2 The reliability test: the Cronbach’s alpha test
Tải bản đầy đủ - 0trang
Table 4.2: The results of Cronbach’ alpha test
Scale Mean if Scale Variance if Corrected
Item Deleted
Item Deleted
Item-Total
Correlation
Squared
Multiple
Correlation
Cronbach's
Alpha if Item
Deleted
Rewards
REWARD01
REWARD02
Cronbach's Alpha
7.4641
3.365
.721
0.823
.532
.723
7.7416
2.779
.710
.527
.729
REWARD03
7.3397
3.370
.620
.386
.812
Policy
POLICY04
POLICY05
POLICY06
Cronbach's Alpha
8.2871
2.744
8.3684
2.676
8.1866
3.028
.757
.710
.706
0.852
.573
.509
.506
.760
.808
.810
Information
IMFORM07
INFORM08
INFORM09
Cronbach's Alpha
6.4498
3.681
5.8517
4.310
6.6077
3.836
.644
.462
.614
0.744
.451
.215
.429
.571
.782
.609
Personalizatio
n
PERSON10
PERSON11
PERSON12
Cronbach's Alpha
7.2057
2.520
7.3206
2.277
7.1483
2.762
.598
.661
.516
0.759
.390
.446
.274
.668
.592
.758
Staff
STAFF13
STAFF14
STAFF15
STAFF16
Cronbach's Alpha
9.9139
10.1962
10.0526
10.2440
.777
.804
.832
.759
0.908
.646
.650
.705
.603
.887
.878
.868
.894
Tangibility
TANGI17
TANGI18
Cronbach's Alpha
3.4593
1.230
3.2249
1.435
.769
.769
0.868
.591
.591
Communicatio
n
COMMU19
COMMU20
Cronbach's Alpha
2.8373
1.185
3.2105
1.109
.663
.663
0.797
.440
.440
Reputation
REPU21
REPU22
Cronbach's Alpha
7.7990
2.979
7.7321
2.986
.711
.733
0.829
.538
.558
6.291
6.245
6.175
6.349
33
.741
.722
REPU23
7.9234
Customer
satisfaction
SATIS24
SATIS25
SATIS26
Cronbach's Alpha
7.7177
2.598
7.8660
2.578
7.8421
2.499
Customer
Loyalty
LOYAL27
LOYAL28
LOYAL29
Cronbach's Alpha
7.0191
4.028
7.0383
3.729
7.5981
3.155
2.869
.628
.395
.831
.774
.758
.764
0.879
.600
.576
.585
.821
.834
.830
.634
.649
.516
0.754
.465
.480
.266
.635
.604
.796
The result performed that 10 scales had the result of Cronbach’s alpha above
0.7, the highest was 0.908 (Staff in the Customer Loyalty Program) and the lowest
was 0.744 (Information in the Customer Loyalty Program). Moreover, the corrected
item-total correlation of each item is above 0.3. This indicates that all scales fit the
requirement for reliability. As a result, these measures were used in establishing the
main survey to test the study hypotheses.
Moreover, in the scale of Information, Cronbach's alpha would increase from
0.744 to 0.782 if item INFORM08 were deleted or not used for computing an overall
task value score. As the same situation, in the scale of Reputation, Cronbach's alpha
would increase from 0.829 to 0.831 if item REPU23 were deleted or not used for
computing an overall task value score. For the scale of Loyalty, Cronbach's alpha
would increase from 0.754 to 0.796 if item LOYAl29 were deleted or not used for
computing an overall task value score. The author decided that these items should be
removed from the scale. Note first that alpha increased by a large degree from deleting
these items. Second, these items did not correlates very well with the composite score
from others in the scales (the item-total correlation for item INFORM08 is 0.462,
REPU23 is 0.628 and LOYAL29 is 0.516).
34
4.3 Exploratory factor analysis (EFA)
After analyzing the Cronbach’s alpha, the author evaluated the measurement
scales by conducting exploratory factor analysis. The purpose of EFA is to define
which set of items go together as a group or are answered similarly by respondents
(Leech et al., 2005). In this study, EFA was run through the Principal Axis Factoring
with Varimax rotation method. As the conceptual model that there are ten factors:
Rewards, Policy, Information, Personalization, Tangibility, Communication, Staff,
Reputation, Customer Satisfaction and Customer Loyalty. The author examined if the
items belonging to one concept actually are in the same group. Based on the test of
assumption, the KMO was 0.893 presenting sufficient items for each factor. KMO test
indicates one whether or not enough items are predicted by each factor. The Bartlett
was significant (0.000 less than 5%) means that the variable are correlated highly
enough to provide a reasonable basis for factor analysis.
Table 4.3: KMO and Bartlett's Test for all variables
KMO and Bartlett's Test
Kaiser-Meyer-Olkin
Measure
Sampling Adequacy.
Bartlett's Test Approx. Chi-Square
of Sphericity
df
Sig.
of .893
3184.526
253
0.000
By doing EFA (Principal Axis Factoring with Varimax rolation method), the
result showed that four factors were extracted from 10 items measuring: perceived
risk, social cost, consumers’ attitude toward purchasing counterfeit products and
purchase intention toward counterfeit products. Moreover, the cumulative of the first
six factors occupied for 74.3 percent of variance. This indicated that major percent of
variance could be explained by six initial items.
35
Table 4.4 : Total Variance Explained for all variables
Total Variance Explained
Factor
Initial Eigenvalues
Total
Extraction Sums of Squared Rotation
Loadings
Sums of
Squared
Loadingsa
Total
%
of Cumulative Total
Variance %
8.619
39.179
39.179
6.415
2.137
9.713
48.893
6.952
1.345
6.113
55.006
6.031
.867
3.940
58.947
2.841
.745
3.387
62.333
3.649
.669
3.041
65.375
3.969
%
of Cumulative
Variance %
1
8.959
40.723
40.723
2
2.472
11.238
51.961
3
1.625
7.388
59.349
4
1.218
5.538
64.887
5
1.082
4.920
69.807
6
1.010
4.589
74.396
Extraction Method: Principal Axis Factoring.
a. When factors are correlated, sums of squared loadings cannot be added to obtain a total
variance.
Additional, the author has eliminated the items of REPU21 and REPU22 due to
the reason that the distance from the max value and the next value are under 0.25. The
Rotated Factor Matrix showed the items and factor loading for rotated factors with
loading higher than 0.5 are significant as requirement. The items clustered into six
groups that they belong to.
There are several changes in factors. This means the perception of interviewees
about some variables is different from the hypotheses of the study and some previous
theory. However, this change reflects the thought and perception of the respondents
about the factors actually influence their satisfaction as well as loyalty in shopping at
the supermarket. After EFA, Customer Satisfaction and Customer Loyalty turn into
the same group that can be name as Customer Outcome. For the same situation, Staff
and Communication are in the same group named Serving, which is same for Rewards
and Policy as belowed:
Customer Outcome (Customer Satisfaction, Customer Loyalty)
Serving (Staff, Communication)
36
Regulation (Rewards, Policy)
These replacement variables including one dependent variable and five
independent variables were entered into the regression equation at the same time. And
the results are presented in the next section
Table 4.5: Rotated Component Matrix for all variables
Customer
Outcome
Serving
Regulation
Personalization
Information
Tangibility
SATIS24
SATIS25
SATIS26
LOYAL27
LOYAL28
STAFF13
STAFF14
STAFF15
STAFF16
COMMU19
COMMU20
REWARD01
REWARD02
REWARD03
POLICY04
POLICY05
POLICY06
PERSON11
PERSON10
IMFORM07
INFORM09
TANGI17
TANGI18
Component
1
2
.783
.787
.785
.709
.712
.794
.821
.824
.798
.644
.706
3
4
5
6
.664
.710
.753
.663
.681
.535
.727
.829
.814
.828
.839
.852
4.4 Multiple regression analysis
After testing the Cronbach’s Alpha Analysis and EFA, the author conducted the
multiple regression analysis in order to define the relationship between six factors
mentioned above. According to Hair et al. (2010), multiple regression analysis helps
the author to predict the level of impact of independent variable on dependent
37
variable. The relationship between the independent variables (such as Serving,
Regulation, Personalization, Information, Tangibility) and Customer Outcome (which
in combined by Customer Satisfaction and Customer Loyalty) was tested by standard
multiple regression analysis. It is necessary for testing correlation between variables
by using Pearson correlations. The results present that the explanatory variables are
not correlated with each other.
Table 4.6: Correlations or all variables
Correlations
Customer
Outcome
Regula
tion
Tangibil
ity
Pearson
1
Correlation
Pearson
.515**
1
Correlation
Tangibility
Pearson
.377**
.170*
1
Correlation
Information Pearson
.503**
.428**
.483**
Correlation
Personalizati Pearson
.522**
.517**
.309**
on
Correlation
Serving
Pearson
.607**
.785**
.272**
Correlation
**. Correlation is significant at the 0.01 level (2-tailed).
Informat Personaliza
ion
tion
Serving
Customer
Outcome
Regulation
1
.424**
1
.446**
.609**
1
The result of running the Multiple Regression was reported to determine how well the
model fit:
Table 4.7: Model Summaryb
Model
R
R Square
Adjusted
Square
1
.768a
.590
.579
R Std. Error of the
Estimate
Durbin-Watson
.50770
1.479
a. Dependent Variable: Customer Outcome
b. Predictors: (Constant), Regulation, Serving, Personalization,
Tangibility, Information
38
According to the Model Summary table, the multiple correlation coefficient (R)
was 0.768, R Square was equal 0.590 and adjusted R Square was 0.579, showing that
57% of the variance in consumers’ outcome could be predicted from five independent
variables.
Table 4.8: ANOVA alpha
ANOVAa
Model
1 Regression
Sum
Squares
75.146
of df
Mean
Square
15.029
5
F
Sig.
58.307
.000b
Residual
52.325
203
.258
Total
127.471
208
a. Dependent Variable: Customer Outcome
b. Predictors: (Constant), Regulation, Serving, Personalization, Tangibility, Information
The value of F was 58.307 and sig<0.05 indicates that the combination of these
variables significantly predicts the dependent variable.
Table 4.9: Coefficients alpha
Coefficientsa
Model
Unstandardized
Coefficients
B
Std. Error
1 (Constant)
.328 .214
Regulation
.196 .055
Serving
.559 .057
Tangibility
.045 .038
Information
.046 .046
Personalization
.110 .050
a. Dependent Variable: Customer Outcome
Standardized
Coefficients
Beta
.200
.537
.063
.054
.117
t
1.534
3.545
9.766
1.207
.984
2.226
Sig.
.127
.000
.000
.229
.326
.027
Collinearity Statistics
Tolerance
VIF
.635
.670
.743
.661
.729
1.574
1.492
1.346
1.513
1.372
Since there new group of scales that have been formed, the author have to revised
conceptual model before testing hypothesis base on the multiple regression analysis.
4.5 Revised research model
39
After checking the result of the reliability of each variable and validity of the
measurement scales testing, there is change of independent factors. The change here
can be understood that specific group of customer shopping at various supermarket in
Ho Chi Minh City did not clearly distinct some factors that were group after EFA
testing. Firstly, the concept of Satisfaction and Loyalty in others sectors are extremely
clear distinct. But most of the respondents suppose that these two factors are the
same. In the terms of loyalty have the satisfaction and vice versa. It can be explained
the same for Rewards and Policy, they assumed that these factors are the regulation of
the Customer Loyalty Program. And for the Staff and Communication factors, the
respondent assumes that they are just the Serving that they gain from the Customer
Loyalty Program. Therefore, the revised and finalized research model is describe in
the following figure.
Figure 4.1: The revised and finalized research model
40
Table 4.10: Finalized hypotheses of the research
Hi
H1(a,b,c,d,e)
H1a
H1b
H1c
H1d
H1e
H2
Hypotheses
There is a positive impact of customer loyalty program service quality on the
customer outcome
There is a positive impact of regulation on the customer outcome
There is a positive impact of serving on the customer outcome
There is a positive impact of tangibility on the customer outcome
There is a positive impact of personalization on the customer outcome
There is a positive impact of information on the customer outcome
Higher the store reputation, higher will be the customer loyalty.
H1a: There is a positive impact of regulation on the customer outcome
As presented in above table, sig value of perceived risk was under 0.05. It
indicated that the factor of regulation have the positive impact on the customer
outcome. Moreover, the regulation has the highest beta coefficient with the value of
0.537 which reflects that it has the greatest contribution to customer outcome.
H1b: There is a positive impact of serving on the customer outcome
As presented in above table, sig value of perceived risk was under 0.05. It
indicated that the factor of serving have the positive impact on the customer outcome.
Moreover, the regulation has the second high beta coefficient with the value of 0.200
which reflects that it also has the great contribution to customer outcome.
H1c: There is a positive impact of tangibility on the customer outcome
With the sig value 0.326 (>0.05), the author could conclude that the factor of
tangibility did not have the relationship with the customer outcome on this research.
H1d: There is a positive impact of personalization on the customer outcome
As presented in above table, sig value of perceived risk was 0.27 (under 0.05).
It indicated that the factor of personalization have the positive impact on the customer
outcome.
H1e: There is a positive impact of information on the customer outcome
41
With the sig value 0.229 (>0.05), the author could conclude that the factor of
information did not have the relationship with the customer outcome on this research.
H2: Higher the store reputation, higher will be the customer loyalty.
Since the elements of factor reputation have been eliminated all throughout the
process of running Cronbach’s Alpha and EFA testing, so the author assumes that
store reputation did not have the relationship with the customer outcome on this
research.
Table 4.11: Summary of hypotheses testing results
No.
H1a
H1b
H1c
H1d
H1e
H2
Hypotheses
There is a positive impact of regulation on the customer outcome
There is a positive impact of serving on the customer outcome
There is a positive impact of tangibility on the customer outcome
There is a positive impact of personalization on the customer
outcome
There is a positive impact of information on the customer
outcome
Higher the store reputation, higher will be the customer loyalty.
Testing result
Supported
Supported
Not supported
Supported
Not supported
Not supported
4.6 Explanation for the finding results of the hypotheses
According to the consumer behavior literature and recent empirical research,
we defined seven variables (Rewards, Policy, Staff, Communication, Information,
Tangibility, Personalization) that impact on customer satisfaction, and then Customer
Satisfaction as well as Store Reputation that influence Customer Loyalty in the
supermarket retail sector. However, our study findings suggest that the anticipated
negative impact of tangibility and information on customer outcome, store reputation
on customer outcome and the most important findings that customer satisfaction and
customer loyalty are the same in the research.
This meant that consumers never pay attention to the store reputation to decide
their selection where to buy, and they do not care much about the differentiation
between satisfaction and loyalty for a retailer. These findings are different with the
42
results from previous empirical. In the best of the effort, the author could not find the
empirical studies in other countries as well as the limited number of researches about
retail loyalty in Vietnam to be an evidence for this finding but the possible explanation
for this outcome might lie in the different context of conducting the research. Most of
the previous studies were studied at different countries which have a different culture,
standard of living, standard of moral, life style, legal…this could lead to different
concepts of consumers in retail loyalty.
Regarding to this finding, in order to explain and confirm this outcome in
practice, the author also conducted the in-depth interviews with 5 respondents about
the reasons why (1) customer satisfaction and customer loyalty are inclusive, (2) staff
and communication are the same, (3) rewards and policy are inclusive and (4) store
reputation does not affect the customer loyalty.
For the first matter of Customer Satisfaction and Customer Loyalty, the point
of customer satisfaction is that people feel comfortable, pleasure and happy when
shopping at the supermarket. From that pleasure, they will also choose that
supermarket again when have the needs to buy goods. The comfort in shopping
includes the tendency to buy again, that mean the satisfaction and the loyalty in this
in-depth interview are inclusive or can be said that they are the same.
For the second matter of Staff and Communication, all of five people in the indepth interview assume that these two factors are the place they get the information or
raise the suggestion about the Customer Loyalty Program Quality. All the promotions
within the program or the process of information feedback, the customer can easily get
through by the Staff and the Website (Communication) of the specific supermarket.
For the third matter of Rewards and Policy, all interviewees declared that they are
fully aware of these two factor but they decide that Rewards or Policy are all in
regulation of Customer Loyalty Program. For example, the gift that they will receive
or how they can get bonus in a period of time are exactly the regulation. For the last
matter of store reputation does not affect the customer loyalty, all interviewees
43