(Nguyen Dinh Tho 2011, page 240). The initial number of questionnaires sent out for collection was 300. The collected questionnaires were screened and checked for validity and conformity to the research scope standards, then the remaining number of questionnaires processed was 200.
Data collection period: From April 1, 2014 to June 30, 2014.
3.3 Regression analysis and regression model testing
3.3.1 Descriptive statistics of data
Questionnaire sent to customers is surveyed in the form of: Direct survey to each customer.
Direct survey: 300 survey forms were sent to 3 branches and affiliated transaction offices in Ho Chi Minh City of MHB Bank, namely Saigon, Cho Lon, Gia Dinh. The results obtained 234 responses (response rate 78%), of which 200 were valid and 34 responses were eliminated due to incomplete or invalid responses.
Thus, the total number of valid samples collected through direct survey of 200 samples is , satisfying the condition that the number of samples must be greater than 145 above, reaching an average of 200/29 = 6.9 observed samples / variable.
3.3.2 Descriptive statistics of qualitative variables
Descriptive statistics of qualitative variables (Appendix C) are as follows:
Regarding gender: In 200 people surveyed, there were 81 men accounting for 40.5%, 119 women accounting for 59.5%. The ratio of female and male customers using MHB bank's e-banking services is not much different, almost equal.
Regarding age: The survey results of the age of using e-banking services show that there are 25 customers under 25 years old, accounting for 12.5%, 116 customers aged 25 to 34, accounting for 58%, 46 customers aged 35 to 45, accounting for 23% and 13 customers over 45, accounting for 6.5%. Customers aged 25-35 and 35-45 account for a high percentage of using e-banking services, possibly because at this age, most customers are working people with stable incomes and are more likely to accept using e-banking services in daily transactions. The number of customers under 25 years old also using this service is quite high and customers over 45 years old are less.
Regarding education level: Statistical results on education level show that customers with college and university degrees account for the highest percentage of 92.5%, customers with postgraduate degrees account for 4.5%, customers with high school degrees account for a relatively low percentage of 3%.
About occupation: Statistics on occupation show that office workers make up the majority with 69.5%, followed by business people with 14.5%, the rest are distributed to other occupations, the lowest is students with 4%.
About income: Information about income shows that the percentage of customers with income from 5 to 10 million accounts for 57.5%, customers with income from 10 to 15 million accounts for 21%, customers with income over 15 million accounts for 16.5%, customers with income under 5 million is 5%.
3.3.3 Assessment of scale reliability
Assess the reliability of a scale based on the Cronbach's Alpha reliability coefficient. A scale is considered valid when it measures exactly what it is supposed to measure, meaning that the measurement method has no systematic or random deviations, specifically by measuring the same thing over and over again to produce stable results (including a reasonable error). A scale has high reliability when it has a Cronbach Alpha coefficient ≥ 0.6. At the same time, the variables must have an item-total correlation coefficient ≥ 0.3 (Hoang Trong and Chu Nguyen Mong Ngoc 2008, page 24).
3.3.3.1 Scale of component concepts
Table 3.1: Summary of Cronbach's alpha coefficient results of the scale of influencing factors
Variable
observe
Medium scale if variable type | Variance scale if variable type | Total variable correlation | Cronbach's alpha if variable type | |
1. Perceived usefulness (PU) | ||||
PU1 | 13.81 | 7.401 | .671 | .810 |
PU2 | 13.97 | 7,336 | .675 | .809 |
PU3 | 14.01 | 7,558 | .647 | .817 |
PU4 | 13.99 | 7,462 | .663 | .812 |
PU5 | 14.02 | 7,572 | .612 | .826 |
Cronbach's Alpha (PU): 0.846 | ||||
Maybe you are interested!
-
Summary of Cronbach'S Alpha Coefficient Results (See Appendix 4) -
Results of Testing Cronbach's Alpha Coefficient of Independent Variable -
Cronbach'S Alpha Coefficient Test Results of the Ability Scale -
Cronbach's Alpha Reliability Coefficient of Components of the Scale of Factors Affecting Investment Capital Attraction for Tourism in Ba Ria Vung Tau Province -
Reliability Testing Using Cronbach'S Alpha Coefficient
2. Perceived Ease of Use (PEU)
PEU1 | 17.25 | 14,761 | .751 | .902 |
PEU2 | 17.43 | 14,910 | .763 | .900 |
PEU3 | 17.34 | 15,070 | .727 | .905 |
PEU4 | 17.43 | 14,859 | .759 | .901 |
PEU5 | 17.31 | 14,418 | .754 | .902 |
PEU6 | 17.29 | 14,315 | .819 | .892 |
Cronbach's Alpha (PEU): 0.916 | ||||
3. Perceived Behavioral Control (PBC) | ||||
PBC1 | 7.19 | 2,087 | .718 | .815 |
PBC2 | 7.23 | 2,100 | .749 | .783 |
PBC3 | 7.22 | 2,293 | .730 | .803 |
Cronbach's Alpha (PBC): 0.857 | ||||
4. System Information (IF) | ||||
IF1 | 3.81 | .768 | .854 | .0 |
IF2 | 3.71 | .870 | .854 | .0 |
Cronbach's Alpha (IF): 0.921 | ||||
5. Subjective norm (SN) | ||||
SN1 | 10.36 | 5,629 | .511 | .671 |
SN2 | 10.41 | 5.308 | .559 | .642 |
SN3 | 10.07 | 5,699 | .595 | .628 |
SN4 | 10.27 | 5,907 | .419 | .726 |
Cronbach's Alpha (SN): 0.728 | ||||
6. Perceived Risk (PR) | ||||
PR1 | 12.63 | 30,605 | .792 | .945 |
PR2 | 12.76 | 30,696 | .818 | .941 |
PR3 | 12.70 | 29,998 | .886 | .933 |
PR4 | 12.85 | 29,626 | .880 | .934 |
PR5 | 12.86 | 31,317 | .836 | .939 |
PR6 | 12.92 | 31,405 | .842 | .939 |
Cronbach's Alpha (PR): 0.948 | ||||
Source: Data compiled from Appendix D
The results of the scale of 6 components with 26 observed variables of the scale of factors affecting the decision to use electronic banking show that:
Useful ingredients felt
The perceived usefulness scale has a Cronbach's Alpha reliability coefficient of 0.846 (> 0.6) and in which the component measurement variables all have small total variable correlation coefficients.
The most is 0.612 (>0.3). That also allows to conclude that these components meet the requirements and can be used for factor analysis in the next step.
Easy to use feel
The perceived ease of use component has an average Cronbach's Alpha value of 0.916 (>0.6) and the total item correlation coefficient meets the allowable standard, the lowest being 0.727 (>0.3). Therefore, all observations of this component are retained and will continue to be used for the EFA exploratory factor analysis in the next step.
Perceived behavioral control component
The Cronbach's Alpha reliability coefficient of this component is 0.857 (>0.6). The total correlation coefficients of the observed variables meet the lowest allowable standard of 0.718 (>0.3), so these variables are retained and used in the next EFA factor analysis.
System information components
The Cronbach's Alpha reliability coefficient of this component is 0.921 (>0.6). The total correlation coefficients of the observed variables meet the lowest allowable standard of 0.854 (>0.3). Thus, no variables are eliminated and they will continue to be used for the EFA exploratory factor analysis in the next step.
Subjective norm component
The Cronbach's Alpha reliability coefficient of this component is 0.728 (>0.6). The total item correlation coefficients meet the allowable standard, the lowest is 0.419 (>0.3), so these variables are still used in the next EFA factor analysis.
Perceived risk component
The perceived risk component has an average Cronbach's Alpha value of 0.948 (>0.6) and the total item correlation coefficient meets the allowable standard, the lowest being 0.792 (>0.3). Thus, no variables are eliminated and they will continue to be used for the EFA exploratory factor analysis in the next step.
3.3.3.2 Decision scale used
Table 3.2: Summary of Cronbach's alpha coefficients of the decision scale used
Observation variable
Average scale if variable type | Scale variance if variable type | Total variable correlation | Cronbach's alpha if excluded variable | |
1. Decision to use (Y) | ||||
Y1 | 13.81 | 7.401 | .671 | .810 |
Y2 | 13.97 | 7,336 | .675 | .809 |
Y3 | 14.01 | 7,558 | .647 | .817 |
Cronbach's Alpha (Y): 0.869 | ||||
Source: Data compiled from Appendix D
The decision-to-use component has a Cronbach's Alpha coefficient of 0.869 (>0.6), in addition, the total correlation coefficient of the observed variables all meet the allowable standard, the lowest is 0.665 (>0.3). Therefore, all observed variables of this component are used in the exploratory factor analysis (EFA) in the next step.
In summary, the results of testing the reliability of the electronic banking decision scales meet the requirements with Cronbach's Alpha coefficient >0.6. All 3 variables in this component will be used in the next EFA analysis.
3.3.4 Exploratory Factor Analysis (EFA)
After assessing the reliability of the scale using Cronbach's alpha coefficient and eliminating variables that do not ensure reliability, we proceed to perform exploratory factor analysis. Exploratory factor analysis is a technique used to reduce, summarize data and find relationships between variables. When performing factor analysis, we calculate two values:
+ Convergent value: Expressed through the percentage of extracted variance and factor loading. Specifically, the factor loading of variable x i must have a high value on the factor that x i is the measuring variable and low on other factors that x i is not supposed to measure.
+ Discriminant value: Shown through the number of extracted factors that are consistent with the initial hypothesis or not, at the same time the number of factors ensures the theory but the internal variables must also ensure the correct position compared to the theory.
Number standards.
+ Test the suitability of factor analysis with sample data through the Kaiser-Meyer-Olkin (KMO) statistic value. The value of KMO is large (0.5 KMO
1) is a sufficient condition for factor analysis to be appropriate, but if this value is less than 0.5, factor analysis may not be appropriate for the data (Hoang Trong and Chu Nguyen Mong Ngoc 2008, page 31).
+ Convergent validity : For the scale to achieve convergent validity, the largest factor loading coefficient (also known as factor weight) of each observed variable must be greater than or equal to 0.5 (Hair et al., 1998) .
+ Variance explained criteria : Total variance explained must be greater than 50% (Nunnally and Bernstein, 1994, quoted from Nguyen Thi Mai Trang and Nguyen Dinh Tho, 2009) .
+ Number of factors: The number of factors is determined based on the Eigenvalue index representing the variation explained by each factor. According to the Kaiser criterion, only factors with Eigenvalue greater than 1 are retained in the analytical model. Factors with Eigenvalue less than 1 will not have a better effect in summarizing information than an original variable, because after standardization, each original variable has a variance of 1 (Hoang Trong and Chu Nguyen Mong Ngoc 2008, page 34) .
3.3.4.1 Exploratory factor analysis of the scale of influencing factors
The scales consisting of a group of 6 factors with 26 initial observed variables after testing reliability by Cronbach's Alpha coefficient were all satisfied and were included in the exploratory factor analysis. The results of the EFA factor analysis are shown as follows:
Table 3.3: Summary of EFA factor analysis results of the impact factor scales
KMO and Bartlett's Test
Kaiser-Meyer-Olkin Measure of Sampling Adequacy.
.866 | |
Approx. Chi-Square | 3751.742 |
Bartlett's Test of Sphericity df | 325 |
Sig. | .000 |
Factor name
Observation variable | Group of factors | |||||
1 | 2 | 3 | 4 | 5 | ||
Perceived Risk (PR) | PR3 | 0.901 | ||||
PR4 | 0.884 | |||||
PR5 | 0.835 | |||||
PR1 | 0.834 | |||||
PR2 | 0.827 | |||||
PR6 | 0.827 | |||||
Perceived Ease of Use (PEU) | PEU6 | 0.834 | ||||
PEU4 | 0.816 | |||||
PEU2 | 0.800 | |||||
PEU5 | 0.738 | |||||
PEU1 | 0.721 | |||||
PEU3 | 0.706 | |||||
Behavior Control, System Information (PBC) | IF1 | 0.857 | ||||
IF2 | 0.822 | |||||
PBC1 | 0.701 | |||||
PBC2 | 0.626 | |||||
PBC3 | 0.619 | |||||
Perceived usefulness (PU) | PU1 | 0.807 | ||||
PU3 | 0.744 | |||||
PU2 | 0.726 | |||||
PU5 | 0.698 | |||||
PU4 | 0.686 | |||||
Subjective norm (SN ) | SN2 | 0.751 | ||||
SN3 | 0.730 | |||||
SN1 | 0.730 | |||||
SN4 | 0.616 | |||||
Individual value | 9,690 | 3,075 | 2.112 | 1,734 | 1,472 | |
Variance extract % | 18,640 | 16,699 | 12,777 | 12,294 | 9,138 | |
Add % | 18,640 | 35,339 | 48,117 | 60,410 | 69,549 | |
Source: Data compiled from Appendix E.
Through the results of EFA analysis, the KMO and Bartlett's Test coefficient of the scale of impact factors is quite high at 0.866 and satisfies the requirement of 0.5 ≤ KMO ≤ 1, with a significance level of 0 (sig=0.000), showing that EFA factor analysis is appropriate. The Eigenvalue level is 1.472 > 1, the observed variables in the table all have weights > 0.5. We have 5 factors extracted from 26 observed variables with a total extracted variance of 69.549%. The number
This shows that the ability to use these 5 factors to explain 26 observed variables is 69.549%, the extracted variance meets the requirement of greater than 50%.
First group of factors – Perceived risk
F1 = PR1+ PR2 + PR3 + PR4 + PR5
No observed variable has a factor loading less than 0.5, so it is not eliminated from the research model.
This group of factors refers to the level of perceived risk of each customer during the process of using e-banking services. The name of this factor is based on the original model as " Perceived risk ". The representative factor is denoted as PR, calculated by the average (Mean) of factors PR1, PR2, PR3, PR4, PR5.
Thus, it can be explained that when customers perceive the system's risk to be higher, they tend to reduce their use of e-banking services.
Second group of factors - Perceived ease of use
F2 = PEU1 + PEU2 + PEU3 + PEU4 + PEU5 + PEU6
These observed variables do not have a loading coefficient less than 0.5 so they are not removed from the model.
This group of factors refers to the individual's perception of the ease of operation and use of the service. The name of this factor is " Perceived ease of use ", with the representative factor denoted as PEU, calculated by the average (Mean) of factors PEU1, PEU2, PEU3, PEU4, PEU5, PEU6.
Thus, it can be explained that the easier the e-banking service is to use, the more customers tend to use the e-banking service.
Third factor group - Perceived behavioral control
F3 = PBC1 + PBC2 + PBC3 + IF1+ IF2
These observed variables do not have a loading coefficient less than 0.5 so they are not removed from the model.
The two components “ Perceived behavioral control ” and “ System information ” were merged into one. Thus, within the scope of customer research, these two factors were unified into one.