CHAPTER 2: ELECTRONIC BANKING SERVICES AT AGRIBANK TAY DO
2.4. Survey results of model affecting the quality of e-banking services at Agribank Tay Do branch
2.4.1. Data surveyed on customers
2.4.1.1. Statistics of customers participating in the survey about ages, genders, education background and income per month
Table 2.14: Information statistics about customer groups at Agribank Tay Do branch
(Source: Summarized by author) The total surveyed was 250 people through the simple sampling method with 25 votes excluded due to lack of data and 225 valid votes, complete with necessary information. The table above shows that there is a certain difference in the distribution of gender, age, education, and income. People surveyed are scattered among age groups, but the highest is 25-40 years old (equivalent to 47.6%) because this is the age group that has a stable source of income from work and contact with modern banking facilities frequently. The proportion of other groups is 4.4% under 18 years old, 28.4% from 18-25 years old, 19.6% over 40 years old. The majority of respondents graduated from high school.
2.4.1.2. Statistics of sources of information customers know about e-banking services at Agribank Tay Do branch
Table 2.15: Information statistics about the awareness of e-banking at Agribank Tay Do branch
(Source: Summarized by author) The survey result shows that out of 225 survey participants, there were 296 answers, of which 92 answered that they know a banking service at the branch through a counselor, accounting for 31.1% of respondents 87 of the respondents know the service through the media at a rate of 29.4%. 100 customers know most of e-banking services at Agribank Tay Do branch through friends, equivalent to 33.8%. Finally only 17 people learn of information about services at Agribank Tay Do from other sources, equals to the rate of 5.7%.
2.4.1.3. Statistics of electronic banking services being used at Agribank Tay Do branch
Figure 2.9: Ratio of users in e-banking products at Agribank Tay Do branch
(Source: Summarized from survey)
Through the statistical results, there are 117 people out of 338 which are using SMS Banking services at Tay Do branch, accounting for 34.6%. Internet Banking has 104 options with 30.8%, ranked second with E-mobile Banking and Bankplus with 74 options and 43 options, accounting for 21.9% and 12.7% respectively. From the collected information, it can be seen that the level of customers' use of E-mobile Banking and Bankplus products and services is much lower than SMS Banking and Internet Banking.
2.4.2. Analyzing of the factor model affecting to the quality of electronic banking services at Agribank Tay Do branch
By theoretical basis combined with the hypotheses about the model, an expected model and description table of the specific observed variables as follows:
STF = β0 + β1REB + β2RPS + β3TGB + β4CPC + β5EPT + β6FL + Ui
The observed component of independent variables is measured by:
Table 2.16: Variable description for customers’satisfaction model
(Source: Survey from Agribank Tay Do branch) 2.4.2.1. Assessing the reliability of the Cronbach’s Alpha scale
In order to check the reliability of the observed variables and remove the non- conforming scales, the author uses the method of assessing reliability through Cronbach’s Alpha coefficient.
Table 2.17: Measurement results by using Cronbach’s Alpha
(Source: Summarized from data surveyed) Based on the results of the reliability test from Cronbach's Alpha, variables with Corrected Item-Total Correlation <0.3 and Cronbach's Alpha <0.6 will be considered as garbage variables and removed from the scale (Nunally and Burnstein, 1994). ..
According to the table, the measurement results have identified two observations including RPS2 (confirming completed transactions) and TGB1 (convenient account opening place) that need to be removed due to Corrected Item-Total Correlation of 0.169 and 0.212 (< 0.3). After removing these two observed variables, Cronbach’s Alpha of the total scale of RPS variable is 0.903> 0.6 and TGB of 0.848> 0.6 is required.
The results of the above analysis show that all 6 independent variables including reliability, responsiveness, tangible facilities, service capacity, empathy and price all affect the satisfaction of Customers for electronic banking services at Agribank Tay Do branch. The sample model is demonstrated as follows:
Figure 2.10: Variables of evaluating customers’satisfaction model
(Source: Summarized by author) 2.4.2.2. Phân tích nhân tố khám phá EFA
Exploratory factor analysis (EFA) is a method that reduces and observes the factors into a more meaningful and less quantitative factors for subsequent regression analysis.
Table 2.18: Exploratory Factor Analysis results
(Source: Summarized from SPSS) The analysis of survey factors (EFA) was conducted with 23 observed variables of 6 independent factors affecting customer satisfaction with the quality of e-banking services at Agribank Tay Do branch (because 2 variables RPS2 and TGB1 are removed due to
non-compliance). The EFA will evaluate two important values, convergent and discriminant. The conditions for EFA to be satisfied are:
- Factor Loading > 0.4 - 0.5 KMO 1 - Sig. < 0.05
- Total Explained Variance 50%
Kết quả phân tích dữ liệu ở bảng trên cho thấy:
KMO = 0.575 [0.5;1]
Sig. = 0.000 < 0.05
Total Explained Variance (% cumulative) = 79.203% > 50%
Eigenvalue = 1.628 > 1 2.4.2.3. Correlation analysis
To analyze the linear relationship between the independent and dependent variables, the author used the correlation test with Pearson coefficients. The absolute value of the Pearson coefficient indicates the severity of these relationships. The result displayed is a square matrix of correlation coefficients. The correlation of a variable with itself has a correlation coefficient of 1 and that is the diagonal of the matrix. Each variable will appear twice in the matrix with the same correlation coefficient in the two upper and lower triangles, symmetrically across the diagonal of the matrix. The condition for there to be no correlation between the independent variables is that the Pearson coefficient between the independent and the independent variables must be <0.05, so that the dependent variable and the independent variables have a correlation, the Pearson coefficient between the secondary variables and each dependent variable must be greater than 0.05.
Table 2.19: Correlation results
Accepted
(Source: Summarized from SPSS) The Pearson test result shows that all Pearson coefficients between the independent variables are smaller than 0.05 and Sig. (2-tailed) between these variables are greater than 0.05 indicates that there is no correlation between independent variables. Meanwhile, Pearson coefficient between dependent variable (STF) and independent variables are greater than 0.05 demonstrates the correlation between STF and independent variables.
Based on this result, it can be concluded that 6 independent variables included in the model can all explain the development of electronic banking services.
2.4.2.4. Regression analysis
Regression analysis is an important step to identify how independent variables affect to dependent variable. Regression model is the type of model used to describe this relationship.
Table 2.20: Regression analysis results
(Source: Summarized from SPSS) When conducting multivariate regression analysis on the quality of E-banking services, there are only 6 factors: reliability (REB), responsiveness (RPS), tangibles (TGB), service capacity (CPC), empathy (EPT), service fee levels (FL). Regression results show that the model has Sig. = 0.000 < 0.05, so that all 6 factors are statistically significant. Therefore, these factors all really affect to customer satisfaction with e- banking services or the quality of e-banking services at Agribank Tay Do branch.
According to the regression results, the reliability factor most affects strongly to customer satisfaction for the quality of e-banking services at Agribank Tay Do, followed by responsiveness, service capacity, empathy, tangibles and ultimately service fee levels.
The sample regression model is rewritten as follows:
STF = 0.063 + 0.011REB + 0.010RPS + 0.012TGB + 0.011CPC + 0.009EPT + 0.011FL + Ui
From the regression model results, it is possible to evaluate the impact of 6 independent variables on customer satisfaction when using e-banking services including Reliability (REB), Service Capacity (CPC), Tangibles (TGB), Responsiveness (RPS),
Empathy (EPT) and Service Fee Levels (FL). Six factors have a relatively uniform impact level, however, tangibles such as technology, infrastructure related to e-banking services are factors that are valued and highly valued by customers, the next is reliability of e-banking services, service capacity of employees for customers and the costs that customers must pay to use services. The regression model also shows that R2 = 0.972, which means that 97.2% of customer satisfaction change is explained by the independent variables in the model, the remaining 2.8% is explained by other variables out of model.
Adjusted R2 = 0.943 < R2 demonstrates that the model was not overblown.
The suitability test of the model
With Sig. = 0.000 < 0.05 Linear regression model is suitable.
Multi-collinearity test
The model has no multi-collinearity when the VIF (Variance Inflation Factor) is smaller than 10.
Table 2.21: Collinearity analysis
(Source: Summarized from SPSS) The results show that VIF of independent variables are smaller than 10, so these independent variables are not closely related to each other and no multi-collinearity occurs. From this, it is reasonable to assume that the relationship between the
independent variables does not significantly affect the interpretation of this model, and the components are involved in the development of electronic banking services.
Residual test: The P-P Plot diagram shows that view points do not disperse to far from the expected straight line, so the normal distribution is not violated.
Figure 2.11: Normal Regression Standardized Residual
(Source: Summarized from SPSS) The independence of residual: Using Durbin-Watson statistical coefficient to verify. This quantity is a value from 0 to 4 which is acceptable. Extracted from the result, the Durbin-Watson coefficient reaches 1.746, which almost equal to 2, thus concluding the residuals independently of each other.
Therefore, the linear regression model built does not violate the necessary assumptions in linear regression.