RESEARCH METHODOLOGY

3.1. Research design

3.1.1. Selection of research models

To make a survey, the authors selected models perceived service quality was building based on SERVQUAL model (Parasuraman et al, 1985).

3.1.2. Questionnaire Design Process

The design phase questionnaire:

The questionnaire was designed and tested through a number of stages to ensure that the information needed to collect reliable service of process data analysis. Develop questionnaire-based raw information to be collected in the theoretical model and the relevant prior studies.

In this step the author also consult experts.

3.1.3. Variable construct

There are 22 items distributing in 5 factors. The customer satisfaction variable these study uses scale of Parasuraman et al, 1985. The table 3.1 shows the variables and its items.

Table 3.1 Variables and items.

Variable Items Code

Reliability

VietcombankATM guarantee security issue RE1

VietcombankATM Provide timely service as committed RE2 The deposit products and services of VietcombankATM were

performed accurately RE3

Vietcombank ATM Provide complete information, accurate

and timely RE4

ATM guarantee security issue RE5

Responsiveness

Diverse products and services to meet customer needs RS1

There are many customer care program RS2

Professional staff to handle rapid accurate RS3 Staff actively advise suitable products for customers RS4

Variable Items Code

Assurance

Staff ready to serve and assist customers AS1

Resolve complaints quickly and rational AS2

Guide to procedures for clients fully understandable AS3 The staff is professional and qualifications AS4

Empathy

Enthusiastic, considerate and fun staffs EM1 Employees interested in the individual requirements of the

customer EM2

Always listen to customers EM3

Always polite, friendly customer EM4

Tangible

The location of transaction is convenience and in accordance

with customer needs TA1

Spacious transaction, using modern technology TA2 Nice layout of the counter, identifiable for Customers

reasonable TA3

The staff dresses polite TA4

Paper forms to be used in transactions were designed easily

TA5

Customer satisfaction

VietcombankATM offers more benefits than my cost CS1 VietcombankATM is the best among competing banks CS2 VietcombankATM is much better than what I expected CS3

Overall I satisfied with service of Vietcombank CS4

Likert scale was used to measure the perceived level of customer and their satisfaction about the quality of ATM services, which from 1 "strong disagree ' to 5 “'strong agree”. The survey was conducted customers during the period from May to July, 2015 in Hanoi.

3.2. Sample

Population: The purpose of the study was to assess the level of customer satisfaction of with ATM of Vietcombank. So the study sample was identified as customer who are using ATM service of Vietcombank

First samples taken at random from 200 to 400 samples, then based on the standard deviation and inference to determine the samples to be taken. Within the scope of this study with

Samples collected will be conducted in accordance with the statistical classification variables.

The basis for classification include: Gender, age, education, work placements and work experience. Also scoring average, maximum value, minimum value and standard deviation of the answers in the questionnaires collected.

3.3.2. Scaling

Factors inspection is done by Cronbach`s Alpha scales. These variations are observed not guarantee minimum reliability will be removed from the scale and does not appear in the analysis of factors. In this study, Cronbach's alpha coefficient was taken as the minimum at 0.6. Variables Cronbach's alpha coefficient of less than 0.3 is considered junk variable.

3.5.3. Factor analysis

After the observed variables were tested using Cronbach's Alpha scale, they will be included in the explore factor analysis (EFA). Factor analysis will help authors observed variables collapsed into more significant variables in explaining the research model.

Some standards apply when analyzing EFA research as follows:

- Inspection of the suitability of factor analysis of sample data via the KMO. If greater than 0.5, the KMO factor analysis is appropriate, whereas if the value is less than 0.5 KMO factor analysis method is not appropriate.

- The number of factors: The number of factors that are determined based on the index eigenvalue represent the fraction of variation explained by each factor.

- The Explained variance criteria: Total variance extracted must be greater than 50%.

- Criticized coefficient method Principal components factor with Varimax rotation to ensure that the amount is the smallest factor.

3.3.4. Analysis of variance

For the test of the difference between the overall research expertise this study use T-test and analysis of variance (ANOVA), testing is also used to directly compare the value p- corresponding value. To consider the difference occurs in the category involving three or more groups are using Post Hoc Test by Tukey to assess value.

To test the relationship among factors and job satisfaction of employees in the research model, this study is using methods of correlated with the correlation coefficient "Pearson correlation coefficient", that is denoted by "r ". Values in the range between -1 ≤ r ≤ 1.

If r> 0 expressed positively correlated. Conversely, r <0 represents the inverse correlation.

Value r = 0 indicates that the two variables do not have a linear relationship.

1: The relationship between the two variables tighter 0: the relationship between two variables as weak

The level of significance "sig" of the correlation coefficient, as follows:

- <5%: fairly tight correlation - <1%: very tight correlation

The next step, using methods of multiple linear regression to determine the meaning and linear correlation of the variables in the model.