# Evaluation of loan service quality for individual customers at National Commercial Joint Stock Bank - 7

Through Table 2.7, we remove the observed variable CT2: "Bank employees serve all customers fairly when they come to the transaction" because there is a correlation coefficient of the total variable less than 0.3. In addition, all the remaining observed variables have a total correlation coefficient greater than 0.3. All coefficients Cronbach's Alpha if the variable type is not greater than Cronbach's Alpha. Besides, all Cronbach's Alphas are greater than 0.6. The above results have met the requirements for evaluating a reliable scale.

2.2.2.2. EFA . exploratory factor analysis

- Factor analysis of independent variables

Before conducting exploratory factor analysis to extract the factors affecting the quality of lending services for science and technology at NCB - Hue from observed variables, I tested the appropriateness of the data. data through the KMO test (Kaiser - Meyer - Olkin) has a value of 0.5 or more and Bartlett's test results in p-value less than 0.05. From the collected data, I conduct exploratory factor analysis. We hypothesize H0: there is no relationship between the observed variables.

Table 2.8: KMO and Bartlett's Test

 KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.814 Bartlett's Test of Sphericity Approx. Chi-Square 1199,470 DF 253 Sig. 0.000

Maybe you are interested!

Source: SPSS data processing

With the KMO test result of 0.814 is greater than 0.5 and Sig. of Bartlett's test is less than 0.05 (observed variables are correlated with each other in the population), thus rejecting H0. The results of EFA analysis showed 5 basic factors. The total variance extracted is 61.555% > 50%, indicating that these 5 factors explain 61.555% of the data variation and the Eigenvalues ​​of the factors are all greater than 1. Bartlett's test has Sig value. = 0.000 < 0.05 should meet the requirements. In this test, no variable is excluded from the model because the factor loading system is > 0.5.

The result has 5 factors with the total variance extracted is 61.555%; ie, the ability to use these 5 factors to explain 23 observed variables is 61.555% (> 50%).

This group of 5 factors is described as follows:

Table 2.9: Result of independent variable factor analysis

 Sign Observed variables Factor loading factor first 2 3 4 5 NLPV1 The bank promptly and fully responded to your loan amount. 0.838 NLPV4 Credit staff fully knowledgeable about banking products and services 0.800 NLPV3 Credit staff answer all your questions enthusiastically and fully 0.739 NLPV2 Credit officers handle loan procedures quickly 0.729 NLPV5 The bank has a hotline serving 24/7 0.681 DU3 Time to review documents, disburse quickly and timely 0.824 DU2 Flexible and suitable loan terms and conditions 0.704 DU1 Simple and clear loan process and procedures 0.695 DU5 Reasonable and competitive loan interest rates and service fees 0.693 DU4 Credit officers provide full information about personal loan services 0.677 TC3 The credit officer executes the transaction correctly 0.751 TC2 The bank builds trust and peace of mind for you. 0.744 TC1 The bank performs the loan service as committed 0.734 TC4 Credit officers protect your personal information well 0.694
 TC5 The bank satisfactorily handles your complaints satisfactorily 0.587 PTTH4 Bank staff have neat and polite clothes 0.831 PTH3 Banks arrange transaction counters, reasonable and convenient signs 0.826 PTTH1 The bank has a convenient transaction location 0.731 PTTH2 The bank has modern facilities and equipment 0.545 CT3 The credit officer executes the transaction correctly 0.850 CT4 Credit officers protect your personal information well 0.775 CT1 The bank performs the loan service as committed 0.721 CT5 The bank satisfactorily handles your complaints satisfactorily 0.669 Eigenvalues 5,809 2.949 2.465 1.618 1.315 Misquote (%) 25,259 38,082 48,801 55,836 61,555 Cumulative Variance (%) 25,259 12,824 10,718 7,035 5.719

Source: SPSS data processing

This factor should be named Responsiveness, denoted by DU.

The fourth factor is drawn with Eigenvalue = 1.618, which explains 55.836% of the variation of the data. This factor has Factor Loading index with variables PTHH1 has Factor Loading of 0.731, PTHH2 has Factor Loading 0.545, PTHH3 has Factor Loading 0.826, PTHH4 has Factor Loading 0.831 . Should name this factor as Tangible means, denoted by PTHH.

The fifth factor is drawn with Eigenvalue = 1.315, which explains 61.555% of the variation of the data. This factor has a Factor Loading index with variables CT1 has a Factor Loading of 0.721, CT3 has a Factor Loading of 0.850, CT4 has a Factor Loading of 0.775, and CT5 has a Factor Loading of 0.669. This factor should be named Sympathy Level, denoted by CT.

- Factor analysis of dependent variable

We hypothesize H0: there is no relationship between the observed variables of the general satisfaction scale on the quality of science and technology lending services. KMO test is 0.685 > 0.5 and and Sig. of Bartlett's test is less than 0.05, thus rejecting H0. Thus, between the observed variables there is a large enough relationship needed for exploratory factor analysis.

Table 2.10: KMO and Bartlett's Test of Dependent Variables

 KMO and Bartlett's Test Kaiser-Meyer-Olkin Measure of Sampling Adequacy. 0.685 Bartlett's Test of Sphericity Approx. Chi-Square 87,800 DF 3 Sig. 0.000

Source: Data processing using SPSS

Table 2.11: Results of factor analysis of dependent variable

 Sign Observed variables Factor loading factor HL3 Would you recommend the Bank's personal loan service to others? 0.831 HL1 In general, you are satisfied with the quality of the Bank's personal loan service 0.826 HL2 In the coming time, will you continue to use the personal loan service of the Bank? 0.787 Eigenvalues 1,992 Extracted Variance (%) 66.392

Source: Data processing using SPSS

The drawn factors have factor loading coefficients all > 0.5. Factor loading coefficients are all high, variables in the same group all have strong load on the factor it measures, the smallest is 0.787. Therefore, not a single component is removed. The total variance extracted is 66.392% > 50%, showing that the explanation is quite high.

The results also show that one factor is extracted and Eigenvalue > 1. There is no separation or displacement of the factors, so there is no change in the number of factors. This factor is drawn with the Eigenvalue = 1,992, this factor explains 66,392% of the variation of the data. This factor has Factor Loading index with variables HL1 has Factor Loading of 0.826, HL2 has Factor Loading 0.787, HL3 has Factor Loading 0.831. This factor should be named Satisfaction, denoted by HL.

2.2.2.3. Regression analysis

- Correlation analysis

The first step in linear regression analysis is to consider the linear correlations between the dependent variable and each independent variable, as well as between the independent variables. The larger the correlation coefficient between the independent variable and the dependent variable, the greater the linear relationship, and the linear regression analysis can be suitable. On the other hand, if there is also a large correlation between the independent variables, it is also a sign that multicollinearity may occur between them in the model we are considering.

Table 2.12: Testing the correlation between independent variable and dependent variable

 HL NLPV DU TC PTH CT HL Correlation coefficients first Sig. NLPV Correlation coefficients 0.32 first Sig. 0.00 DU Correlation coefficients 0.426 0.00 first Sig. 0.00 0,500 TC Correlation coefficients 0.433 0.00 0.00 first Sig. 0.00 0.00 0,500 PTH Correlation coefficients 0.285 0,500 0.00 0.00 first Sig. 0.00 0.00 0,500 0,500 CT Correlation coefficients -0.092 0.00 0.00 0.00 0.00 first Sig. 0.148 0,500 0,500 0,500 0,500

Source: Data processing using SPSS

Table 2.12 shows that all variables have Sig significance level. < 0.05, except for the variable Level of sympathy with significance Sig. > 0.05 means that the variable Level of sympathy is not correlated with the variable Satisfaction. Therefore, remove the Sympathy variable from the model before regression analysis. After conducting exploratory factor analysis, grouping variables according to each factor, we conduct regression. The applied regression model is a multivariable regression model (multiple regression model) to consider the relationship between the dependent variable and the independent variables. When analyzing regression, the results will show the factors affecting satisfaction when using science and technology lending services at NCB-Hue and their impact level.

Specifically, regression analysis was performed with 4 independent variables: (1) Service capacity, (2) Responsiveness, (3) Reliability, (4) Tangible means and auxiliary variables. of satisfaction. One-pass input method (Enter method) was used for regression analysis. The values ​​of the factors used to run the regression are the mean values ​​from the factors. The model is written as follows:

HL= 0 + 1*NLPV + 2*DU + 3*TC + 4*PTHH

Inside:

0: coefficient of freedom

ßi: partial regression coefficient corresponding to independent variables.

HL: the value of the dependent variable is customer satisfaction about the quality of science and technology lending services

NLPV: The first independent variable value is Service capacity

DU: The second independent variable value is Responsiveness

TC: The third independent variable value is Confidence level

PTTH: The fourth independent variable value is Tangible Means

- Evaluate the fit of the regression model

To evaluate the fit of the model, we use the adjusted coefficient of determination R2. The coefficient of determination R2 adjusted for this model is 55.3%, showing that 4 independent variables in the model explain 55.3% of the variation of the dependent variable. With this value, the fit of the model is acceptable.

Table 2.13: Summary model using the Enter . method

 Paradigm CHEAP R2 R2 adjustable Standard error of the estimate Durbin-Watson first 0.744 0.553 0.539 0.67893804 1,713

Source: Data processing using SPSS