After the scale has been adjusted and supplemented by qualitative research, a pilot survey was conducted with 30 customers. The purpose of this study is to adjust and perfect the official questionnaire. The adjustment method is carried out through two main techniques, the Cronbach's Alpha reliability coefficient method (Cronbach, 1951) and EFA exploratory factor analysis. Accordingly, variables with item-total correlation coefficients below 0.30 in Cronbach's Alpha analysis will be eliminated, observed variables with weights (Factor loading) less than 0.50 in EFA will continue to be eliminated and the total square root (≥ 50%) will be tested (Nunnally & Burnstein, 1994). The remaining variables (complete scales) will be included in the questionnaire for official quantitative research.
Step 2: Formal quantitative research
After conducting preliminary quantification, the official scale results will be included in the survey questionnaire used for official quantitative research. The data collection results will be included in multivariate regression analysis to test the hypotheses and research models. The results of this research step aim to discover the model of factors affecting customer loyalty using BIDV Dong Sai Gon banking services.
The completed survey questionnaire after being evaluated by quantitative method includes 22 specific observation variables: 8 observation variables on perceived service quality, 3 observation variables on satisfaction, 4 observation variables on trust, 3 observation variables on corporate image, and 4 observation variables on loyalty. Personal information such as gender, age, income, and educational level are also designed in the questionnaire to compare the differences of these factors when evaluating the impact of factors on customer loyalty to DVNHBL.
For all observed variables of the scales, to assess the level of customer agreement, the author uses a 5-point Likert scale from (5): completely agree to (1): completely disagree.
3.2. Research process
Identify the research problem
Research objectives
Theoretical basis:
- Theory of Customer Loyalty
- Factors affecting loyalty
- Relevant studies have been conducted
Draft scale
The Thi Thien An Institute of Qualitative Research
- Discuss and ask for expert opinion
-
-
Mock interview
Calibrate and add scales
Official scale
Quantitative research
- Select model
- Data collection
Data processing
- Reliability analysis: Check alpha coefficient
- EFA factor analysis: Eliminate variables with small variable-total correlation coefficients, Eliminate variables with small EFA weights.
- Regression analysis
- Hypothesis testing
Research results
Conclusion and recommendations
Figure 3.1 Research process
3.3. Building and adjusting the scale
As presented in the previous sections, the scales in this study are based on theories and existing scales in the world. They are adjusted and supplemented to suit the implementation of this study in Vietnam based on the results of qualitative research with group discussion techniques. There are 5 research concepts used in this study, which are (1) Perceived service quality, (2) Satisfaction, (3) Trust, (4) Bank image, (5) Loyalty.
3.3.1. Perceived service quality scale
Perceived service quality is measured and adjusted according to the service quality assessment criteria of Parasuraman, Zeithaml and Berry (1991). As a result, the scale of perceived service quality includes 8 observed variables:
CL1: The bank looks quite modern and convenient CL2: Bank staff uniforms are neat
CL3: The bank provides services on time as committed and without errors.
CL4: The bank provides fast and accurate service. CL5: Bank staff are enthusiastic in helping customers. CL6: Bank staff are always polite, courteous and qualified.
CL7: Bank employees are able to understand and grasp customer needs. CL8: The bank shows personal attention to customers (sending birthday gifts, cards and New Year calendars...).
3.3.2. Satisfaction scale
To measure customer satisfaction with retail banking service quality, the study used a scale developed by Fornell (1992) consisting of three observation variables, measuring overall satisfaction with the bank, the extent to which customer expectations are reinforced, and the gap between customers' perceptions of banking services after consumption experience and the ideal banking model they desire.
HL1: I found this to be the ideal bank to transact with HL2: The bank performed as expected
HL3: I am satisfied with the quality of service of the bank.
3.3.3. Confidence scale
The customer trust scale was built and adjusted according to the research of Delgago (2003) - Lin and Wang (2006). In which, the observation variable "I find the bank meets my expectations" was omitted because this variable has a similar meaning to the variable in the satisfaction scale. As a result, the trust scale includes 4 observation variables:
TC1: My bank is honest and sincere in taking care of my needs TC2: I can trust my bank to handle my financial problems.
TC3: I find the Bank to have vision. TC4: I find the Bank to understand the market.
3.3.4. Bank image scale:
The bank image scale was built according to the original scale of Lee, 2004, including 3 observed variables:
HA1: The Bank's advertisement is very impressive.
HA2: The bank's transaction manager is a good leader. HA3: The bank treats its employees kindly.
3.3.5. Customer loyalty scale for DVNHBL
Loyalty scale is a measure of repeat purchasing behavior with trust, love and willingness to introduce to friends and relatives when they have a need. The loyalty scale is built according to the original scale of Yoo & ctg (2000) including 4 observed variables:
LTT1: I don't like to change to another bank because I appreciate this bank.
LTT2: I will continue to do business with this bank even if other banks have more incentives.
LTT3: I would recommend these banking services to my friends and relatives LTT4: The bank is my first choice when thinking about banking services
3.4. Selecting research samples:
3.4.1. Sampling method
The research sample size is based on Tabachnick's formula N≥50 +8p, where p is the number of independent variables (Nguyen Dinh Tho, 2011, page 499). The model has 4 independent variables, so the minimum sample size is 82 samples.
The sample was selected by convenience and non-proportional method. There were a total of 350 questionnaires (Appendix 2) sent to customers via email (100), and inside/outside the transaction counter (250) at BIDV Dong Sai Gon. And to ensure that the interviewees are the appropriate group for the survey in the field of retail banking services, who have been and are transacting at BIDV Dong Sai Gon, the customers will be asked first about the place of transaction, the frequency and time of using traditional services such as deposit, withdrawal, savings, ATM card, internet banking, SMS banking... As a result, the author collected 215 valid samples from direct interviews and 74 valid samples from emails. Thus, the final sample size for processing is 289.
3.4.2. Description and analysis of research sample
From 350 questionnaires distributed via email and direct survey, the author collected 314 questionnaires (response rate reached 89%). After eliminating inappropriate questionnaires due to having many blank boxes or checking many boxes, the author collected 289 questionnaires that met the requirements, which was 82 more than the required sample size. Thus, the number of collected samples ensured representativeness for the research.
Table 3.2 Research sample information
Sample information
Characteristic | Quantity | Rate (%) | |
Sex | Male | 136 | 47.06 |
Female | 153 | 52.94 | |
Total | 289 | 100.00 | |
Age | <30 | 185 | 64.01 |
30-50 | 76 | 26.30 | |
>50 | 28 | 9.69 | |
Total | 289 | 100.00 | |
Income | < 4 million | 85 | 29.41 |
4-10 million | 136 | 47.06 | |
> 10 million | 68 | 23.53 | |
Total | 289 | 100.00 | |
Education level | No college degree | 54 | 18.69 |
College/University | 197 | 68.17 | |
Above university | 38 | 13.15 | |
Total | 289 | 100.00 | |
Job | Office | 142 | 49.13 |
Business | 81 | 28.03 | |
Other | 66 | 22.84 | |
Total | 289 | 100.00 |
Maybe you are interested!
-
Factors affecting customer loyalty to international payment services at Asia Commercial Joint Stock Bank - 12 -
Solutions to increase individual customer loyalty at Vietnam Joint Stock Commercial Bank for Industry and Trade by 2018 - 2 -
Relationship marketing factors affecting customer loyalty to Vietnam Joint Stock Commercial Bank for Investment and Development in Ho Chi Minh City - 2 -
Factors affecting customer loyalty to retail banking services at Dong A Commercial Joint Stock Bank, District 5 branch - 2 -
Evaluating customer satisfaction with savings deposit services at National Citizen Commercial Joint Stock Bank NCB Tan Huong Transaction Office - 3
Sex
Male
Female 47%
53%
Age |
>50 T |
10% |
30-50 T |
26% |
<30 T |
64% |
Income
Postgraduate 13% | Level College/University 68% | No college degree 19% |
> 10 < 4 | ||
million million | ||
24% 29% | ||
4-10 | ||
million | ||
47% |
Job
Other 23%
Business 28%
Office 49%
Figure 3.2: Diagram showing characteristics of the research sample
Regarding gender : among 289 valid customers, there were 136 male customers (accounting for 47.06%) and 153 female customers (accounting for 52.94%). Thus, the gender ratio between men and women in the research sample did not have much difference.
In terms of age : there are 185 customers under 30 years old (accounting for 64.01%), 76 customers from 30-50 years old (accounting for 26.30%) and 28 customers over 50 years old (accounting for 9.69%). The age structure of the sample shows that the majority of the surveyed sample is young people.
Regarding income : 85 customers have income under 4 million VND (accounting for 29.4%), 136 customers have income from 4 to 10 million VND (accounting for
47.6% rate), and 68 customers have income over 10 million VND (accounting for 23.5%).
Regarding education level : there are 54 customers who have not attended college or university (accounting for 18.7%), the number of people with college or university degrees is relatively high (68%), and 38 people with university degrees (accounting for 13%).
Regarding occupation: The survey subjects were mainly office workers at 49% (equivalent to 142 customers), business subjects accounted for 28% (equivalent to 81 customers), the rest were from other professions.
3.5. Scale evaluation:
3.5.1. Cronbach Alpha reliability analysis
The scales are tested for reliability using the Cronbach Alpha tool. Cronbach's coefficient is a statistical test of the degree of rigor with which the items in the scale are correlated with each other, helping to eliminate inappropriate variables and scales. Many researchers agree that when Cronbach Alpha is from 0.8 or higher to nearly 1, the measurement scale is good, from 0.7 to nearly 0.8 is usable. There are also studies suggesting that Cronbach Alpha from 0.6 or higher can be used in cases where the concept being measured is new or new to the respondents in the research context (Hoang Trong & Chu Nguyen Mong Ngoc, 2005).
From there, the author tested the reliability of the scale based on the observed variables with item-total correlation coefficients less than 0.3 will be eliminated and the scale selection criteria when Cronbach Alpha is 0.7 or higher.
The results of reliability testing of the scales showed that all scales achieved acceptable reliability, so they were all used in the next EFA and regression analysis steps (see details in Appendix D).