Testing the Reliability of the Scale Using Cronbach'S Alpha Coefficient

Bank image

Vietinbank constantly improves service quality	HA.1
Vietinbank has a good brand compared to other banks.	HA.2
Vietinbank has a reputation in providing services. online banking	HA.3
Decision to use	Will use online banking when eligible	SUDUNG.1
	Vietinbank will be the first choice bank when registering to use/utilize banking services. online shop	SUDUNG.2
	Will register to use Vietinbank's online banking service in the near future.	SUDUNG.3
	Will recommend Vietinbank online banking service to friends and relatives use together	SUDUNG.4

Maybe you are interested!

2.3.3 Analysis of research results

2.3.3.1 Research sample information

The survey was conducted during the period from July to October 2013 with 250 questionnaires distributed, the survey results collected 239 questionnaires. After eliminating unsatisfactory responses and cleaning the data, the remaining research sample was 230 observations.

Table 2.2: Characteristics of the study sample

Allocation by sample

Quantity	Percentage in sample
Sex	Male	89	38.7
Sex	Female	141	61.3
Age	Under 22 years old	33	14.3
	22 to 35 years old	109	47.4
	36 to 45 years old	57	24.8
	Over 45 years old	31	13.5

Level

Below high school	9	3.9
From high school to college/university	92	40
Undergraduate/Graduate	129	56.1
Job	Office staff	106	46.1
	Manager/Director	26	11.3
	Merchant	56	24.3
	Housewife	25	10.9
	Other	17	7.4
Income	Under 6 million VND	20	8.7
	From 6 to under 12 million VND	162	70.4
	From 12 to under 25 million VND	36	15.7
	From 25 million VND or more	12	5.2
	Total	230	100

(Source: Appendix C – Survey sample characteristics)

Regarding the characteristics of the survey sample, among 230 customers who answered the questionnaire, the ratio between men and women was different, accordingly, the number of women surveyed was higher with 141 customers (accounting for 61.3%) and there were 89 male respondents (accounting for 38.7%). Regarding age, the majority of respondents were between 22 and 35 years old (109 people, accounting for 47.4%), 57 customers were between 36 and 45 years old (accounting for 24.8%), the age group under 22 and over 45 had a low ratio of 14.3% and 13.5% respectively, especially since they had access to modern banking services such as Internet banking. Statistics by occupation, out of 230 surveys collected, the main occupations are: office workers 106 customers (accounting for 46.1% of the sample size), followed by business people (accounting for 24.3% of the sample size), the lowest percentage is the survey respondents belonging to other groups with 17 customers. The level of surveyed customers

The majority of respondents were university and postgraduate students (56.1% of the sample size), while the common income level of surveyed customers was 6 to under 12 million VND (accounting for 70.4%).

2.3.3.2 Testing the reliability of the scale using Cronbach's Alpha coefficient

Before conducting factor analysis to extract the influencing factors of the model, the study will test the reliability of the scale through the Cronbach's Alpha coefficient on the SPSS software program, as well as test the correlation between observed variables. According to many researchers, a scale with a Cronbach Alpha coefficient of 0.8 or higher is a good measurement scale. However, in cases where the concept under study is a new concept or new to the respondents in the research context, a coefficient of 0.6 or higher is still acceptable (Hoang Trong and Chu Nguyen Mong Ngoc, 2008). In addition, variables with an item-total correlation coefficient of less than 0.3 will also be eliminated from the scale.

Table 2.3: Cronbach's Alpha test results of the scales

Observation variable

Average scale if variable type	Scale variance if variable type	Total variable correlation	Cronbach's Alpha if variable is excluded
Security (ATBM) - Cronbach Alpha = 0.77
ATBM.1	13.66	6,959	0.558	0.722
ATBM.2	13.13	7,803	0.484	0.746
ATBM.3	13.73	6,798	0.598	0.707
ATBM.4	13.11	7,909	0.503	0.741
ATBM.5	13.59	7.203	0.566	0.719

Bank Image (HA) - Cronbach Alpha = 0.728

HA.1	6.7	2,169	0.546	0.647
HA.2	6.77	2,333	0.576	0.611
HA.3	6.83	2,345	0.529	0.664
Pricing Policy (GIA) - Cronbach Alpha = 0.798
GIA.1	10.23	4.021	0.675	0.714
GIA.2	10.16	4,770	0.509	0.793
GIA.3	10.11	4,310	0.606	0.750
GIA.4	10.26	3.905	0.656	0.724
Expected Performance (HQMD) - Cronbach Alpha = 0.866
HQMD.1	16.00	12,432	0.662	0.844
HQMD.2	16.10	11,825	0.755	0.827
HQMD.3	15.96	12,374	0.679	0.841
HQMD.4	16.13	12,554	0.590	0.857
HQMD.5	16.01	12,236	0.667	0.843
HQMD.6	16.13	12,388	0.624	0.850
Ease of Use (DSD) -Cronbach Alpha = 0.808
DSD.1	9.90	5,493	0.561	0.789
DSD.2	10.11	4,581	0.741	0.701
DSD.3	9.96	5,182	0.607	0.768
DSD.4	10.04	4,789	0.601	0.774

Decision to use (QDSD)- Cronbach Alpha = 0.764

SUDUNG.1	10.0783	3,941	0.444	0.777
SUDUNG.2	10.1913	3,623	0.624	0.674
SUDUNG.3	10.3174	3,947	0.573	0.704
SUDUNG.4	10,213	3,784	0.632	0.673

(Source: Appendix C - Testing scale reliability using Cronbach's Alpha)

Preliminary testing shows that the component variables measuring security, bank image, pricing policy, expected efficiency, ease of use and decision to use have Cronbach's Alpha coefficients of 0.77; 0.728; 0.798; 0.866; 0.808; 0.764 respectively, all greater than 0.7, and the total correlation coefficient of the observed variables is greater than 0.3. Therefore, it can be concluded that the theoretical scales ensure reliability.

2.3.3.3 Exploratory Factor Analysis (EFA)

All observed variables were subjected to exploratory factor analysis (EFA). Some criteria to be considered in exploratory factor analysis (EFA) are as follows:

- KMO coefficient (Kaiser-Mayer-Olkin) is greater than or equal to 0.5

- The significance level of Bartlett's test is less than or equal to 0.05

- Factor loading coefficient is greater than 0.4, if any observed variable has a factor loading coefficient less than 0.4 it will be eliminated.

- The scale is accepted when the total extracted variance is greater than or equal to 50%.

- Eigenvalue coefficient is greater than 1.

Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

Approx. Chi-Square

Bartlett's Test of

.913

2132.841

231

.000

Table 2.4: KMO and Bartlett test of independent variables

Sphericity

Sig.

(Source: Appendix C – Results of exploratory factor analysis EFA)

Table 2.5: EFA analysis results for influencing factor scales

Component

Matrix a

Component
1	2	3	4	5
HQMD.1	.625
HQMD.2	.766
HQMD.3	.653
HQMD.4	.654
HQMD.5	.741
HQMD.6	.683
DSD.1		.695
DSD.2		.797
DSD.3		.604
DSD.4		.609
GIA.1			.738
GIA.2			.592
GIA.3			.697
GIA.4			.677
ATBM.1				.757
ATBM.2				.622
ATBM.3				.562
ATBM.4				.570
ATBM.5				.525
HA.1					.806
HA.2					.815
HA.3					.786

(Source: Appendix C – Results of exploratory factor analysis EFA)

EFA analysis results for independent variables: According to tables 2.4 and 2.5, the five original factors are still kept. The KMO coefficient = 0.913, so EFA is suitable for the data and the Chi-quare statistic of Bartlett's test reaches the value of 2132.841 with a significance level of 0.000, so the observed variables are correlated with each other in the overall scope. The extracted variance is 61.947%, showing that the extracted factors explain 61.947% of the variation in the data. The specific factor components after running EFA are as follows:

- Expected performance (HQMD): measured by 6 observed variables (From HQMD.1 to HQMD.6)

- Ease of use (DSD): measured by 4 observation variables (From DSD.1 to DSD.4)

- Pricing policy (GIA): measured by 4 observed variables (From GIA.1 to GIA.4)

- Safety and security (ATBM): measured by 5 observed variables (From ATBM.1 to ATBM.5)

- Bank image (HA): Measured by 3 observed variables (From HA.1 to HA.3)

The results of EFA analysis for the dependent variable are shown in Table 2.6 and

2.7. The KMO coefficient reached 0.731, the Chi-Square statistic of Bartlett's test reached 251.866 with a significance level of 0.000, the extracted variance was 59.406%. All factor loading coefficients met the requirements.

Kaiser-Meyer-Olkin Measure of Sampling Adequacy.

Approx. Chi-Square

Bartlett's Test of

.731

251,866

.000

Table 2.6: KMO and Bartlett test of dependent variable

Sphericity

Sig.

Table 2.7: EFA analysis results for the usage decision scale

Component Matrix a

Component
1
SUDUNG.1	.647
SUDUNG.2	.809
SUDUNG.3	.784
SUDUNG.4	.830

(Source: Appendix C - Results of exploratory factor analysis EFA)

2.3.3.4 Testing the influence of personal factors

Test for the difference between two population means (Independent sample T-test)

Hypothesis H 0 : There is a difference between men and women in the decision to use Vietinbank's online banking services.

Table 2.8: Independent sample T-test results

QDSD

Levene's Test for Equality

of Variances

T-test for Equality of Means

Sig.

Sig. (2-

tailed)

Mean Difference

Std. Error Difference

95% Confidence

Interval of the Difference

Lower

Upper

Equal variances assumed

Equal variances not assumed

.045

.832

-.394

-.392

228

182,824

.694

.696

-.05663

.14358

.14461

-.33953

-.34195

.22628

.22870

(Source: Appendix C – T-test)

The Sig value in Levene test is 0.832>0.05, which shows that the variance between the two genders is not different, so we consider the next t-test result in the "Equal variances assumed" section. The Sig value in t-test is 0.694>0.05, which shows that there is no difference between men and women in the decision to use Vietinbank's online banking services.

One way ANOVA test

- Hypothesis H 01 : There is no difference in the decision to use online banking services of individual customers at Vietinbank according to age.

Table 2.9: Results of variance test for age groups

QDSD

Levene Statistic