 4. Descriptive statistics of the service quality scales
In 21 scales of service quality (Table 4), 5 scales that have been agreed by customers are as INT04 - the bank considers customers rights as their prime concerns (Mean: 4.60); INT03 - The bank has annual meeting to show appreciation for customers’ contribution (Mean: 4.24); TAN01 – The bank has modern equipment and technology (Mean: 4.15); COU05 – the bank staff are polite and consistently courteously with customers (Mean: 4.08); INT01 – The bank has hotline 24/24 for customers’ inquiries answers (Mean: 4.06).
Price competitiveness
Minimum Maximum Mean Std. Deviation
PRI01 1 5 3,81 1,017
PRI02 1 5 3,80 1,028
PRI03 2 5 4,25 ,846
Valid N (listwise)
Table 5. Descriptive statistics of the price competitive scales
Result of descriptive statistics (Table 5) shows that customers agree most on PRI03 - The bank has flexible pricing policies (Mean: 4.25) and PRI01 - The bank offer competitive interest rates (Mean: 3.81). Most of customers are voted for the price PRICE COMPETITIVE factor which can be decided that customers are very interested in the price when deciding to use the service (in line with customer expectations results as presented in Table 3)
Corporate image
Descriptive Statistics
Minimum Maximum Mean Std. Deviation
IMA01 1 5 3,51 1,077
IMA02 2 5 3,87 ,885
IMA03 1 5 3,70 ,982
IMA04 3 5 4,66 ,618
Valid N (listwise)
Table 6. Descriptive statistics of the corporate image scales
The Table 6 shows that for measuring corporate image, most customers agree two scales IMA04 - The bank launches marketing activities effectively and efficiently (Mean: 4.66), and IMA02 - The bank honors commitment to customers (Mean: 3.87) Measurement of customer satisfaction
Descriptive Statistics
Minimum Maximum Mean Std. Deviation
Overall service quality 3 5 4,73 ,548
Ability to meet customer needs
3 5 4,70 ,502
Continuity to use bank services
3 5 4,71 ,566
Valid N (listwise)
Table 7. Descriptive statistics of customer satisfaction scales
The Table 7 (through the survey) shows the level of customer satisfaction to the bank is very high, in which the three observed variables of customer satisfaction scale have Mean > 4.7 and customers confirm to continue to use bank services in the future.
This is a good signal for the performance and reputation of the bank for years. This also requires that the bank needs to make more efforts to further improve the service quality and maintain customer satisfaction.
Analysis of the measurement scales Cronback’s alpha
Reliability analysis by Cronbach’s alpha is a common measure of internal consistency (reliability) of a test or scale. Internal consistency describes the extent to which all the items in a test measure the same concept or construct and hence it is connected to the inner-relatedness of the items within the test (Tavakol et al., 2011).
The value of alpha (a) may be between negative infinity and 1. However, only positive values of alpha have meaning. In general, alpha coefficient ranges in value from 0 to 1, and the increase of this value means that the correlations between the items increase (Amit and Choudhury, 2010) which have Cronbach’s alpha coefficient greater than or equal to 0.6 will be accepted.
Besides assessing the reliability of scales, Cronbach’s alpha analysis also helps to check whether any item is not consistence with the rest of the scale through item-total correlations. Variables which have greater than 0.3 item-total correlations will be accepted; the others which have smaller than 0.3 item-total correlations will be eliminated from analysis data.
A commonly accepted rule of thumb for describing internal consistency using Cronbach's alpha is as follows, (George and Mallery, 2003; Kline, 1999). however, a greater number of items in the test can artificially inflate the value of alpha (Cortina, 1993) and so this rule of thumb should be used with caution:
Cronbach's alpha Internal consistency
α ≥ 0.9 Excellent
0.8 ≤ α < 0.9 Good 0.7 ≤ α < 0.8 Acceptable 0.6 ≤ α < 0.7 Questionable 0.5 ≤ α < 0.6 Poor
α < 0.5 Unacceptable
RELIABILITY ANALYSIS
Item-total Statistics Scale Scale Corrected
Mean Variance Item- Alpha if Item if Item Total if Item Deleted Deleted Correlation Deleted
CON01 13,5745 9,4319 ,5696 ,7349 CON02 13,6667 8,9524 ,6289 ,7145 CON03 13,7163 9,4618 ,4812 ,7634 POR01 13,7305 9,1268 ,5693 ,7339 POR02 13,9362 8,8602 ,5332 ,7478
Alpha = ,7798
TAN01 11,4296 5,6936 ,3398 ,7309 TAN02 12,0141 4,2126 ,4661 ,6802 TAN03 11,7042 4,1956 ,5945 ,5877 TAN04 11,5845 4,2729 ,6237 ,5725
Alpha = ,7121
COU01 11,5141 4,7197 ,7928 ,7842 COU02 11,6690 4,8187 ,7180 ,8194 COU03 11,5000 5,2872 ,6957 ,8264 COU05 11,2254 5,9489 ,6387 ,8507
Alpha = ,8604
INT01 12,7042 3,1601 ,4181 ,5694 INT02 12,8944 2,9320 ,4101 ,5801 INT03 12,5211 2,9038 ,4406 ,5543 INT04 12,1620 3,6970 ,4434 ,5738
Alpha = ,6384
REL01 7,2042 2,9013 ,6050 ,5724 REL02 7,3239 3,3837 ,5083 ,6885 REL03 7,3028 2,7516 ,5450 ,6521
Alpha = ,7274
PRI01 8,0493 2,4302 ,4756 ,5411 PRI02 8,0634 2,4144 ,4698 ,5504 PRI03 7,6056 2,9356 ,4561 ,5749
Alpha = ,6533
IMA01 7,5634 2,6023 ,5099 ,6568 IMA02 7,2113 3,2316 ,4770 ,6856 IMA03 7,3803 2,6061 ,6189 ,5091
Alpha = ,7119
CSL01 9,4155 ,9254 ,5278 ,7625 CSL02 9,4366 ,9428 ,6011 ,6843 CSL03 9,4296 ,7858 ,6731 ,5944
Alpha = ,7653
Table 8. Results of the analysis of Cronbach's coefficient alpha
The Table 8 shows as follows:
For the factor of CONVENIENCE, the three variables have correlation coefficients of variables fit>0.3. Therefore, these are selected. Meanwhile, the SERVICE PORTFOLIO factor does not match and it is removed. However, when CONVENIENCE factor are combined with SERVICE PORTFOLIO factor, a set of 5 variables CON01, CON02, CON03, POR01 and POR02 have correlation coefficients of > 0.3 and a high coefficient of 0.7611 Alpha (the alpha of the others of CON01, CON02, CON03 are at 0.6779) suitable for analysis implementation. Therefore, the CONVENIENCE factor is a combination of two variables of CONVENIENCE and SERVICE PORTFOLIO.
For the factor TANGIBLES, all variables have correlation coefficients > 0.3 and alpha coefficient > 0.6 (0.7121) and this is satisfied the creditability requirement to be analyzed.
For the COURTESY, the variable COU04 does not match correlation coefficient of 0.1410 < 0.3, and it is removed from the analysis. The other variables with correlation coefficients of > 0.3 and the Alpha coefficient of 0.8604 are reaching to analyze. (Are used for further research/analysis)
For the CUSTOMER CONTACT, four variables are matching the correlation coefficient of higher 0.3 and the Alpha coefficient 0.6384 and are use to analyze.
For the REALIABILITY factor, three variables have correlation coefficients of >
0.3 and the Alpha coefficient 0.7274 are suitable for analysis
For the PRICE COMPETITIVENESS, all variables are eligible for the reliability analysis (correlation coefficient of > 0.3 and the Alpha coefficient 0.6533) should be included in the analysis
For the CORPORATE IMAGE, the IMA04 variable with correlation coefficient of 0.0551 < 0.3 will be removed, three other variables including IMA01, IMA02, IMA03 reaches the satisfactory correlation coefficient of > 0.3 and the Alpha coefficient of 0.7119 should fit into the analysis
For the CUSTOMER SATISFACTION factors, three variables meet the requirements of the correlation coefficient of > 0.3 and the Alpha coefficient 0.7653 are also selected to be included in the analysis.
Therefore, there are 26 variables (Table 8) of 7 scales included in the analysis in comparison with 28 variables factor in the 8 initial scales (2 variables COU04, and IMA04 are removed). In addition, three variables of the level of customer satisfaction are also done in the Exploratory Factor Analysis
Exploratory Factor Analysis (EFA)
Exploratory factor analysis is a powerful statistical technique which is used for data reduction and summarization. The sampling adequacy of factor analysis is base on Kaiser-Meyer-Olkin (KMO) Measure.
By performing exploratory factor analysis, investigator can decide the number of factors to extract in the model. The Kaiser creation states that investigator should use a number of factors equal to the number of the eigenvalues of the correlation matrix that are greater than one (DeCoster, 1998).
An important part in exploratory factor analysis is interpreting factor matrixes.
This research will use Varimax rotation process to produce multiple group factors.
Factor loadings which indicate correlations between the variables and the factors are required to have greater than 0.5 values. Then, a factor can be interpreted in terms of the variables that have high load on it. The factor loading is less than 4.4 will continue
to be excluded from the group variables to ensure convergence between the variables in a factor; stops when Initial Eigenvalue greater than 1 and the total variance larger than 0.5. (Gerbing and Andserson, 1988). In this study, Extraction method - Principal Axis Factoring with Promax rotation and Regression method will be used for the EFA. The EFA is carried out with the following steps:
Step 1: All variables that have passed the test of reliability is put into the EFA analysis (26 variables on customer satisfaction and three variables on the customer satisfaction measurement). This process is called the first EFA (Appendix 2B) with the following results:
For factors influencing to customer satisfaction: KMO is 0.776 and 2 variables INT03 and PRI03 has been removed (factor loading <0.45), other variables will be taken into the 2nd second EFA.
For the level of customer satisfaction: KMO achieved at 0.665, Eigenvalue> 1 and total variance > 50% (54.058%) are qualified to be analyzed.
Therefore, the EFA result of the level of customer satisfaction (Appendix E5) shows 3 variables CSL01, CSL 02, and CSL 03 with factor loading > 0.45 and are used to explain the measurement scale of customer satisfaction logically.
Step 2: With the 24 variables that haves been done with the first EFA will be do the 2nd the second EFA (Appendix 3C) for KMO results of .765 and one variable are also removed (TAN01)
Step 3: The third EFA has been done with the 23 variables (Appendix 4D) and provides the results as follows:
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy. ,765
Approx. Chi-Square 1240,104
df 253
Bartlett's Test of Sphericity
Sig. ,000
Total variance : 51.49%
Number of factors : 6
COU: consists of four variables of COU and 1 of INT
CON: consists of 3 variables of COU and 2 of POR
REL: consists of 3 variable of REL and 2 of INT
TAN: consists of 3 variable of TAN
IMA: consists of 3 variable of IMA
PRI: consists of 2 variable of PRI
Rotated Component Matrixa
Factor
1 2 3 4 5 6
COU01 ,908
COU02 ,787
COU03 ,793
COU05 ,687
INT02 ,686
CON01 ,683
CON02 ,777
CON03 ,475
POR01 ,686
POR02 ,561
REL01 ,790
REL02 ,612
REL03 ,673
INT01 ,654
INT04 ,455
TAN02 ,551
TAN03 ,710
TAN04 ,805
IMA02 ,517
IMA01 ,619
IMA03 ,913
PRI01 ,796
PRI02 ,515
Table 9. EFA result
The generalized research model The research model
After analyzing with Cronbach's Alpha and EFA, research models are adjusted containing six independent variables (Courtesy, Convenience, Reliability, Tangibles, Corporate image, and price competitiveness) to measure the dependent variable - the customer satisfaction. All six variables are has impact to the increase or decrease the customer satisfaction with the scale as described in Table 9 and the generalized research model will be adjusted as follows:
Figure 9. The overall research model Hypotheses
H1: The more convenient, the higher satisfaction
H2: The better tangibles, the higher satisfaction
H3: The better courtesy, the higher satisfaction
H4: The better price competitiveness, the higher satisfaction
H5: The higher reliability, the higher satisfaction
H6: The better business image, the higher satisfaction
Courtesy
Reliability
Price competitive
Corporate image Convenience
s Tangibles
Customer satisfaction
Testing the research model
Pearson’s correlation coefficient analysis
Correlation between variables is a measure of how well the variables are related.
(http://www.statisticshowto.com). The most common correlation measurement in statistics is the Pearson Correlation (technically called the Pearson Product Moment Correlation or PPMC), which shows the linear relationship between two variables.
Two letters are used to represent the Pearson correlation: Greek letter rho (ρ) for a
population and the letter “r” for a sample.
Results are between -1 and 1. A result of -1 means that there is a perfect negative correlation between the two values at all, while a result of 1 means that there is a perfect positive correlation between the two variables. A result of 0 means that there is no linear relationship between the two variables. There is a very rarely correlation of 0, -1 or 1. (http://www.statisticshowto.com).
The closer the value of r gets to zero, the greater the variation the data points are around the line of best fit (High correlation: .5 to 1.0 or -0.5 to 1.0; Medium correlation: .3 to .5 or -0.3 to .5; Low correlation: .1 to .3 or -0.1 to -0.3) (http://www.statisticshowto.com).
If correlation between two variables related, the multi collinear regression analysis should be concerned. In Pearson correlation analysis, there is no distinction between the independent variables and the dependent variable. In multiple regression analysis, the multi-collinear is accessed by SPSS with Collinearity Diagnostics (http://www.statisticshowto.com).
As in the matrix of independent variables (Table 10), COURTESY factor is the most strongly correlated with the RELIABILITY (0.307). Later on, the
CONVENIENCE factor was significantly correlated with PRICE COMPETITIVENESS factor (0.258) and CORPORATE IMAGE factor is most correlated to the COURTESY (0.260), while TANGIBLES factor is most correlated with SERVICE PORTFOLIO factor (0.285).
In addition, LEVEL OF CUSTOMER SATISFACTION factor has a very tight linear correlation with all six independent variables (COURTESY, CONVENIENCE,
RELIABILITY, CORPORATE IMAGE, TANGIBLES, and PRICE
COMPETITIVENESS). Therefore, with the independent variables with the weak linear correlation of Pearson’s coefficient <0.3 (SERVICE PORTFOLIO and RELIABILITY are excluded) are satisfied to analyze with regression analysis and the multi-collinear of two variables of REL and COU are also concerned.
Correlations
COU CON REL IMA PRI TAN SHLCSL
COU
Pearson
Correlation 1 ,214(*) ,307(**) ,260(**) ,178(*) ,285(**) ,620(**) CON
Pearson
Correlation ,214(*) 1 ,145 ,228(**) ,258(**) ,172(*) ,564(**) REL
Pearson
Correlation ,307(**) ,145 ,228(*) ,176(*) ,216(**) ,198(*) ,576(**) IMA
Pearson
Correlation ,260(**) ,228(**) ,176(*) 1 ,244(**) ,195(*) ,604(**) PRI
Pearson
Correlation ,178(*) ,258(**) ,216(**) ,244(**) 1 ,186(*) ,608(**) TAN
Pearson
Correlation ,285(**) ,172(*) ,198(*) ,195(*) ,186(*) 1 ,542(**) CSL
Pearson
Correlation ,620(**) ,564(**) ,576(**) ,604(**) ,608(**) ,542(**) 1
* Correlation is significant at the 0.05 level (2-tailed)
** Correlation is significant at the 0.01 level (2-tailed)
Table 10. Results of Pearson analysis on factor of the Customer Satisfaction With only three variables on the customer satisfaction, they also correlate closely with the Pearson coefficient in all variables at > 0.4. Therefore, there happens a phenomenon of the multi-collinear and that is unsuitable for regression analysis, in
which customer satisfaction is always associated with the ability to meet customer needs of (0.814) and continuity to use bank services (0.871)
Overall service quality
Ability to meet customer needs
Service continuity
Customer satisfaction Overall service quality Pearson
Correlation 1 ,424(**) ,520(**) ,719(**)
Ability to meet customer needs
Pearson
Correlation ,424(**) 1 ,621(**) ,814(**)
Service continuity with bank
Pearson
Correlation ,520(**) ,621(**) 1 ,871(**)
Customer satisfaction Pearson
Correlation ,791(**) ,814(**) ,871(**) 1
** Correlation is significant at the 0.01 level (2-tailed)
Table 11. Results of Pearson analysis (Correlation) of Customer Satisfaction factor Regression analysis
Regression analysis is a modeling technique for analyzing the relationship between dependent variable (customer satisfaction) and independent variables (tangible, reliability, responsiveness, assurance, and empathy). Then, base on regression function, we can assess the impact of each independent variable on dependent variable as well as predict the change in dependent variable when there is any change in independent variables.
At first, it is necessary to test assumptions for regression analysis. The principal assumption is that there is a linearity of the relationship between dependent and independent variables. This research investigates the model with more than one independent variables, the correlation among independent variables (multi- collinearity) should be checked through Variance inflation factor (VIF). Regression model accept variables which have VIF smaller than 10. In addition, it is assumed that the error terms 8 are independent, normally distributed random variables with mean value of 0, and constant variances. As long as these assumptions are not seriously violated, regression model will be established.
Once regression function was given, the research can investigate relationship between service quality and customer satisfaction at the bank. R-square (coefficient of determination) will provide a goodness-of-fit measure. With higher R-square value, the model is higher fit for analysis.
Regression analysis will determine the causal relationship between the dependent variables (SATISFACTION) and the independent variables (COURTESY, CONVENIENCE, RELIABILITY, CORPORATE IMAGE, TANGIBLES, and PRICE COMPETITIVENESS). The regression analysis model will describe the kind of the relationship and thereby it is easily to predict the extent of the dependent variables with the informed value of the independent variable. The selected method is Stepwise method with PIN standard 0.05 and the POUT standard 0.1. The Table 10 are shown the analysis results.
Evaluation of the adaptable rate of the multiple linear regression model
Determination coefficient R2 has been shown to be a function without reduction of the number of independent variables in the model (6 variables). However, the model usually do not match the actual data as value R2 (0.989). For this issue, R2 is adjusted R2 is used to reflect more closely adaptable rate of the multiple linear regression model (0.988) because it is independent with the inflation deviation of R2. When comparing the adjusted 2 R2 and R2 values in table Table 12 it is shown that R2 is adjusted smaller and used it to evaluate the suitability of the model which is more securable without the exaggerate of the suitability of the model. Therefore, the 0.988 adjusted R2 shows that the compatibility of the model with variables is very large and the dependent variable of customer satisfaction is great explained by the six independent variables in the model.
Testing the model
The F test used in the variance analysis of is a test of the hypothesis on the adaptability of the overall linear regression model to explore the dependent variables with a linear relationship of all independent variables. In the Table 12, it shows that the F- statistic value is calculated from the full R2 value (not rezo), sig. value s which are small. From that consequence, the model used is appropriate and the variables achieve acceptable standards (Tolerance> 0.0001).
In addition, the item of Collinearity diagnostics with Variance inflation factor VIF of independent variables in the model were < 2 (1 to 1.182) representing that the Collinearity of the independent variables are not significant and the variables in the model are accepted.
At last, Durbin Watson coefficients used to test the correlation chain shows that the model that does not violate when using the multiple regression method as d value is achieved 1.944 (almost 2) and accept the hypotheses of no - correlation chain in the model. Therefore, the multiple regression reaches the measurement and variables in the model are accepted.
Testing the interpretation of the equation
From the regression analysis (Table 12), it shows the relationship between the dependent variable of customer satisfaction and the six independent variables presented in the following equation:
CSL= -,102+0.309PRI+0.296IMA+0.286REL+0.275CON+0.272COU+0.245TAN In which:
CSL: Customer Satisfaction
PRI: The price competitiveness
IMA: Corporate Image
REL: Reliability
CON: Convenience
COU: Courtesy
TAN: Tangibles
According to the above regression equation shows that customer satisfaction has a linear relationship with the price competitiveness factor (standardized beta coefficient is 0.309), the corporate image (standardized beta coefficient is 0.296), the Reliability (standardized beta coefficient is 0.286), the convenience (standardized beta coefficient is 0.275), the courtesy (standardized beta coefficient is 0.272), and the tangibles (standardized coefficient is 0.245).
Furthermore, the standardized Beta coefficients are > 0, meaning of which independent variables has positively impact to the customer satisfaction. These results also confirm the hypotheses laid out in the research model (H1-H6) accepted and tested accordingly. Therefore, the banks must strive to improve these factors to enhance the customer satisfaction.
variables entered/removed(a)
Model Variables Entered
Variables
Removed Method
1 COU . Stepwise (Criteria: Probability-of-F-to-enter <= ,050,Probability-of-F-to- remove>=,100).
2 PRI . Stepwise (Criteria: Probability-of-F-to-enter <= ,050,Probability-of-F-to- remove>=,100).
3 IMA . Stepwise (Criteria: Probability-of-F-to-enter <= ,050,Probability-of-F-to- remove>=,100).
4 REL . Stepwise (Criteria: Probability-of-F-to-enter <= ,050,Probability-of-F-to- remove>=,100).
5 CON . Stepwise (Criteria: Probability-of-F-to-enter <= ,050,Probability-of-F-to- remove>=,100).