Home

Evaluation of Electricity Bill Collection Service Quality at Vietnam Technological and Commercial Joint Stock Bank - Hue Branch

2.2. ASSESSMENT OF THE QUALITY OF ELECTRICITY BILL COLLECTION SERVICES AT VIETNAM TECHNOLOGICAL AND COMMERCIAL JOINT STOCK BANK - HUE BRANCH

2.2.1. Descriptive statistics

With the data collected from the survey, after checking and cleaning the data, there were 110 valid ballots used as data for the following analysis. The results of synthesizing personal information of customers are described in detail as follows:

Table 2.4: Descriptive statistics of survey sample



Target

Absolute frequency (people)

Relative frequency (%)

Valid Frequency (%)

Cumulative frequency (%)

1. Classification by gender

+ Male

+ Female


36

74


32.7

67.3


32.7

67.3


32.7

100.0

2. Classified by age

+ 18 - 25 years old

+ 26 -35 years old

+ 36 - 55 years old

+ 55 -65 years old


2

14

63

31


1.8

12.7

57.3

28.2


1.8

12.7

57.3

28.2


1.8

14.5

71.8

100.0

3. Classification by income (VND/month)

+ Under 2 million

+ 2 - 5 million

+ 5 - 7 million

+ Over 7 million


3

31

51

25


2.7

28.2

46.4

22.7


2.7

28.2

46.4

22.7


20.7

30.9

77.3

100.0

4. Classification by occupation

+ Staff

+ Business

+ Workers

+ Other


57

25

8

20


51.8

22.7

7.3

18.2


51.8

22.7

7.3

18.2


51.8

74.5

81.8

100.0

Total

110

100.0

100.0


Maybe you are interested!

Evaluation of Electricity Bill Collection Service Quality at Vietnam Technological and Commercial Joint Stock Bank - Hue Branch

- Regarding gender: Thus, there are a total of 74 female customers, accounting for a very high proportion.

67.3% and male customers are 36 people, accounting for 32.7% of the total 110 customers surveyed. Through the above analysis, it can be seen that women are very interested in the service, often being the ones who manage the family's spending. This result also reflects the true nature of the electricity bill collection product.

- Regarding age: Table 2.4 above shows that there are 2 customers in the age group of 18 - 25 years old, accounting for 1.8%; 14 customers in the age group of 26 - 35 years old, accounting for 12.7%; 63 customers in the age group of 36 - 55 years old, accounting for 57.3% and the group of customers from 55 - 65 years old is 31 people, accounting for 28.2% of the total of four age groups surveyed. Through these figures, it can be seen that middle-aged customers are customers who are very interested in the service, as well as suitable for the financial capacity of customers.

- Regarding income: The results of the income survey of 110 customers show that: customers with income under 2 million VND are 3 people, corresponding to a rate of 2.7%; 31 people with income from 2 to 5 million VND, accounting for a rate of 28.2%; 51 people with income from 5 to 7 million VND with the highest corresponding rate of 46.4% and the income level over 7 million VND is 25 people, corresponding to a rate of 22.7%. The analysis shows that the financial capacity of customers is mainly in the range of 5 to 7 million.

- Regarding occupation: The occupations of the surveyed customers were divided into 3 groups, including: the staff group had 57 people, accounting for the highest percentage of 51.8%; the business group had 25 people, accounting for 22.7%; the workers group had 8 people, accounting for the lowest percentage of 7.3%; the remaining group of other occupations had 20 people, accounting for 18.2%. Through this result, it is clear that the need to use the electricity bill collection service is responded to by the staff, because of the nature of the work according to administrative hours, it is difficult to pay the electricity bill using the old method.

2.2.2. Analysis of scale reliability using Cronbach's Alpha coefficient

Cronbach's Alpha reliability test is used to eliminate garbage variables before conducting factor analysis. The reliability test of variables in the scale measuring electricity bill collection service quality and customer satisfaction is based on the Cronbach's Alpha coefficient of the scale components and the Cronbach's Alpha coefficient of each measurement variable.

Variables with a Corrected Item Total Correlation coefficient of less than 0.3 will be eliminated. A scale has good reliability when it varies within the range

[0.7 – 0.8]. If Cronbach's Alpha ≥ 0.6, the scale is acceptable in terms of reliability. (Nunnally and Bernstein (1994)). The analysis results are shown in Table 2.5.

- Tangibles: includes 4 observed variables (PT1, PT2, PT3, PT4), in which 3 variables PT1, PT2 and PT3 all have a Corrected Item-Total Correlation coefficient greater than 0.3, so they are accepted. However, the 4th variable (PT4) has a Corrected Item-Total Correlation coefficient less than 0.3, so it is not accepted. In addition, the Cronbach's Alpha coefficient of the scale is 0.637. If the PT4 variable is removed, the Cronbach's Alpha coefficient increases to 0.790 (> 0.6), so the Tangibles scale meets the requirements. Except for the PT4 variable, the remaining variables are included in the next factor analysis.

- Reliability: includes 4 observed variables (TC1, TC2, TC3 and TC4), in which the first 3 variables (TC1, TC2, TC3) all have a Corrected Item-Total Correlation coefficient greater than 0.3, so they are accepted. Only variable TC4 has a Corrected Item-Total Correlation coefficient of -0.096 (< 0.3), so it is not accepted. After removing variable TC4, the Cronbach's Alpha coefficient is 0.905 (> 0.6), so the Reliability scale meets the requirements.

- Responsiveness: 6 observed variables (DU1, DU2, DU3, DU4, DU5, DU6). All 6 variables have Corrected Item-Total Correlation greater than 0.3 so they are accepted. In addition, Cronbach's Alpha coefficient is 0.924 (> 0.6) so the Responsiveness scale meets the requirements.

- Service capacity : includes 4 observed variables: NL1, NL2, NL3, NL4. The first 3 variables (NL1, NL2, NL3) all have a total variable correlation coefficient greater than 0.3, so they are accepted. However, the 4th variable (NL4) has a total variable correlation coefficient of -0.090 (< 0.3), so the NL4 variable does not meet the requirements. After removing the NL4 variable, the Cronbach's Alpha coefficient is 0.875 (> 0.6), so the Service Capacity scale meets the requirements. The variables NL1, NL2 and NL3 are included in the next factor analysis.

- Sympathy: in the 4 observed variables (CT1, CT2, CT3, CT4), the first 3 variables (CT1, CT2, CT3) all have Corrected Item-Total Correlation greater than 0.3 so they are accepted. Variable E4 has a Corrected Item-Total Correlation coefficient of 0.137 (< 0.3) so it is not accepted. In addition, the Cronbach's Alpha coefficient of the scale is 0.639

and if E4 is removed, the Cronbach's Alpha coefficient increases to 0.737 (> 0.6), so the Empathy scale meets the requirements. These variables will be included in the next factor analysis.

Conclusion: After evaluating the reliability of the scale using Cronbach's Alpha coefficient, 4 variables were eliminated: PT4, TC4, NL4 and CT4.

Table 2.5: Results of Cronbach's Alpha coefficient analysis of scale components



Item-Total

Scale Mean if Item Deleted

Scale Variance if

Item Deleted

Corrected Item-Total

Correlation

Cronbach's Alpha if Item

Deleted

Reliability: Alpha = .744

TC1

9.38

2,705

.753

.554

TC2

9.70

2,394

.717

.568

TC3

9.53

2,432

.845

.482

TC4

11.69

5,172

-.096

.905

Response: Alpha = .924

DU1

17.94

11,915

.791

.910

DU2

18.65

10,788

.789

.909

DU3

18.57

10,698

.790

.909

DU4

18.53

10,900

.786

.909

DU5

18.61

11,015

.789

.909

DU6

18.02

12,162

.786

.912

Service Capacity: Alpha = .622

NL1

9.64

2,952

.716

.348

NL2

10.25

2,604

.574

.403

NL3

9.88

2,824

.655

.362

NL4

11.23

4,738

-.090

.875

Sympathy: Alpha = .639

CT1

9.95

2,898

.476

.530

CT2

10.38

2,741

.473

.529

CT3

10.33

2,421

.626

.403

CT4

11.43

3,815

.137

.737

Tangibles: Alpha = .637

PT1

11.26

2,914

.641

.390

PT2

11.35

2,842

.515

.490

PT3

11.25

3.162

.593

.440

PT4

11.60

5.179

-.018

.790

(Source: Extracted from data processing results using SPSS software)

2.2.3. Exploratory factor analysis EFA

After assessing the reliability of the scale using Cronbach's Alpha coefficient and eliminating variables with uncertain reliability, the study went into factor analysis. EFA is a statistical analysis method to reduce a set of interdependent observed variables into a smaller set of variables (called factors) so that they are more meaningful but still contain most of the information content of the original set of variables (Hair et al., (1998)). The basis of this reduction is based on the linear relationship between factors and observed variables.

This study used the Principal Components extraction method with Varimax rotation.

- Pay attention to the standard: │Factor Loading│maximum of each Item ≥ 0.5. At each Item, the difference between the maximum │Factor Loading│and any │Factor Loading│must be ≥ 0.3 (Jaboun and Al-Tamimi, (2003)).

- Total variance extracted ≥ 50% (Gerbing and Anderson, (1988)).

- KMO ≥ 0.5, Bartlett's test is statistically significant (Sig. < 0.05). The KMO coefficient (Kaiser-Meyer-Olkin) is an index used to examine the appropriateness of factor analysis. A large KMO value means that factor analysis is appropriate.

The authors Mayers, LSGamst, G.Guarino AJ (2000) mentioned that: “In factor analysis, the method of extracting Pricipal Components Analysis combined with Varimax rotation is the most commonly used method.”

According to Hair et al. (1998), Factor loading (factor loading coefficient or factor weight) is an indicator to ensure the practical significance of EFA:

- Factor loading > 0.3 is considered to reach the minimum level.

- Factor loading > 0.4 is considered important.

- Factor loading > 0.5 is considered to have practical significance.

The conditions for exploratory factor analysis must satisfy the following requirements:

- Factor loading > 0.5.

- 0.5 ≤ KMO ≤ 1.

- Bartlett's test is statistically significant (Sig. < 0.05). This is a statistical measure to examine the hypothesis that variables are not correlated in the population. If this test is statistically significant (Sig < 0.05), the observed variables are correlated with each other in the population.

- Percentage of variance > 50%: Shows the percentage of variation of observed variables. That is: considering the variation as 100%, this value shows how much % the factor analysis can explain.

2.2.3.1. Results of service quality scale analysis according to the SERVPERF model

Exploratory factor analysis was conducted with 18 items. The extraction method chosen for factor analysis was the Principal Components method with Varimax rotation. EFA factor analysis was used to re-evaluate the convergence of observed variables according to the components.

After conducting EFA, KMO and Bartlett's tests in factor analysis showed that: KMO coefficient was quite high (= 0.878 > 0.5) with significance level Sig. = 0.000 (Sig. < 0.05) showing that EFA factor analysis was very suitable. At Eigenvalues ​​levels greater than 1 and with the method of extracting Principal Components with Varimax rotation, factor analysis extracted 4 factors from 18 observed variables and with a total variance extracted of 74.759% (> 50%) meeting the requirements.

Table 2.6 shows that the Tangibles and Trust scales have been combined into one factor because these two components did not achieve discriminant value. Thus, the 5 service quality components according to the theoretical model have become 4 components when evaluating the quality of electricity bill collection service as follows:

- Factor 1: includes variables PT1, PT2, PT3, TC1, TC2 and TC3 Tangibles and Reliability (PTTC).

- Factor 2: includes variables DU1, DU2, DU3, DU4, DU5 and DU6 Response

(DU).


- Factor 3: includes variables NL1, NL2 and NL3 Service capacity (NL).

- Factor 4: includes variables CT1, CT2 and CT3 Sympathy (CT).

Table 2.6: Results of factor analysis of service quality scale

KMO and Bartlett's Test


KMO and Bartlett's Test

Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

Bartlett's Test of Sphericity Approx. Chi-Square

df Sig.

.878 1.472E3

153

.000


Total Variance Explained



Component

Initial Eigenvalues

Extraction Sums of Squared

Loadings

Rotation Sums of Squared

Loadings

Total

% of

Variance

cumulative

%

Total

% of

Variance

cumulative

%

Total

% of

Variance

cumulative

%

1

8.102

45,009

45,009

8.102

45,009

45,009

4,493

24,962

24,962

2

2,643

14,683

59,692

2,643

14,683

59,692

4,384

24,358

49,320

3

1,597

8,870

68,562

1,597

8,870

68,562

2,485

13,807

63,127

4

1,115

6,196

74,759

1,115

6,196

74,759

2,094

11,632

74,759

5

.698

3,878

78,637







6

.649

3.604

82,241







7

.485

2,692

84,933







8

.428

2,379

87,312







9

.374

2,078

89,390







10

.315

1,747

91,137







11

.280

1,557

92,693







12

.259

1,438

94,131







13

.243

1,350

95,482







14

.221

1,227

96,709







15

.175

.973

97,681







16

.169

.942

98,623







17

.145

.804

99,427







18

.103

.573

100,000







Extraction Method: Principal Component Analysis.

Total Variance Explained Rotated Component Matrix


Component

1

2

3

4

TC3 TC2 TC1 PT3 PT1 PT2 DU5 DU1 DU6 DU4 DU3 DU2 NL3 NL1 NL2 CT3 CT1

CT2

.912

.869

.861

.790

.760

.727


.


.835

.812

.805

.798

.789

.756


.880

.855

.778


.846

.659

.658

(Source: Extracted from data processing results using SPSS software)

2.2.3.2. Analysis of factors affecting customer satisfaction

Table 2.7: Results of Cronbach's Alpha coefficient analysis of satisfaction variables



Item-Total

Scale Mean if Item Deleted

Scale Variance if Item Deleted

Corrected Item- Total

Correlation

Cronbach's Alpha if Item

Deleted

Alpha = .802

HL1

7.5804

2,066

.677

.702

HL2

7.6071

1,916

.730

.643

HL3

7.8125

2,064

.550

.839

(Source: Extracted from data processing results using SPSS software)

The customer satisfaction scale consists of 3 observed variables HL1, HL2 and HL3. These variables will be put into factor analysis to check the level of convergence. The KMO test achieved a value of 0.666, Eigenvalues ​​>1 and the total variance used to explain the factor > 50% (72.265%) meeting the conditions of factor analysis.

Thus, the results of factor analysis of customer satisfaction show that all 3 observed variables have loading factors > 0.5 and are reasonable to use to explain the customer satisfaction scale.

Table 2.8: Results of factor analysis of customer satisfaction scale


Variable encoding

Factor

1

HL1

.899

HL2

.872

HL3

.774

Eigenvalues

2,168

Cumulative %

72.265%

Cronbach's Alpha

.802

(Source: Extracted from data processing results using SPSS software)

2.2.4. Building regression models and testing hypotheses

2.2.4.1. Regression model

Based on the analysis in the above section, the general regression equation is constructed as follows:

HL = β 0 + β 1 *PTTC + β 2 *DU + β 3 *NL + β 4 *CT

In which: HL Customer satisfaction.

PTTC Tangibles and Reliability. DU Responsive.

NL Service capacity. CT Sympathy.

2.2.4.2. Assessment of model fit

A statistical coefficient called Pearson Correlation Coefficient is used to quantify the strength of the linear relationship between two quantitative variables. If there is a strong correlation between two variables, then the problem of multicollinearity must be considered when performing regression analysis.

Table 2.9: Correlation between variables in the regression model of the scale

Correlations



PTTC

DU

NL

CT

HL

PTTC

Pearson Correlation

1

.457**

.337**

.446**

.573**

Sig. (2-tailed)


.000

.000

.000

.000

N

110

110

110

110

110


DU

Pearson Correlation

.457**

1

.494**

.568**

.577**

Sig. (2-tailed)

.000


.000

.000

.000

N

110

110

110

110

110


NL

Pearson Correlation

.337**

.494**

1

.485**

.505**

Sig. (2-tailed)

.000

.000


.000

.000

N

110

110

110

110

110


CT

Pearson Correlation

.446**

.568**

.485**

1

.652**

Sig. (2-tailed)

.000

.000

.000


.000

N

110

110

110

110

110



HL

Pearson Correlation

.573**

.577**

.505**

.652**

1

Sig. (2-tailed)

.000

.000

.000

.000


N

110

110

110

110

110


(Source: Extracted from data processing results using SPSS software)

In Table 2.12, we see a close correlation between the dependent variable HL (satisfaction) and the independent variables in the PTTC model (Tangibles and Trust), DU (Responsiveness) and CT (Empathy). The correlation coefficient between the Satisfaction variable and the other variables is greater than 0.3.

2.2.4.3. Regression analysis

Multiple regression analysis is not just a description of observed data. From the results in the sample, the causal relationship between the dependent variable (HL) and the independent variables will be determined. The regression analysis model will describe the form of the relationship and thereby help predict the level of the dependent variable when the value of the independent variable is known. To build the regression model, select the Enter method with the following analysis results:

- R square = 0.570 with a very small observed significance level (Sig. = 0) shows that the multiple linear regression model fits the data set well and can be used. However, the model often does not fit the actual data as the R 2 value shows. In this situation, the adjusted R 2 (0.554) from R 2 is used to more closely reflect the goodness of fit of the multiple linear regression model because it does not depend on the exaggerated bias of R 2 . Thus, the adjusted R 2 coefficient of 0.554 shows that the compatibility of the model with the observed variable is very large with about 55.4% of the variation of the dependent variable HL can be explained by the 4 independent variables in the model.

Table 2.10: Regression analysis results

Model Summary b


Model

R

R Square

Adjusted R

Square

Std. Error of

the Estimate

1

.755a

.570

.554

.45240

a. Predictors: (Constant), CT, PTTC, NL, DU

b. Variable: HL

ANOVA b


Model

Sum of

Squares

Df

Mean

Square

F

Sig.


1

Regression Residual

Total

28,989

21,900

50,889

4

106

110

7,247

.205

35,410

.000a

a. Predictors: (Constant), CT, PTTC, NL, DU

b. Dependent Variable: HL

Coefficients a



Model

Unstandardized

Coefficients

Standardized

Coefficients


T


Sig.

Collinearity

Statistics

B

Std.

Error

Beta

Tolerance

VIF

1

(Constant)

.243

.306

.286


.169


.154

.795

.428

.735



PTTC

.263

.068

3,870

.000

.578

1,360


DU

.172

.085

2,032

.045

.690

1,729


NL

.144

.071

2.017

.046

.590

1,448


CT

.367

.086

.353

4,277

.000

.735

1,694

a. Dependent Variable: HL

Residuals Statistics a



Minimum

Maximum

Mean

Std.

Deviation

N

Predicted Value

2.5080

4.9729

3.8333

.51104

110

Residual

-1.13978

1.10836

.00000

.44418

110

Std. Predicted

Value

-2.593

2,230

.000

1,000

110

Std. Residual

-2.519

2,450

.000

.982

110

(Source: Extracted from data processing results using SPSS software)

Comment


Agree Privacy Policy *