Professional Documents
Culture Documents
au
la
na
A
za
d
Li
br
ar
y,
A
lig
ar
h
Chapter – 5
M
us
lim
Analysis and Interpretation
U
ni
ve
rs
i ty
CHAPTER–5: ANALYSIS AND INTERPRETATION
CHAPTER – 5
ty
characteristic and nature of the overall customer relation management at the two
i
rs
colossal public and private sector banks of India, selected for this study, viz.
ve
Allahabad Bank and HDFC Bank belonging to public and private sector respectively,
has been carried out at length in the previous chapter. It has been found that the
ni
U
customer relation management effectiveness and practices obtaining at two entities
lim
meet the standard and ensemble the nature and requirement of their customer to
provide world class banking facilities with minimum service charge. The impact,
us
however, of these customer relation management policies and practices remains to be
examined with the application of statistical tools.
M
h
The present chapter is consequently devoted to finding out the impact of customer
ar
relation management effectiveness and practices in force at the Allahabad Bank and
lig
loyalty and cross buying in both the selected banks representing India’s banking
y,
industry for the present study. To this end, data is collected through administration of
ar
well-designed questionnaire from the customers belonging to both public and private
br
sector banks. Also, the data is analyzed with the application of statistical tools and
Li
interpreted to derive logical results which are used to test the hypotheses of this study.
d
za
5.1 Universe
A
The population or universe represents the entire group of units which is the focus of
na
the study. Thus, the population could consist of all the persons in the country, or those
la
depending on the purpose and coverage of the study. The population could also
M
consist of non-human units such as farms, houses or business establishment. Bless and
Higson-Smith(2006) view the universe as “the entire set of objects or group of people,
which is the object of research and about which the researcher wants to determine
some characteristics”. On the other hand, Neuman (2000) refers to a universe as “the
broad class of units that are covered in hypotheses.” Therefore, it can be concluded
143
CHAPTER–5: ANALYSIS AND INTERPRETATION
that universe refers to all the units to which the findings of a specific study might get
generalized directly or indirectly.
The universe for the survey of this study is finite. All the customers of Allahabad
Bank and HDFC Bank in India are considered as the universe. It includes all the
customers of both the banks who reside in different cities of Delhi NCR, and form
ty
part of this study. However, due to time & other resource constraints collection of
i
rs
complete data from all the customers from the universe is not possible. Hence, a select
ve
sample of geographical areas as well as of customers is chosen from the universe for
ni
data collection for this study.
U
5.2 Sample
lim
The sample represents the subset of the population. A sample is a finite part of a
us
statistical population whose properties are studied to gain information about the
M
whole. When dealing with people it can be defined as a set of respondents selected
h
from a large population for the purpose of a survey. For this research, sample has to
ar
be taken since it is not possible to cover such a vast sector as Banking Industry on
lig
account of impediments of time, finance and other resources required for the purpose.
A
Hence, the geographical area for the surveys as well as the customers of the bank
y,
(respondents) to the questionnaire have been selected in such a manner that the whole
ar
country can be represented and the conduct of research work gets handy and possible
br
As far as geographical coverage for this study is concerned, the researcher tried to
d
za
draw the sample in such a way that it may represent the whole population judiciously.
A
Four cities namely Ghaziabad, Noida, Delhi, and Gurugram, have been chosen as
na
sample for geographical coverage. All select cities have a huge presence of customers
of these two banks. Respondents have been drawn from the four cities to a
la
manageable handling limit of 564 customers spread over the cities. This convenience
au
sampling method has been used for the study. While the sample selection of four
M
cities for geographical coverage is based on proximity and convenience, the selection
of respondents from the four cities is at random.
144
CHAPTER–5: ANALYSIS AND INTERPRETATION
5.3 Questionnaire
The current survey has been conducted with the help of a well-designed prepared
questionnaire, which is placed as Appendix-1 to this study. The questionnaire was
equipped taking into deliberation the requirement of this study to assess the impact of
customer relation management effectiveness and practices on organizational
ty
commitment as well as customer’s satisfaction, Loyalty, and Cross buying in both the
i
rs
banks namely Allahabad Bank and HDFC bank and making a comparative study of
ve
both the select banks. The design, pattern, and structure of the questionnaire have
ni
been simple, study-specific, unambiguous, minimal, and capable of invoking the
U
objective type of answers through marking a tick only on the statement which
lim
customers consider fit as per their agreement and disagreement. In designing the
us
question items, the researcher adapted and modified the items of questionnaires from
previous researchers to suit the need of the present study. The Questionnaire first
M
seeks to elicit perception of customers towards select customer relation management
h
ar
dimensions and their practices. Each question is rated on a five-point Likert scale. It is
lig
commitment, satisfaction, loyalty and cross buying. Further, the questionnaire seeks
y,
ar
the respondents has been obtained which include age, gender, educational
qualification, income level, profession type, (professional or non professional) work
d
za
experience, marital status etc, so as to get the overall background of the customers for
A
drawing the fruitful conclusion. This questionnaire contains one hundred twelve
na
statements which measure the twenty independent variables and five dependent
variables selected for the present study.
la
au
The research instrument used is Likert’s 5-point scale. Responses to items were
scored in such a way that a response indicative of most favorable was given the
highest scores and vice versa (strongly agree = 5 and strongly disagree = 1).The
whole questionnaire is divided into five parts, the first part deals with demographics,
second part concerns with organizational commitment, third segment starts with
145
CHAPTER–5: ANALYSIS AND INTERPRETATION
customer satisfaction, fourth part deals with customer loyalty and finally the fifth and
last part is related to cross buying. In order to obtain information on various
parameters relating to customer relation management effectiveness and practices in
the two selected banks, the questionnaire poses issue-wise statements (queries) for
answers from the respondents. Thus, the questionnaire poses to the respondent queries
ty
for assessing the customer perception. with regard to organizational commitment (Six
i
rs
statements),customer experience (Six statements) process driven(five statements),
ve
Reliability (Five items),Technological Orientation(Five statements), Customized
ni
Offer (Six statements), Perceived quality(five items),Competence (5 statements),
U
Trust (5 statements), Communication (5 statements), Strong Relationship(5
lim
statements), Conflict Handling(5 statements), Stronger Relational Bond(5
statements),Customer Influence(5 statements), Promoting Customer Dependency(5
us
statements), Interaction Frequency( 5 statements), Customer Commitment(5 items),
M
Customer Expanding(5 items), Product Patronage( 5 statements), Extremely Loyal
h
Customers(5 statements), Additional Product Purchase( 4 items) . Items were drawn
ar
from the model questionnaire propounded by Singh, K. (2004), Rao and Abraham
lig
(1991). However, some changes in the items of the questionnaire were made to suit
A
the present study. The reliability and validity of the scale in the Indian context has
y,
been tested by authors and it is found to be satisfactory. The study follows the survey
ar
research technique. Each respondent is required to tick the option that represents the
br
perception most truly and faithfully. The questionnaire, at the outset, mentions clear-
Li
cut instructions to the respondents for filling up the questionnaire properly. This
d
5.5 Response
la
Customers of both, the Allahabad Bank and HDFC Bank gave an encouraging and
au
useful response. About 1500 questionnaires were floated among customers of both the
M
banks at different locations, in order to cover the sample size of 1128. Out of these
300 did not respond while 72 respondents incompletely filled the questionnaire and,
hence, these were rejected. Thus, the received and rejected responses amounted to 372
out of 1500 leaving 1128 purposeful responses with adequate information as per the
required sample of 1128 for this study. This yielded an aggregate response rate of
146
CHAPTER–5: ANALYSIS AND INTERPRETATION
75.2 %, which is said to be satisfactory in a study based on primary data. Data were
analyzed on the basis of responses provided by 1128 respondents completely in all
respects. Therefore, the exact sample was 1128. It has been observed that during the
course of data collection, the respondents took very cooperative and sincere interest
and care in appropriately giving their perceptions on the scale of items provided in the
ty
questionnaire and returned the questionnaire duly and completely filled. Distribution
i
rs
of questionnaire is explained in the following Table 5.1
ve
TABLE 5.1
ni
Questionnaire Distribution and Response Rate
U
lim
S. Name of Distributed Table Usable, Response Methodology
No. District/ Questionnaire Returned and Rate (%) adopted in
us
State completed Distributing
questionnaires Questionnaires
147
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
presents the preliminary analysis of data. Under preliminary analysis demographic
i
rs
profile of the respondents, common method variance, reliability analysis of research
ve
instrument, exploratory factor analysis, and confirmatory factor analysis have been
ni
discussed. Whereas the Part two deals with the objective-wise analysis of data which
U
examines the impact of customer relation management effectiveness and practices on
lim
organizational commitment, customer satisfaction, customer loyalty, and cross buying
us
among the customers of selected banks.
The adapted questionnaire (Singh, 2004), for the purpose of this research, was utilized
lig
with sample size (n=100). As in adapted questionnaire, the items are changed as per
A
the demographic and cultural basis, so it becomes mandatory to explore the new items
y,
within the existing constructs (Malik, 2016). And that is why the exploratory factor
ar
analysis was used to see the nature and structure of the research instrument. Factor
br
analysis is used to reduce a large set of measured variables in terms of relatively few
Li
among the variables this technique allows the researcher to group the variables into
A
factors and the factors are known as new variables or latent variables (Kothari, 2012).
na
terms of their common underlying dimensions (factors). Table 5.2 below shows the
au
148
CHAPTER–5: ANALYSIS AND INTERPRETATION
The exploratory factor analysis was conducted by using KMO to find the Sampling
Adequacy of the data, it indicates the variables of the dependent and independent
nature were placed to the respondents and they have responded well. Further the total
variance explained by all the variables was calculated in order to find out the
ty
threshold limit which is indicated by Hair and Black 2006.
i
rs
TABLE 5.2
ve
KMO and Bartlett's Test
Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy. .781
ni
Approx. Chi-Square 1962.929
U
Bartlett's Test of Sphericity Df 351
Sig. .000
lim
us
TABLE 5.3
Communalities
OC_1
Initial
1.000
M Extraction
.484
h
OC_2 1.000 .670
ar
149
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
TOTAL VARIANCE EXPLAINED
i
rs
Component Initial Eigen values Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings
ve
Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative %
1 6.410 23.741 23.741 6.410 23.741 23.741 4.217 15.617 15.617
ni
2 5.137 19.027 42.768 5.137 19.027 42.768 4.084 15.127 30.744
U
3 3.239 11.997 54.765 3.239 11.997 54.765 3.690 13.666 44.410
4 2.457 9.099 63.865 2.457 9.099 63.865 3.516 13.022 57.432
lim
5 1.467 5.432 69.297 1.467 5.432 69.297 3.203 11.864 69.297
6 1.167 4.322 73.618
us
7 .856 3.171 76.790
8 .748 2.770 79.560
M
9 .702 2.600 82.159
10 .614 2.275 84.434
h
ar
11 .561 2.077 86.511
12 .472 1.749 88.260
lig
13 .436 1.616 89.876
14 .373 1.383 91.259
A
15 .358 1.325 92.583
y,
16 .319 1.183 93.766
ar
17 .291 1.078 94.844
18 .225 .833 95.676
br
19 .181 .670 96.346
Li
150
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
threshold limit which is indicated by Hair and Black 2006.
i
rs
ve
TABLE 5.4
ni
KMO and Bartlett's Test
U
Kaiser-Meyer-Olkin (KMO) Measure of Sampling Adequacy. .773
lim
Approx. Chi-Square 627.306
Bartlett's Test of Sphericity
Df 171
us
Sig. .000
M
h
TABLE 5.5
ar
Communalities
lig
Initial Extraction
A
151
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
TABLE 5.6
i
rs
Total Variance Explained
ve
Component Initial Eigen values Extraction Sums of Squared Loadings Rotation Sums of Squared Loadings
ni
Total % of Variance Cumulative % Total % of Variance Cumulative % Total % of Variance Cumulative %
U
1 5.497 28.930 28.930 5.497 28.930 28.930 2.764 14.549 14.549
lim
2 1.727 9.088 38.018 1.727 9.088 38.018 2.752 14.482 29.031
3 1.645 8.658 46.676 1.645 8.658 46.676 2.424 12.758 41.789
us
4 1.480 7.787 54.463 1.480 7.787 54.463 1.966 10.349 52.138
M
5 1.177 6.195 60.658 1.177 6.195 60.658 1.619 8.520 60.658
6 .953 5.015 65.673
h
ar
7 .828 4.360 70.033
lig
8 .771 4.057 74.090
9 .762 4.008 78.098
A
10 .655 3.450 81.548
y,
11 .587 3.091 84.639
12 .503 2.649 87.288
ar
br
13 .459 2.417 89.705
Li
152
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.7
Exploratory Factor Analysis
ty
EXPLORATORY FACTOR ANALYSIS HDFC bank Allahabad bank
i
rs
KMO Measure of Sampling Adequacy .781 .773
ve
p-Value 0.000 0.000
ni
Total Variance Explained 69.29 60.65
U
lim
The Kaiser-Meyer-Olkin measure of sampling adequacy and Bartlett's test of
us
sphericity:
M
The Kaiser-Meyer-Olkin measure of sampling adequacy is an index used to examine
h
the appropriateness of factor analysis. High values indicate that factor analysis is
ar
appropriate (Malhotra and Dash, 2010). The KMO statistic varies between 0 and 1. A
lig
value of 0 indicates that the sum of partial correlations is large relative to the sum of
A
are relatively compact and so factor analysis should yield distinct and reliable factors.
br
Kaiser (1974) recommends accepting values greater than 0.5 as acceptable (values
Li
below this should lead to either collect more data or rethink which variables to
d
za
include). Furthermore, values between 0.5 and 0.7 are mediocre, values between 0.7
and 0.8 are good, values between 0.8 and 0.9 are great and values above 0.9 are
A
in the case of HDFC Bank is 0.781 and in the case of Allahabad Bank, it is .773
la
which falls into the range of being acceptable and good respectively: there is
au
confidence that that factor analysis is appropriate for the present study.
M
Bartlett’s measure tests the null hypothesis that the original correlation matrix is an
identity matrix. For factor analysis to work there is a need for some relationships
between variables and if the R-matrix were an identity matrix then all correlation
coefficients would be zero. Therefore, this test to be significant should have a
153
CHAPTER–5: ANALYSIS AND INTERPRETATION
significance value less than 0.05. A significant test tells us that the R-matrix is not an
identity matrix; therefore, there are some relationships between the variables included
in the analysis. For these data, Bartlett's test is highly significant (p < 0.001), and
therefore factor analysis is appropriate. Therefore the data is found suitable for further
analysis.
ty
5.8 Demographic Profile
i
rs
It is essential to know the demographic profile of the particular respondent. The
ve
profile of the customers includes their gender, marital status, educational qualification
ni
occupation, household income, age, association of bank.
U
5.8.1 Gender of the Respondents
lim
TABLE 5.8
us
Distribution of the customers based on Gender
Number of Customers M
h
Total Number
ar
% % %
y,
The present study analyses the gender among the customers in the two groups (male
na
and female) of public (Allahabad) and private (HDFC) banks. When a profile of
respondents was generated based on gender, it can be seen from the table given above
la
au
that the male respondents’ outnumbered female employees as far as respondents for
this study are concerned. In the present study, there are ‘764’males and ‘364’females.
M
Table 5.8 above also reveals that out of total 1128 respondents, 764 customers are
male while female employees account for 364. In the case of Allahabad Bank out of
total 564 customers, 400 are male while female employees stand at 164 giving a
response rate of 71 % and 29% respectively. However, in the case of HDFC Bank out
154
CHAPTER–5: ANALYSIS AND INTERPRETATION
of total 564 customers, 364 customers are male while females constitute 200
customers yielding a response rate of 64.5% and 35.5% respectively.
ty
Allahabad bank of the public sector and HDFC bank of the private sector.
i
rs
ve
ni
U
lim
us
M
h
ar
lig
Age of the customers constitutes one of the important profile variables of the
br
effectiveness dimensions regarding their customers. Customers between the age group
d
of 15-25 are most prone to raiding and poaching by other banks. Hence, it becomes
za
extremely important for the banks to give due attention to the specified age group of
A
of the banks. The age group of the customers in the present study includes customers
la
of the age group from15-25, 25-35, 35-45, 45-55, and 55-65 and above 65.
au
M
155
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.9
Number of customers
Age
(in Allahabad bank HDFC bank Total
ty
years) Percentage
Number
Number Percentage Number Percentage
i
rs
15-25 50 8.8 80 14.2 130 11.5
ve
ni
25-35 182 33 200 35 382 33.9
U
35-45 232 41 104 18.8 336 29.8
lim
45-55 50 8.8 80 14.3 130 11.5
us
55-65 30 5 70 12.4 100 8.8
From Table 5.9 above, it is evident that in the HDFC Bank the age group of 25-35
y,
years constitutes the highest number of customers being 35% of the total respondents
ar
drawn from the bank. In the case of Allahabad Bank, however, the highest number of
br
banks customer respondents fall in the age group of 35-45 years constituting 41% of
Li
the total respondents drawn from the bank. However, when total respondents merged
d
together of both the banks, the statistics indicate that the largest number of customers
za
pertains to the age group of 25-35 years constituting 33.9% of the total sample of
A
The bar diagrams below shows the break-up in percentage terms of each category of
la
age groups of the respondents in both the Allahabad Bank and the HDFC Bank under
au
study.
M
156
CHAPTER–5: ANALYSIS AND INTERPRETATION
232
200
182
130 130
104
ty
100
80 80 70
50 50 50
i
30 20 30
rs
ve
15-25 25-35 35-45 45-55 55-65 >65
ni
U
Allahabad bank HDFC bank Total Number
lim
Fig. 5.3: Age Wise Distribution of Customers of Allahabad Bank and HDFC Bank
us
M
The level of education among customers is one of the significant criteria which need
to be given utmost attention while formulating customer relation management
h
ar
whether a particular customer wants to be in their career path. Hence the level of
A
education of the customers is included as one of the profile variables. The highly
y,
educated customers may be more prone to attrition while lower educated are
ar
complacent with what s/he is getting. The education level of the customer in the
br
TABLE 5.10
za
Number of customers
na
Total
Education Allahabad bank HDFC bank
la
Under
180 16 80 7.1 260 23
M
Graduate
Post
164 14 184 16.4 348 31
Graduate
Professionals 220 20 300 26.5 520 46
Total 564 50 564 50 1128 100.00
157
CHAPTER–5: ANALYSIS AND INTERPRETATION
Table 5.10 above reveals that professionals form the highest percentages of customers
in both the banks followed by post graduate and undergraduate which constitute 20 %
in case of Allahabad bank and 26.5% in case of HDFC bank.
The bar diagram below shows the Public Sample size, percentage and educational
Fig. 4. Distribution of the customers on the basis of Level
level of the respondents. of Education
ty
i
rs
520
ve
348
300
ni
260
220
180 164 184
U
80
lim
us
Allahabad bank HDFC bank Total
The marital status of the customer respondents drawn from both the Allahabad Bank
ar
and the HDFC Bank has been computed in two broad heads, viz, ‘Married’ and
br
TABLE 5.11
d
Number of Customer
na
Total
Marital
Allahabad bank HDFC bank
la
Status
au
158
CHAPTER–5: ANALYSIS AND INTERPRETATION
Table 5.11 above exhibit the distribution of the respondent’s marital status in case of
Allahabad Bank as well as HDFC Bank. Statistic in table clearly shows that 53%
respondent belong to the category of married class while rest are unmarried which
constitute 47% of the total respondents. This outcome is also highlighted with the help
of bar diagram below.
i ty
rs
ve
53
47
ni
35
U
32
lim
18
15
us
M
h
Allahabad bank HDFC bank Total
ar
Married Unmarried
lig
A
Fig. 5.5: Distribution of the customer on the basis of marital status (In Percentage)
y,
The occupation status of the customer respondents has also been elicited. It has been
br
TABLE 5.12
d
za
Number of Customers
na
Occupation Total
Status Allahabad bank HDFC bank
la
159
CHAPTER–5: ANALYSIS AND INTERPRETATION
Table 5.12 above depicts the occupation status of respondents in both the banks.
Employment status of the respondents has been divided into four group viz, govt jobs,
private jobs, businessman, students. In the case of Allahabad bank 31% customers are
in Government job while in the case of HDFC Bank they are 6%. The 7% customers
of Allahabad bank belong to private sector jobs. On the other hand customer of HDFC
ty
bank doing private jobs is 13%. The business class favoured HDFC bank for their
i
rs
banking services which are clearly indicated from the table which constitute 27%,
ve
while they are 3% in the case of public sector Allahabad banks. Participation of
ni
student community in both the banks is not very much noticeable.
U
The bar diagrams below shows the occupation- wise status of total respondents in the
lim
two Public vs. Private sector banks in terms of sample size, percentage and
us
occupation of the respondent respectively.
414
M
h
ar
334
lig
A
230
y,
Total Number
ar
37
d
30
20 13
za
160
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
TABLE 5.13
i
rs
Distribution of the Respondents Income
ve
Number of Customers
ni
Household Total
Monthly Allahabad bank HDFC bank
U
Income
lim
Number Percentage Number Percentage Number Percentage
Rs 15000-
us
50 4 34 3 84 7
25000
Rs 25000-
132 12 150
M 13 282 25
h
35000
ar
Rs 35000-
lig
Rs 45000-
100 9 90 8 190 17
y,
55000
ar
Rs 55000-
br
80 7 50 4.5 130 12
65000
Li
Rs >65000 20 2 50 4.5 70 6
d
za
The table 5.13 above shows the different categories of respondents income-wise. The
la
161
CHAPTER–5: ANALYSIS AND INTERPRETATION
33
25
16 17 17
ty
12 13 12
i
9 8
rs
7 7
4 3 4.5 4.5 6
2
ve
ni
15000-25000 25000-35000 35000-45000 45000-55000 55000-65000 >65000
U
Allahabad bank HDFC bank Total
lim
us
Fig. 5.7: Distribution of customers on the basis of Monthly Income (In Percentage)
required for researcher to analyse the common method variance in order to see the
lig
pattern of biases that may exist in the study, as in every primary study there may be
A
some chances of biases of the raters or respondents to the items (independent and
y,
variance can create systematic extent inaccuracy, which further may lead to both
br
Types I and Type II errors”(Podsakoffet., al, 2003). “The magnitude of biases present
Li
in any study, in regard to this Harman’s one-factor test (Harman, 1960) is a technique
d
uses exploratory factor analysis”, where all variables are loaded onto a single factor
za
and fixed invariance of 1 from the total number of variables i.e. there is no rotation
A
(Podsakoffet al, 2003) (as shown in table 5.14 and table 5.15) in case of HDFC bank
na
and Allahabad bank respectively. “This process through which new factor comes is
la
not the researchers model but it is required to be done only for the purpose of analysis
au
to check or see as if the total variables which are fixed into 1 variance (no rotation) is
M
explaining the variance more than 50%, then common method bias may be present”
162
CHAPTER–5: ANALYSIS AND INTERPRETATION
Harman’s Common method bias was tested and it was found that the total variance
explained is on 9.420 %, which means biases exist but within the acceptable range.
Table 5.14
Total Variance Explained
ty
Component Initial Eigen values Extraction Sums of Squared Loadings
i
rs
Total % of Cumulative Total % of Cumulative
Variance % Variance %
ve
1 10.550 9.420 9.420 10.550 9.420 9.420
2 8.837 7.891 17.310
ni
3 6.784 6.057 23.368
4 6.533 5.833 29.201
U
5 5.946 5.309 34.510
6 5.190 4.634 39.144
lim
7 4.586 4.094 43.238
8 4.163 3.717 46.955
us
9 3.688 3.293 50.248
10 3.579 3.195 53.443
11
12
3.200
3.026
2.857
2.702
56.300
59.002 M
h
13 2.912 2.600 61.603
ar
163
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
54 .316 .282 93.731
55 .295 .263 93.995
i
rs
56 .280 .250 94.245
57 .276 .246 94.491
ve
58 .264 .236 94.726
59 .259 .231 94.958
ni
60 .239 .213 95.171
61 .235 .210 95.381
U
62 .232 .207 95.588
lim
63 .222 .198 95.786
64 .213 .190 95.977
65 .207 .184 96.161
us
66 .197 .176 96.337
67 .194 .173 96.510
68
69
.192
.186
.172
.166
96.681
96.848 M
h
70 .176 .157 97.005
ar
164
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
110 .022 .020 99.969
111 .018 .016 99.986
i
rs
112 .016 .014 100.000
ve
Extraction Method: Principal Component Analysis.
ni
a) Harman common method Bias calculated by using principal component extraction method and
fixing all the independent and dependent variables into one.
U
b) The total variance explained by all the independent and dependent variable should be less than 50
%. Results showing 9.420 %, which is within the acceptable range.
lim
5.9.2 Harman’s Common Method (Allahabad bank.)
us
Haman’s Common method bias was tested and it was found that the totalvariance
M
explained is only 12.797%, which means that biases exist but within the acceptable
h
range.
ar
Table 5.15
lig
165
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
29 1.171 1.045 74.676
i
30 1.114 .995 75.671
rs
31 1.074 .959 76.630
ve
32 1.048 .936 77.567
33 1.013 .904 78.471
ni
34 .982 .877 79.348
U
35 .908 .810 80.158
36 .871 .778 80.936
lim
37 .864 .771 81.708
38 .835 .745 82.453
us
39 .804 .718 83.170
40
41
.793
.732
.708
.653
83.878
84.532 M
h
42 .727 .649 85.180
ar
166
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
76 .196 .175 97.477
i
77 .191 .171 97.648
rs
78 .172 .153 97.801
ve
79 .165 .148 97.949
80 .159 .142 98.091
ni
81 .155 .139 98.229
U
82 .151 .135 98.364
83 .144 .129 98.493
lim
84 .135 .120 98.613
85 .130 .116 98.729
us
86 .123 .110 98.839
87
88
.121
.112
.108
.100
98.946
99.047 M
h
89 .110 .098 99.145
ar
167
CHAPTER–5: ANALYSIS AND INTERPRETATION
d) The total variance explained by the entire independent and dependent variable should be less
than 50 %. This study sho12.797%, which is within the acceptable range.
5.9.3 Comparison of Harman’s Common Method Bias in HDFC Bank and
Allahabad Bank
The comparative analysis indicated the common method variance of 9.420 %
and 12.797 % in case of HDFC Bank and Allahabad Bank respectively. Taking the
ty
inferences from the analysis shown in Table 5.16, it can be stated that the biases are
i
rs
found to be more in the instrument of Allahabad bank than HDFC bank, but both are
ve
under the prescribed limit, required for the present study.
ni
TABLE 5.16
U
Comparison of Common Method Variance (CMV)
lim
HDFC Bank CMV (in %) Allahabad Bank CMV (in %)
(N=564) (N=564)
us
Total number of
Total number of
Independent and 9.420
Dependent M
Independent and 12.97
h
Dependent variables
variables
ar
lig
Reliability test is inevitable for the study based on a questionnaire or based on survey.
y,
Reliability refers to the extent to which a scale produces consistent results if repeated
ar
instrument to produce the same result consistently (Tavakol & Dennick, 2011).The
Li
measure the strength of the scale before examining the relationship between the
za
computation of coefficient alpha” (Cronbach, 1951). The value of Cronbach alpha 0.9
na
is excellent, 0.8 is good, 0.7 is acceptable, 0.6 is questionable, 0.5 is poor and less
la
coefficient alpha should be above 0.7 (Nunnally, 1978). The reliability of items in
M
168
CHAPTER–5: ANALYSIS AND INTERPRETATION
1, it means the internal consistency of the item in the scale is greater. The purpose of
reliability analysis is to find out the internal consistency of scale. The coefficient
values of the individual item were under the acceptable range (0.906 – 0.909). The
accumulated Cronbach’s alpha of HDFC bank for the 112 items was found to be
0.909 as shown in Table -5.17.
ty
5.10.1 Reliability Analysis of Research Instrument of HDFC bank
i
rs
TABLE 5.17
ve
Reliability Statistics
ni
Reliability Analysis HDFC Bank
Total No. of Items Combined Cronbach's Alpha Value
U
112 0.909
lim
Observed Variables Cronbach's Alpha
us
OC_1 .908
OC_2 .908
OC_3
OC_4
M .908
.907
h
ar
OC_5 .907
OC_6 .908
lig
CE_1 .907
CE_2 .908
A
CE_3 .908
y,
CE_4 .908
ar
CE_5 .908
CE_6 .908
br
PDA_1 .908
Li
PDA_2 .907
PDA_3 .908
d
PDA_4 .909
za
PDA_5 .909
A
REL_1 .908
REL_2 .909
na
REL_3 .908
REL_4 .908
la
REL_5 .909
au
TOR_1 .908
TOR_2 .908
M
TOR_3 .908
TOR_4 .909
TOR_5 .908
CO_1 .908
CO_2 .909
CO_3 .908
CO_4 .909
169
CHAPTER–5: ANALYSIS AND INTERPRETATION
CO_5 .908
PQ_1 .909
PQ_2 .909
PQ_3 .908
PQ_4 .907
PQ_5 .908
PQ_6 .908
ty
COM_1 .908
i
COM_2 .908
rs
COM_3 .908
ve
COM_4 .908
COM_5 .908
ni
TRU_1 .908
U
TRU_2 .908
TRU_3 .908
lim
TRU_4 .907
TRU_5 .908
us
COMM_1 .908
COMM_2
COMM_3 M .908
.908
h
COMM_4 .908
ar
COMM_5 .908
lig
SR_1 .908
SR_2 .908
A
SR_3 .908
SR_4 .908
y,
SR_5 .908
ar
CH_1 .908
br
CH_2 .908
CH_3 .908
Li
CH_4 .908
d
CH_5 .908
za
SRB_1 .908
SRB_2 .908
A
SRB_3 .908
na
SRB_4 .908
SRB_5 .908
la
CI_1 .908
au
CI_2 .908
CI_3 .908
M
CI_4 .908
CI_5 .907
PCD_1 .907
PCD_2 .907
PCD_3 .908
PCD_4 .908
PCD_5 .908
170
CHAPTER–5: ANALYSIS AND INTERPRETATION
IF_1 .907
IF_2 .907
IF_3 .907
IF_4 .908
IF_5 .908
CC_1 .908
ty
CC_2 .908
CC_3 .908
i
rs
CC_4 .908
ve
CC_5 .908
CEX_1 .908
ni
CEX_2 .908
U
CEX_3 .908
lim
CEX_4 .909
CEX_5 .908
us
PP_1 .909
PP_2 .908
PP_3 M .908
h
PP_4 .908
ar
PP_5 .907
lig
ELC_1 .908
ELC_2 .908
A
ELC_3 .908
y,
ELC_4 .908
ar
ELC_5 .908
APP_1 .908
br
APP_2 .908
Li
APP_3 .909
d
APP_4 .908
za
APP_5 .908
AB_1 .908
A
AB_2 .908
na
AB_3 .909
AB_4 .909
la
au
TABLE 5.18
M
171
CHAPTER–5: ANALYSIS AND INTERPRETATION
OC_2 .907
OC_3 .907
OC_4 .905
OC_5 .906
OC_6 .906
CE_1 .905
CE_2 .906
ty
CE_3 .905
i
CE_4 .905
rs
CE_5 .905
ve
CE_6 .905
PDA_1 .905
ni
PDA_2 .905
U
PDA_3 .905
PDA_4 .905
lim
PDA_5 .905
REL_1 .905
us
REL_2 .905
REL_3
REL_4 M .905
.906
h
REL_5 .905
ar
TOR_1 .905
lig
TOR_2 .906
TOR_3 .905
A
TOR_4 .905
TOR_5 .905
y,
CO_1 .904
ar
CO_2 .905
br
CO_3 .906
CO_4 .905
Li
CO_5 .904
d
PQ_1 .905
za
PQ_2 .905
PQ_3 .905
A
PQ_4 .905
na
PQ_5 .905
PQ_6 .906
la
COM_1 .906
au
COM_2 .905
COM_3 .907
M
COM_4 .905
COM_5 .905
TRU_1 .906
TRU_2 .905
TRU_3 .905
TRU_4 .905
TRU_5 .905
172
CHAPTER–5: ANALYSIS AND INTERPRETATION
COMM_1 .905
COMM_2 .904
COMM_3 .905
COMM_4 .905
COMM_5 .905
SR_1 .905
SR_2 .905
ty
SR_3 .905
i
SR_4 .905
rs
SR_5 .905
ve
CH_1 .906
CH_2 .905
ni
CH_3 .906
U
CH_4 .905
CH_5 .904
lim
SRB_1 .905
SRB_2 .905
us
SRB_3 .905
SRB_4
SRB_5 M .905
.905
h
CI_1 .905
ar
CI_2 .905
lig
CI_3 .904
CI_4 .905
A
CI_5 .904
PCD_1 .904
y,
PCD_2 .905
ar
PCD_3 .905
br
PCD_4 .905
Li
PCD_5 .905
IF_1 .906
d
IF_2 .906
za
IF_3 .905
A
IF_4 .906
na
IF_5 .906
CC_1 .907
la
CC_2 .907
au
CC_3 .905
CC_4 .904
M
CC_5 .905
CEX_1 .906
CEX_2 .906
CEX_3 .906
CEX_4 .907
CEX_5 .907
173
CHAPTER–5: ANALYSIS AND INTERPRETATION
PP_1 .918
PP_2 .907
PP_3 .907
PP_4 .907
PP_5 .906
ELC_1 .906
ty
ELC_2 .906
ELC_3 .906
i
rs
ELC_4 .906
ve
ELC_5 .906
APP_1 .907
ni
APP_2 .907
U
APP_3 .906
lim
APP_4 .906
APP_5 .907
us
AB_1 .907
AB_2 .908
AB_3 M .907
h
AB_4 .907
ar
lig
Table 5.19
ar
In the above table 5.19 reliability analysis through Cronbach’s alpha is required to be
za
done in order to measure the strength of the scale before examining the relationship
A
Cronbach alpha 0.9 is excellent, 0.8 is good, 0.7 is acceptable, 0.6 is questionable, 0.5
au
is poor and less than0.5 is unacceptable (George and Mallery, 2003). A measure to be
M
acceptable of coefficient alpha should be above 0.7 (Nunnally, 1978). The reliability
of items in different variables was assessed by computing the coefficient alpha
(Cronbach, 1951).The instrument is subjected to the computation of coefficient alpha”
(Cronbach, 1951). It is mandatory that the researcher should estimate the Cronbach
174
CHAPTER–5: ANALYSIS AND INTERPRETATION
alpha in order to make accuracy in the interpretation of the data. Cronbach alpha
reliability coefficient range between 0 to 1.
ty
it represents the model, as up to what extent the measurable variables represent the
i
constructs, and when these outcome resultants are combined with constructs validity
rs
test then it signifies the better understanding of the quality of measures” (Hair et. al,
ve
2006)
ni
TABLE 5.20
U
Confirmatory Factor Analysis of HDFC Bank Ltd
lim
Fit Indices Parsimonious Improved Model
us
Indicator Model Indices
CMIN/DF
1
3.43
2
3.20 M 3
3.10
4
2.52
h
2 311.98 310.16 289.71 140.6
ar
Moderate Fit
br
Model Decision
Model
Li
The measurement model, which is also called confirmatory factor analysis was run on
za
the latent variables and observed variables of the research questionnaire. The purpose
A
for using the CFA was to identify the validity of the questionnaire items in order to
na
a) CFA CRME
M
To validate the measurement model the fit of model, reliability and validity check was
performed. The back to back model fit indices was improved with the series of
modification indices and some items were deleted on the ground of low factor
loadings. The Parsimonious run results of CFA indicated poor fit model with values
(CMIN/DF= 3.43, χ ² =311.98, DF= 144, GFI= 0.708, NFI = 0.775, CFI = 0.721,
175
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
GFI= 0.784, NFI = 0.789, CFI = 0.784, RMSEA = 0.101, p =0.000), in another
i
rs
subsequent the result indicated moderate fit model with values (CMIN/DF= 2.52, χ ²
ve
= 140.6, DF= 138, GFI= 0.804, NFI = 0.800, CFI = 0.815, RMSEA = 0.082,
ni
p=0.000).
U
TABLE 5.21
lim
Confirmatory Factor Analysis Goodness of fit of CRME Latent and Observed
us
Variables –HDFC Bank
Fit Indices
Indicator
Parsimonious
Model Indices
M
Improved Model
h
ar
1 2 3 4
CMIN/DF 3.43 3.20 3.10 2.52
lig
Moderate Fit
Model Decision
Model
d
za
TABLE 5.22
na
1. OC_1 DELETED
2. OC_2 0.72
3. OC_3 0.76
4. OC-4 0.97
176
CHAPTER–5: ANALYSIS AND INTERPRETATION
5. OC_5 0.79
6. OC_6 DELETED
7. CE_1 DELETED
8. CE_2 0.67
9. CE_3 0.86
ty
10. CE_4 0.88
i
rs
11. CE_5 0.67
ve
12. CE_6 DELETED
ni
13. PDA_1 0.72
U
14. PDA_2 0.89
lim
15. PDA_3 0.78
us
16. PDA_4 0.82
17. PDA_5
MDELETED
h
18. REL_1 0.95
ar
Reliability**
M
177
CHAPTER–5: ANALYSIS AND INTERPRETATION
constructs, because “high Discriminant validity provides that the construct is unique”.
In the current research study, the Discriminant validity was assessed by using the
pattern method of (Fornell and Larcker, 1981), which states that, “Discriminant
validity exists when Squared Inter Construct Correlation (SIC) is less than average
variance extracted (AVE)”. For this purpose, the entire three constructs correlation
ty
matrix was drawn through the Spearman’s correlation method. Therefore, these
i
rs
correlation values were squared and put above the diagonal lines and then compared
ve
with the AVE, which indicated that all the Squared Inter Construct Correlations (SIC)
ni
are less than AVE as shown in the following table. Hence, the constructs are truly
U
distinct from others.
lim
Table 5.23
us
Showing Discriminant Validity of CRME
Constructs AVE M
Squared Inter Construct Correlation (SIC)
h
OC CE PDA REL TOR
ar
178
CHAPTER–5: ANALYSIS AND INTERPRETATION
b) CFA CS
To validate the measurement model the fit of model, reliability and validity check was
performed. The back to back model fit indices was improved with the series of
modification indices and some items were deleted on the ground of low factor
loadings. The Parsimonious run results of CFA indicated poor fit model with values
ty
(CMIN/DF= 3.53, χ ² =351.98, DF= 244, GFI= 0.714, NFI = 0.785, CFI = 0.741,
i
rs
RMSEA = 0.115, p =0.000). Subsequently, a series of 3 improved run of CFA was
ve
carried on and the results of statistics for fit indices shown improvement in every
ni
steps and final resultant came to moderate fit as- (CMIN/DF= 3.20, χ ² = 340.16,
U
DF= 240, GFI= 0.715, NFI = 0.787, CFI = 0.780, RMSEA = 0.101, p =0.000) then
lim
additional run was done and values were (CMIN/DF= 3.15, χ ² = 280.72, DF= 239,
us
GFI= 0.794, NFI = 0.799, CFI = 0.794, RMSEA = 0.100, p =0.000), in another
M
subsequent the result indicated moderate fit model with values (CMIN/DF= 2.75, χ ²
h
= 175.4, DF= 225, GFI= 0.801, NFI = 0.801, CFI = 0.821, RMSEA = 0.090, p
ar
=0.000).
lig
TABLE 5.24
A
1 2 3 4
za
179
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
No. Items
CO PQ COM TRU COMM SR CH
i
rs
1 OC_1 0.68
ve
2 OC_2 0.99
ni
3 OC_3 0.71
U
4 OC-4 -
lim
5 OC_5 -
6 PQ_1 0.56
us
7 PQ_2 0.64
8 PQ_3 0.81 M
h
9 PQ_4 0.85
ar
10 PQ_5 -
lig
11 PQ_6 -
A
12 COM_1 0.87
y,
13 COM_2 0.76
ar
14 COM_3 0.58
br
15 COM_4 -
Li
16 COM_5 -
d
17 TRU_1 0.74
za
18 TRU_2 0.95
A
19 TRU_3 0.79
na
20 TRU_4 0.59
la
21 TRU_5 -
au
22 COMM_1 0.59
M
23 COMM_2 0.60
24 COMM_3 0.99
25 COMM_4 0.79
26 COMM_5 -
27 SR_1 0.50
180
CHAPTER–5: ANALYSIS AND INTERPRETATION
28 SR_2 0.73
29 SR_3 0.95
30 SR_4 -
31 SR_5 -
32 CH_1 0.98
ty
33 CH_2 0.82
i
rs
34 CH_3 0.60
ve
35 CH_4 -
ni
36 CH_5 -
U
Average 64.6 52.2 55.6 60.5 57.5 56.6 66.3
Variance
lim
Extracted (in %)
us
Construct 6.74 0.79 0.74 0.79 0.79 0.73 0.74
Reliability**
M
* AVE = Individual loading square sum / Sum of number of items* 100
h
** CR = (Sum of Items loading)2/(Sum of Items loading)2 + Std. error variance
ar
lig
TABLE 5.26
y,
181
CHAPTER–5: ANALYSIS AND INTERPRETATION
ity
rs
ve
ni
Fig. 5.9: Confirmatory Factor Analysis of Customer Satisfaction
U
lim
c) CFA CL
us
To validate the measurement model the fit of model, reliability and validity check was
M
performed. The back to back model fit indices was improved with the series of
h
modification indices and some items were deleted on the ground of low factor
ar
loadings. The Parsimonious run results of CFA indicated poor fit model with values
lig
(CMIN/DF= 3.43, χ ² =1322.8, DF= 144, GFI= 0.814, NFI = 0.797, CFI = 0.781,
A
carried on and the results of statistics for fit indices shown improvement in every
ar
steps and final resultant came to moderate fit as- (CMIN/DF= 3.23, χ ² = 1318.8,
br
DF= 140, GFI= 0.815, NFI = 0.799, CFI = 0.789, RMSEA = 0.100, p =0.000) then
Li
additional run was done and values were (CMIN/DF= 3.11, χ ² = 1316.8, DF= 139,
d
za
GFI= 0.824, NFI = 0.810, CFI = 0.824, RMSEA = 0.09, p =0.000), in another
A
subsequent the result indicated moderate fit model with values (CMIN/DF= 2.79, χ ²
na
= 1312.8, DF= 134, GFI= 0.830, NFI = 0.812, CFI = 0.827, RMSEA = 0.08, p
la
=0.000).
au
M
182
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.27
Confirmatory Factor Analysis Goodness of Fit of Customer Loyalty Latent and
Observed Variables – HDFC Bank
ty
Indicators Model Indices
i
rs
1 2 3 4
ve
CMIN/DF 3.43 3.23 3.11 2.79
2 1322.8 1318.8 1316.8 1312.8
ni
U
DF 144 140 139 134
lim
GFI 0.814 0.815 0.824 0.830
NFI 0.797 0.799 0.810 0.812
us
CFI 0.781 0.789 0.824 0.827
RMSEA 0.101 0.100
M 0.09 0.08
h
p-value 0.000 0.000 0.000 0.000
ar
Moderate Fit
Model Decision
lig
Model
A
y,
Constructs
br
TABLE 5.28
Li
1 SRB_1 0.67
na
2 SRB_2 0.76
la
3 SRB_3 0.90
au
4 SRB_4 0.76
M
5 SRB_5 -
6 CL_1 0.60
7 CL_2 0.72
8 CL_3 0.83
9 CL_4 0.77
10 CL_5 -
183
CHAPTER–5: ANALYSIS AND INTERPRETATION
11 PCD_1 -
12 PCD_2 -
13 PCD_3 0.78
14 PCD_4 0.86
15 PCD_5 0.78
ty
16 IF_1 0.56
i
17 IF_2 0.63
rs
18 IF_3 0.84
ve
19 IF_4 0.89
ni
20 IF_5 0.71
U
21 CC_1 0.95
lim
22 CC_2 0.81
23 CC_3 0.53
us
24 CC_4 -
25 CC_5
M -
h
Average Variance 60.5 53.7 65.5 58.8 61.6
ar
Extracted (in %)
lig
TABLE 5.29
Li
SRB CI PCD IF CC
A
na
184
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
ni
U
lim
us
M
Fig. 5.10: Confirmatory Factor Analysis of Customer Loyalty
h
ar
d) CFA CB
lig
To validate the measurement model the fit of model, reliability and validity check was
A
performed. The back to back model fit indices was improved with the series of
y,
modification indices and some items were deleted on the ground of low factor
ar
loadings. The Parsimonious run results of CFA indicated poor fit model with values
br
(CMIN/DF= 3.23, χ ² =749.4, DF= 90, GFI= 0.814, NFI = 0.797, CFI = 0.777,
Li
carried on and the results of statistics for fit indices shown improvement in every
A
steps and final resultant came to moderate fit as- (CMIN/DF= 3.20, χ ² = 735, DF=
na
86, GFI= 0.815, NFI = 0.799, CFI = 0.789, RMSEA = 0.102, p =0.000) then
la
additional run was done and values were (CMIN/DF= 3.00, χ ² = 732.2, DF= 85,
au
GFI= 0.824, NFI = 0.810, CFI = 0.824, RMSEA = 0.101, p =0.000), in another
M
subsequent the result indicated moderate fit model with values (CMIN/DF= 2.11, χ ²
= 729.4, DF= 80, GFI= 0.857, NFI = 0.805, CFI = 0.822, RMSEA = 0.09, p =0.000).
185
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.30
Confirmatory Factor Analysis Goodness of fit of Cross Buying Latent and
Observed Variables –HDFC Bank
ty
1 2 3 4
i
rs
CMIN/DF 3.23 3.20 3.00 2.11
ve
2
ni
749.4 735.0 732.2 729.4
U
DF 90 86 85 80
lim
GFI 0.814 0.815 0.824 0.857
us
NFI 0.797 0.799 0.810 0.805
TABLE 5.31
br
1 CEX -
A
2 CEX 0.68
na
3 CEX 0.95
la
4 CEX 0.66
au
5 CEX -
M
6 PP -
7 PP -
8 PP 0.77
9 PP 0.87
10 PP 0.73
11 ELC 0.79
186
CHAPTER–5: ANALYSIS AND INTERPRETATION
12 ELC -
13 ELC 0.66
14 ELC 0.68
15 ELC -
16 APP 0.74
ty
17 APP 0.93
i
rs
18 ELC 0.74
19 ELC -
ve
20 ELC -
ni
21 AB 0.65
U
22 AB 0.91
lim
23 AB 0.62
us
24 AB -
Average Variance 60.0 58.3 50.6 65.3 54.6
Extracted (in %) M
h
Construct 0.73 0.74 0.74 0.71 0.74
ar
Reliability**
lig
TABLE 5.32
Li
187
CHAPTER–5: ANALYSIS AND INTERPRETATION
ity
rs
ve
ni
U
lim
us
M
Fig. 5.11: Confirmatory Factor Analysis of Cross Buying
h
ar
called confirmatory factor analysis was run on the latent variables and observed
A
variables of the research questionnaire. The purpose for using the CFA was to identify
y,
the validity of the questionnaire items in order to have the proper reliability of the
ar
study.
br
a) CFA CRME
Li
To validate the measurement model the fit of model, reliability and validity check was
d
za
performed. The back to back model fit indices were improved with the series of
modification indices and some items were deleted on the ground of low factor
A
loadings. The Parsimonious run results of CFA indicated poor fit model with values
na
(CMIN/DF= 3.01, χ ² =289.71, DF= 144, GFI= 0.808, NFI = 0.799, CFI = 0.821,
la
au
steps and final resultant came to moderate fit as- (CMIN/DF= 2.34, χ ² = 269.81,
DF= 140, GFI= 0.815, NFI = 0.827, CFI = 0.841, RMSEA = 0.109, p =0.000) then
additional run was done and values were (CMIN/DF= 2.10, χ ² = 234.24, DF= 139,
GFI= 0.884, NFI = 0.869, CFI = 0.884, RMSEA = 0.100, p =0.000), in another
188
CHAPTER–5: ANALYSIS AND INTERPRETATION
subsequent the result indicated good fit model with values (CMIN/DF= 2.04, χ ² =
229.70, DF= 138, GFI= 0.932, NFI = 0.875, CFI = 0.892, RMSEA = 0.095, p
=0.000).
TABLE 5.33
Confirmatory Factor Analysis Goodness of fit of CRME Latent and Observed
ty
Variables – Allahabad Bank
i
rs
Fit Indices Parsimonious Improved Model
ve
Indicators Model Indices
ni
1 2 3 4
U
CMIN/DF 3.01 2.34 2.10 2.04
lim
2 289.71 269.81 234.24 229.70
us
DF 144 140 139 38
M
h
GFI 0.808 0.815 .884 0.932
ar
CFI 0.892
y,
TABLE 5.34
CRME of Allahabad Bank Questionnaire Constructs
la
au
1 OC_1 DELETED
2 OC_2 DELETED
3 OC_3 DELETED
4 OC_4 DELETED
189
CHAPTER–5: ANALYSIS AND INTERPRETATION
5 OC_5 DELETED
6 OC_6 DELETED
7 CE_1 DELETED
8 CE_2 DELETED
9 CE_3 DELETED
ty
10 CE_4 DELETED
i
rs
11 CE_5 DELETED
ve
12 CE_6 DELETED
ni
13 PDA_1 0.67
U
14 PDA_2 0.74
lim
15 PDA_3 0.52
us
16 PDA_4 0.67
17 PDA_5 0.60
M
h
18 REL_1 0.72
ar
19 REL_2 0.54
lig
20 REL_3 0.60
A
21 REL_4 DELETED
y,
22 REL_5 DELETED
ar
23 TOR_1 DELETED
br
24 TOR_2 DELETED
Li
25 TOR_3 0.70
d
26 TOR_4 0.68
za
27 TOR_5 0.49
A
Average
na
Extracted (in %)
au
Construct
Reliability**
M
190
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
ni
U
lim
Fig 5.12: Confirmatory Factor Analysis of CRME of Allahabad Bank
us
b) CFACS
M
h
TABLE 5.35
ar
1 2 3 4
br
Li
191
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
S. Customer Satisfaction HDFC Bank
i
No. Items
rs
CO PQ COM TRU COMM SR CH
ve
1 CO_ DELETED
2 CO_ DELETED
ni
3 CO_ DELETED
U
4 CO_ DELETED
lim
5 CO_ DELETED
us
6 CO_ DELETED
7 PQ .74
8 PQ .76 M
h
9 PQ .69
ar
10 PQ .64
lig
11 PQ -
A
12 COM 0.72
y,
13 COM 0.80
ar
14 COM 0.68
br
15 COM -
Li
16 COM -
d
17 TRU -
za
18 TRU 0.63
A
19 TRU 0.72
na
20 TRU -
21 TRU 0.56
la
22 COMM -
au
23 COMM -
M
24 COMM .59
25 COMM .84
26 COMM .66
27 SR -
28 SR -
192
CHAPTER–5: ANALYSIS AND INTERPRETATION
29 SR .74
30 SR .86
31 SR .68
32 CH -
33 CH -
ty
34 CH .69
i
35 CH .73
rs
36 CH .61
ve
Average Variance - 46.75 54.00 41.00 49.00 59.00 46.00
ni
Extracted (in %)
U
Construct 0.81 0.75 0.74 0.74 0.74 0.75
Reliability**
lim
* AVE = Individual loading square sum / Sum of number of items*100
us
** CR = (Sum of Items loading)2/ (Sum of Items loading)2 + Std. error variance
TABLE 5.37
lig
193
CHAPTER–5: ANALYSIS AND INTERPRETATION
ity
rs
ve
ni
U
lim
us
M
h
ar
lig
A
c) CFA CL
br
To validate the measurement model the fit of model, reliability and validity check was
Li
performed. The back to back model fit indices was improved with the series of
d
modification indices and some items were deleted on the ground of low factor
za
loadings. The Parsimonious run results of CFA indicated poor fit model with values
A
(CMIN/DF= 3.43, χ ² =1329.8, DF= 58, GFI= 0.875, NFI = 0.775, CFI = 0.789,
na
carried on and the results of statistics for fit indices shown improvement in every
au
steps and final resultant came to moderate fit as- (CMIN/DF= 3.13, χ ² = 1319.8,
M
DF= 55, GFI= 0.885, NFI = 0.798, CFI = 0.799, RMSEA = 0.108, p =0.000) then
additional run was done and values were (CMIN/DF= 2.89, χ ² = 1311.8, DF= 52,
GFI= 0.898, NFI = 0.820, CFI = 0.828, RMSEA = 0.106, p =0.000), in another
194
CHAPTER–5: ANALYSIS AND INTERPRETATION
subsequent the result indicated good fit model with values (CMIN/DF= 221, χ ² =
394.1, DF= 48, GFI= 0.904, NFI = 0.830, CFI = 0.846, RMSEA = 0.093, p =0.000).
Confirmatory Factor Analysis Goodness of fit of Customer Loyalty Latent and
Observed Variables –Allahabad Bank
TABLE 5.38
i ty
Parsimonious
rs
Fit Indices Model Improved Model
ve
Indicators Indices
ni
1 2 3 4
U
CMIN/DF 3.43 3.13 2.89 2.21
lim
χ² 1329.8 1319.8 1311.8 394.1
DF 58 55 52 48
us
GFI 0.875 0.885 0.898 0.904
NFI 0.775 M
0.798 0.820 0.830
h
CFI 0.789 0.799 0.828 0.846
ar
Constructs
br
TABLE 5.39
Li
No. Items
SRB CI PCD IF CC
A
1 SRB .76
na
2 SRB .93
la
3 SRB .88
au
4 SRB -
5 SRB -
M
6 CI -
7 CI .52
8 CI .83
9 CI .63
10 CI -
195
CHAPTER–5: ANALYSIS AND INTERPRETATION
11 PCD .74
12 PCD .71
13 PCD .60
14 PCD -
15 PCD -
ty
16 IF -
17 IF .34
i
rs
18 IF .96
ve
19 IF .73
ni
20 IF -
U
21 CC DELETED
lim
22 CC DELETED
23 CC DELETED
us
24 CC DELETED
25 CC
Average Variance 73.7 45.0 47.0 M 52.0
DELETED
DELETED
h
Extracted (ix %)
ar
Reliability**
A
TABLE 5.40
Li
SRB CI PCD IF CC
na
196
CHAPTER–5: ANALYSIS AND INTERPRETATION
d) CFA CB
To validate the measurement model the fit of model, reliability and validity check was
performed. The back to back model fit indices were improved with the series of
modification indices and some items were deleted on the ground of low factor
loadings. The Parsimonious run results of CFA indicated poor fit model with values
ty
(CMIN/DF= 3.23, χ ² =749.4, DF= 90, GFI= 0.814, NFI = 0.799, CFI = 0.794,
i
rs
RMSEA = 0.105, p =0.000). Subsequently, a series of 3 improved run of CFA was
ve
carried on and the results of statistics for fit indices shown improvement in every
ni
steps and final resultant came to moderate fit as- (CMIN/DF= 3.21, χ ² = 735, DF=
U
86, GFI= 0.815, NFI = 0.801, CFI = 0.802, RMSEA = 0.104, p =0.000) then
lim
additional run was done and values were (CMIN/DF= 3.02, χ ² = 732.2, DF= 85,
GFI= 0.824, NFI = 0.812, CFI = 0.834, RMSEA = 0.102, p =0.000), in another
us
subsequent the result indicated good fit model with values (CMIN/DF= 2.06, χ ² =
M
654.5, DF= 108, GFI= 0.888, NFI = 0.867, CFI = 0.886, RMSEA = 0.095, p =0.000).
h
ar
Table 5.41
Confirmatory Factor Analysis Goodness of fit of Cross Buying Latent and
lig
1 2 3 4
br
Li
DF 90 86 85 108
A
197
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
3 .78
i
rs
4 .61
ve
5 .48
ni
6 .75
U
7 -
lim
8 .20
us
9 -
10 .72
11 .83 M
h
ar
12 .83
lig
13 .87
14 .60
A
15 .48
y,
16 --
ar
17 --
br
18 --
Li
19 --
d
za
20 --
21 .91
A
22 .61
na
23 .62
la
24 .78
au
Extracted (ix %)
Construct 0.82 0.71 0.82 Deleted 0.82
Reliability**
* AVE = Individual loading square sum / Sum of number of items*100
** CR = (Sum of Items loading)2/ (Sum of Items loading)2 + Std. error variance
198
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.42
Discriminant Validity of cross buying- Allahabad bank
ty
CEX 0.73 1 .007 0.025 Deleted 0.000
i
rs
PP 0.45 0.088** 1 0.012 Deleted 0.000
ve
ELC 0.47 0.161 ** 0.114 ** 1 Deleted Deleted
ni
U
APP Deleted Deleted Deleted Deleted 1 0.036
lim
AB 0.52 0.019 0.022 Deleted -0.190** 1
us
** Correlation is significant at the 0.01 level (2-tailed)
M
* Correlation is significant at the 0.05 level (2-tailed).*
h
ar
lig
A
y,
ar
br
Li
d
za
A
na
Fig. 5.14: Confirmatory Factor Analysis Goodness of fit of Cross Buying Latent
and Observed Variables – Allahabad Bank
la
aspects i.e. GOFI, Convergent validity and the Discriminant validity. As indicated in
the data given below in the table -5.43 , the data of the Allahabad and HDFC banks
both shows the moderate fit model , which is common on both the sides, but the data
of Allahabad bank was more refined as compared to HDFC bank. The GOFI of
HDFC bank were found to be better than the Allahabad bank, which mean the
199
CHAPTER–5: ANALYSIS AND INTERPRETATION
applicability of the latent and observed variables of HDFC are having more practical
implication than that of HDFC bank. Further, the convergent validity and
Discriminant validity was measured as a part of CFA to find out the convergence i.e.
their reliability of the constructs in the form of average variance extracted and
constructs reliability and further, the squared inter-construct correlation were
ty
computed in order to know the Discriminant validity of the constructs. If the (AVE >
i
rs
SIC), it means there is existence of Discriminant validity among the constructs.
ve
ni
U
lim
us
M
h
ar
lig
A
y,
ar
br
Li
d
za
A
na
la
au
M
200
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
TABLE 5.43
ni
A Comparison of CFA Goodness of fit of CRME model of Allahabad bank and HDFC Bank
U
lim
Final Indices of Convergent Validity Discriminant Validity
Goodness of Fit
us
FIT INDICES Allahabad HDFC Constructs Allahabad Bank HDFC Bank Allahabad Bank HDFC Bank
M
Bank Bank
CRME
AVE (in CR AVE (in %) CR AVE SIC AVE SIC
h
%)
ar
lig
CMIN/DF 2.52 2.04 OC 48.75 0.80 DEL DEL 0.48 0.000-0.077 DEL DEL
X2 140.6 229.70 60.25 0.79 DEL DEL 0.60 0.22-0.435 DEL DEL
A
CE
y,
DF 138 38 PDA 64.25 0.80 41.4 0.83 0.64 0.017-0.071 0.41 0.200-0.256
201
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
Allahabad and HDFC banks both shows the moderate fit model , which is common on
i
rs
both the sides, but the data of Allahabad bank was more refined as compared to
ve
HDFC bank. The GOFI of HDFC bank were found to be better than the Allahabad
ni
bank, which mean the applicability of the latent and observed variables of HDFC are
U
having more practical implication than that of HDFC bank. Further, the convergent
lim
validity and Discriminant validity was measured as a part of CFA to find out the
us
convergence i.e. their reliability of the constructs in the form of average variance
extracted and constructs reliability and further, the squared inter-construct correlation
M
were computed in order to know the Discriminant validity of the constructs. If the
h
ar
(AVE > SIC), it means there is existence of Discriminant validity among the
lig
constructs.
A
y,
ar
br
Li
d
za
A
na
la
au
M
202
CHAPTER–5: ANALYSIS AND INTERPRETATION
ity
rs
ve
TABLE 5.44
ni
A Comparison of CFA Goodness of fit of Customer Satisfaction model of Allahabad bank and HDFC Bank
U
lim
Final Indices of Convergent Validity Discriminant Validity
Goodness of Fit
us
FIT INDICES Allahabad HDFC Constructs Allahabad Bank HDFC Bank Allahabad Bank HDFC Bank
M
Bank Bank
Customer
AVE (In CR AVE (in %) CR AVE SIC AVE SIC
h
Satisfaction
%)
ar
lig
CMIN/DF 2.60 2.75 CO Deleted Deleted 64.6 .74 Deleted Deleted 0.64 .000-.066
A
1057.2 1745.4 PQ
y,
DF 139 225 COM 54.75 .75 55.6 .74 0.54 .003-.094 0.55 .001-.028
CFI 0.853 0.821 SR 59.00 .74 56.6 .73 0.59 .007 0.56 .331
d
za
203
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
Allahabad and HDFC banks both shows the moderate fit model , which is common on
i
rs
both the sides, but the data of Allahabad bank was more refined as compared to
ve
HDFC bank. The GOFI of HDFC bank were found to be better than the Allahabad
ni
bank, which mean the applicability of the latent and observed variables of HDFC are
U
having more practical implication than that of HDFC bank. Further, the convergent
lim
validity and Discriminant validity was measured as a part of CFA to find out the
us
convergence i.e. their reliability of the constructs in the form of average variance
extracted and constructs reliability and further, the squared inter-construct correlation
M
were computed in order to know the Discriminant validity of the constructs. If the
h
ar
(AVE > SIC), it means there is existence of Discriminant validity among the
lig
constructs.
A
y,
ar
br
Li
d
za
A
na
la
au
M
204
CHAPTER–5: ANALYSIS AND INTERPRETATION
ity
rs
ve
TABLE 5.45
ni
A Comparison of CFA Goodness of fit of Customer Loyalty model of Allahabad bank and HDFC Bank
U
Final Indices of Convergent Validity Discriminant Validity
lim
Goodness of Fit
FIT INDICES Allahabad HDFC Constructs Allahabad Bank HDFC Bank Allahabad Bank HDFC Bank
us
Bank Bank Customer
Loyalty AVE (in %) CR AVE (in %) CR AVE SIC AVE SIC
M
CMIN/DF 2.21 2.79 SRB 73.7 0.75 64.5 0.79 0.73 .009-.222 0.64 .008-.032
h
ar
X2 394.1 1312.8 CI 45.0 0.74 53.7 0.76 0.45 .029-.172 0.53 .002-.079
lig
DF 48 134 PCD 47.0 0.75 65.5 0.74 0.47 .034 0.65 .065-.346
A
GFI 0.904 0.830 IF 52.0 0.72 55.8 0.82 0.52 - 0.55 .276
y,
NFI 0.830 0.812 CC Deleted Deleted 61.6 0.73 Deleted Deleted 0.61 -
ar
br
CFI 0.846 0.827
Li
205
CHAPTER–5: ANALYSIS AND INTERPRETATION
A comparison on the constructs and indicators of Cross Buying was done on several
aspects i.e. Goodness of Fit indices, Convergent validity and the Discriminant
validity. As indicated in the data given below in the table – 5.44 , the data of the
ty
Allahabad and HDFC banks both shows the moderate fit model , which is common on
i
rs
both the sides, but the data of Allahabad bank was more polished as compared to
ve
HDFC bank. The GOFI of HDFC bank were found to be better than the Allahabad
ni
bank, which mean the applicability of the latent and observed variables of HDFC are
U
having more practical implication than that of HDFC bank. Further, the convergent
lim
validity and Discriminant validity was measured as a part of CFA to find out the
us
convergence i.e. their reliability of the constructs in the form of average variance
extracted and constructs reliability and further, the squared inter-construct correlation
M
were computed in order to know the Discriminant validity of the constructs. If the
h
ar
(AVE > SIC), it means there is existence of Discriminant validity among the
lig
constructs.
A
y,
ar
br
Li
d
za
A
na
la
au
M
206
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
TABLE 5.46
ni
A Comparison of CFA Goodness of fit of Cross buying model of Allahabad bank and HDFC Bank
U
Final Indices of Convergent Validity Discriminant Validity
lim
Goodness of Fit
us
FIT Allahabad HDFC Constructs Allahabad Bank HDFC Bank Allahabad Bank HDFC Bank
INDICES Bank Bank
Cross
M
AVE (In CR AVE CR AVE SIC AVE SIC
Buying
%) (in %)
h
ar
CMIN/DF 2.06 2.11 CEX 55.0 0.82 60.0 .73 0.55 0.000-.025 0.60 .001-.007
lig
X2 654.5 729.4 PP 37.3 0.71 58.3 .74 0.37 0.000-.012 0.58 .005-.088
A
DF 108 80 ELC 55.0 0.82 55.6 .74 0.55 deleted 0.55 .004-.015
y,
GFI 0.888 0.857 APP Deleted Deleted 65.3 .71 Deleted .036 0.65 .012
ar
br
NFI 0.867 0.805 AB 55.0 0.82 54.6 .74 0.55 - 0.54 -
Li
207
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
dependent variables. Though there are many ways to describe SEM, it is most
i
commonly thought of as a hybrid between some form of analysis of variance
rs
(ANOVA)/regression and some form of factor analysis. In general, it can be said that
ve
SEM allows performing some type of multilevel regression/ANOVA on factors.
ni
Structural equation modeling involves the measurement of the structural model and
U
with the structural model, the relationship between latent construct in nature of
lim
independent and dependent variables are measured. SEM provides a better way of
us
empirically examining a theoretical model by involving both measurement model and
M
structural model in one analysis” (Hair et al 2010). Thus, SEM is the accumulation of
h
measurement and structural model in total. The main goal of measurement theory
ar
under SEM is to depict a visual diagram in the form of casual relationship among
lig
parameters estimates which are also called path estimates, wherein the inferences are
y,
The Structural modeling was utilized for the measurement of cause and effect
br
reliability, and technological orientation of HDFC Bank with a sample size of 564
za
respondents.
A
na
Before proceeding to the causal relationship of CRME with customer satisfaction the
M
goodness of fit was measured. In the measurement of structural model of the GOF
indices the model was found to be perfectly good fit. The statistics for structural
model of GOF are as- (CMIN/DF = 1.84, Chi- Square = 27.60, DF = 15, GFI = 0.992,
NFI =0.978, CFI = 0.990, RMSEA = 0.039, P-Value =0.000) given in table 5.47.
208
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
proposed hypotheses. Hence, several inferences were drawn from the results. Initially
i
rs
CRME effect on customer satisfaction was found relatively positively significant
ve
explaining the variance with regression weight which is mentioned below in the table
ni
5.47.
U
TABLE 5.47
lim
Goodness of fit indices and CRME impact on Customer satisfaction HDFC Bank
us
Data
Fit Indices
M
Statistics for structured model
h
CMIN/DF 1.84
ar
DF 15
GFI 0.992
A
NFI 0.978
y,
CFI 0.990
RMSEA 0.039
ar
p-Value 0.000
br
209
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
TRU <--- PDA .030 .534 Not Supported
i
COMM <--- PDA .084 .064 Not Supported
rs
SR <--- PDA .131 .004 Supported
ve
CH <--- PDA .165 *** Supported
ni
CO <--- REL -.221 *** Not Supported
U
PQ <--- REL -.113 .016 Not Supported
lim
COM <--- REL -.009 .852 Not Supported
us
TRU <--- REL .121 .010 Supported
COMM <--- REL .079 .076 Not Supported
SR <--- REL .228 *** M Supported
h
CH <--- REL .143 .002 Supported
ar
Fig. 5.15: Goodness of fit indices and CRME impact on Customer satisfaction
HDFC Bank Data
210
CHAPTER–5: ANALYSIS AND INTERPRETATION
Before proceeding to the causal relationship of CRME with customer satisfaction the
goodness of fit was measured. In the measurement of structural model of the GOF
indices the model was found to be perfectly good fit. The statistics for structural
model of GOF are as- (CMIN/DF = 2.59, Chi- Square = 10.39, DF = 4, GFI = 0.996, NFI
ty
=0.990, CFI = 0.994, RMSEA = 0.053, P-Value =0.000) given in table 5.48.
i
rs
Moreover, there are 5 latent variables under CRME (organizational commitment,
ve
customer experience, process driven approach, reliability, technological orientation)
ni
which are independent variable to the customer satisfaction. On the other hand, the
U
dependent variable was customer loyalty having 5 constructs. The multiple mixtures
lim
of significant and insignificant effects were highlighted by the analysis of proposed
us
hypotheses. Hence, several inferences were drawn from the results. Initially CRME
M
effect on customer satisfaction was found relatively positively significant explaining
h
the variance with regression weight which is mentioned below in the table 5.48.
ar
TABLE 5.48
lig
Goodness of fit indices and CRME impact on Customer Loyalty HDFC Bank
A
Data
y,
CMIN/DF 2.59
br
DF 4
d
GFI 0.996
za
NFI 0.990
CFI 0.994
A
RMSEA 0.053
na
p-Value 0.000
la
211
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
CI <--- PDA .094 .040 Supported
i
rs
PCD <--- PDA .094 .048 Supported
ve
IF <--- PDA .085 .069 Not Supported
ni
CC <--- PDA .123 .010 Supported
U
SRB <--- REL .064 .163 Not Supported
lim
CI <--- REL .072 .111 Not Supported
PCD <--- REL .100 .034 Supported
us
IF <--- REL .148 .001 Supported
CC <--- REL .109 .020 M Supported
h
CI <--- TOR .230 *** Supported
ar
Fig. 5.16: Goodness of fit indices and CRME impact on Customer Loyalty HDFC
Bank Data
212
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
0.996, NFI =0.990, CFI = 0.994, RMSEA = 0.050, P-Value =0.000) given in table 5.49.
i
rs
Moreover, there are 7 latent variables under customer satisfaction (customized offer,
ve
perceived quality, competence, trust, communication, conflict handling) which are
ni
independent variable to the customer satisfaction. On the other hand, the dependent
U
variable was customer loyalty having 5 constructs. The multiple mixtures of
lim
significant and insignificant effects were highlighted by the analysis of proposed
us
hypotheses. Hence, several inferences were drawn from the results. Initially customer
M
satisfaction effect on customer loyalty was found relatively positively significant
h
explaining the variance with regression weight which is mentioned below in the table
ar
5.49.
lig
TABLE 5.49
A
CMIN/DF 2.41
Li
GFI 0.996
NFI 0.990
A
CFI 0.994
na
RMSEA 0.050
p-Value 0.000
la
213
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
CI <--- COM .055 *** Supported
i
rs
PCD <--- COM -.074 .108 Not Supported
ve
IF <--- COM .098 .908 Not Supported
ni
CC <--- COM .180 *** Supported
U
SRB <--- TRU .134 *** Supported
lim
CI <--- TRU -.016 .019 Not Supported
PCD <--- TRU .093 *** Supported
us
IF <--- TRU .553 .001 Supported
CC <--- TRU .525 M
*** Supported
h
SRB <--- COMM .210 .161 Not Supported
ar
214
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
0.986, NFI =0.943, CFI = 0.960, RMSEA = 0.057, P-Value =0.000) given in table 5.50.
i
rs
ve
Moreover, there are 7 latent variables under customer satisfaction (customized offer,
perceived quality, competence, trust, communication, conflict handling) which are
ni
independent variable to the customer satisfaction. On the other hand, the dependent
U
variable was cross buying having 4 constructs. The multiple mixtures of significant
lim
and insignificant effects were highlighted by the analysis of proposed hypotheses.
us
Hence, several inferences were drawn from the results. Initially customer satisfaction
M
effect on cross buying was found relatively positively significant explaining the
h
variance with regression weight which is mentioned below in the table 5.50.
ar
TABLE 5.50
lig
CMIN/DF 2.81
br
DF 17
GFI 0.986
d
za
NFI 0.943
CFI 0.960
A
RMSEA 0.057
na
p-Value 0.000
Paths of constructs showing Customer Estimate P Hypothesis Decision
la
215
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
ELC <--- COM .115 .019 Supported
i
rs
APP <--- COM -.037 .465 Not Supported
ve
AB <--- COM .041 .411 Not Supported
ni
CEX <--- TRU .032 .436 Not Supported
U
PP <--- TRU -.024 .565 Not Supported
lim
ELC <--- TRU .092 .027 Supported
us
APP <--- TRU .096 .024 Supported
AB <--- TRU .144
M
*** Supported
h
CEX <--- COMM .163 *** Supported
ar
216
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
ni
U
lim
us
M
Fig. 5.17: Goodness of fit indices and Customer Satisfaction impact on Cross
h
Buying HDFC Bank Data
ar
lig
Before proceeding to the causal relationship of customer loyalty with cross buying the
y,
goodness of fit was measured. In the measurement of structural model of the GOF
ar
indices the model was found to be perfectly good fit. The statistics for structural
br
model of GOF are as- (CMIN/DF = 2.68, Chi- Square = 21.4, DF = 8, GFI = 0.993, NFI
Li
=0.973, CFI = 0.982, RMSEA = 0.055, P-Value =0.006) given in table 5.51
d
Moreover, there are 5 latent variables under customer loyalty (stronger relational
za
customer commitment) which are independent variable to the cross buying. On the
na
other hand, the dependent variable was cross buying having 4 constructs. The multiple
la
proposed hypotheses. Hence, several inferences were drawn from the results. Initially
M
217
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.51
Goodness of fit indices and Impact of Customer Loyalty on Cross Buying HDFC
Bank Data
ty
Chi- Square 21.4
i
DF 8
rs
GFI 0.993
ve
NFI 0.973
ni
CFI 0.982
U
RMSEA 0.055
p-Value 0.006
lim
us
Paths of constructs showing Estimate P Hypothesis Decision
Customer Loyalty impact on Cross
Buying
M
h
CEX <--- SRB .134 .002 Supported
ar
218
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
AB <--- CC .002 .972 Not Supported
i
rs
AB <--- SRB .098 .025 Supported
ve
ni
Structural Model of Allahabad Bank Data
U
a) Impact of CRME on Customer Satisfaction
lim
Before proceeding to the causal relationship of CRME with customer satisfaction the
us
goodness of fit was measured. In the measurement of structural model of the GOF
M
indices the model was found to be perfectly good fit. The statistics for structural
h
model of GOF are as- (CMIN/DF = .501, Chi- Square = 1.03, DF = 2, GFI = 0.991, NFI
ar
=0.999, CFI = 0.999, RMSEA = 0.000, P-Value =0.000) given in table 5.52
lig
which are independent variable to the customer satisfaction. On the other hand, the
ar
proposed hypotheses. Hence, several inferences were drawn from the results. Initially
d
za
explaining the variance with regression weight which is mentioned below in the
na
table5.52
la
au
M
219
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.52
Goodness of fit indices and CRME impact on Customer satisfaction Allahabad
Bank Data
Fit Indices Statistics for structured model
CMIN/DF 0.501
Chi- Square 1.03
ty
DF 2
i
GFI 0.991
rs
NFI 0.999
ve
CFI 0.999
ni
RMSEA 0.000
U
p-Value 0.000
lim
Paths of constructs showing CRME Impact Estimate P Hypothesis Decision
us
on Customer Satisfaction
Not Supported
y,
220
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
ni
U
lim
Fig. 5.18: Goodness of fit indices and CRME impact on Customer satisfaction
us
Allahabad Bank Data
b) Impact of CRME on Customer Loyalty
M
h
Before proceeding to the causal relationship of CRME with customer satisfaction the
ar
goodness of fit was measured. In the measurement of structural model of the GOF
lig
indices the model was found to be perfectly good fit. The statistics for structural
A
model of GOF are as- (CMIN/DF = 2.73, Chi- Square = 9.39, DF = 44, GFI = 0.991, NFI
y,
=0.995, CFI = 0.990, RMSEA = 0.010, P-Value =0.000) given in table 5.53
ar
br
dependent variable was customer loyalty having 5 constructs. The multiple mixtures
A
hypotheses. Hence, several inferences were drawn from the results. Initially CRME
effect on customer satisfaction was found relatively positively significant explaining
la
au
the variance with regression weight which is mentioned below in the table 5.53
M
221
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.53
Goodness of fit indices and CRME impact on Customer Loyalty Allahabad Bank
Data
CMIN/DF 2.73
ty
Chi- Square 9.39
i
rs
DF 44
ve
GFI 0.991
ni
NFI 0.995
U
CFI 0.990
RMSEA 0.010
lim
p-Value 0.000
us
Paths of constructs showing CRME Estimate P Hypothesis Decision
Impact on Customer Satisfaction
M
h
SRB <--- PDA -.024 .625 Not Supported
ar
222
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
ni
U
Fig. 5.19: Goodness of fit indices and CRME impact on Customer Loyalty
lim
Allahabad Bank Data
us
c) Impact of Customer Satisfaction on Customer Loyalty
M
Before proceeding to the causal relationship of customer satisfaction with customer
h
loyalty the goodness of fit was measured. In the measurement of structural model of
ar
the GOF indices the model was found to be perfectly good fit. The statistics for
lig
structural model of GOF are as- (CMIN/DF = 1.06, Chi- Square = 3.20, DF = 3, GFI =
A
0.999, NFI =0.997, CFI = 0.999, RMSEA = 0.011, P-Value =0.000) given in table 5.54
y,
Moreover, there are 7 latent variables under customer satisfaction (customized offer,
ar
independent variable to the customer satisfaction. On the other hand, the dependent
Li
hypotheses. Hence, several inferences were drawn from the results. Initially customer
A
explaining the variance with regression weight which is mentioned below in the table
la
5.54
au
M
223
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.54
Goodness of fit indices and Customer Satisfaction on Customer Loyalty
Allahabad Bank Data
ty
Fit Indices Statistics for structured model
i
rs
ve
CMIN/DF 1.06
Chi- Square 3.20
ni
DF 3
U
GFI 0.999
lim
NFI 0.997
CFI 0.999
us
RMSEA 0.011
p-Value
M
0.000
h
ar
224
CHAPTER–5: ANALYSIS AND INTERPRETATION
i ty
rs
ve
ni
U
lim
us
M
h
ar
lig
A
y,
buying the goodness of fit was measured. In the measurement of structural model of
za
the GOF indices the model was found to be perfectly good fit. The statistics for
A
structural model of GOF are as- (CMIN/DF = 1.46, Chi- Square = 4.39, DF = 3, GFI =
na
0.998, NFI =0.988, CFI = 0.996, RMSEA = 0.029, P-Value =0.000) given in table 5.55
la
au
Moreover, there are 7 latent variables under customer satisfaction (customized offer,
M
225
CHAPTER–5: ANALYSIS AND INTERPRETATION
effect on cross buying was found relatively positively significant explaining the
variance with regression weight which is mentioned below in the table 5.55
TABLE 5.55
Goodness of fit indices and Customer Satisfaction impact on Cross Buying of
Allahabad Bank Data
ity
Fit Indices Statistics for structured model
rs
ve
CMIN/DF 1.46
Chi- Square 4.39
ni
DF 3
U
GFI 0.998
lim
NFI 0.988
us
CFI 0.996
RMSEA 0.029
p-Value M
0.000
h
ar
226
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
PP <--- CH -.031 .703 Not Supported
i
rs
ELC <--- CH .033 .660 Not Supported
ve
ni
AB <--- CH .098 .124 Not Supported
U
lim
us
M
h
ar
lig
A
y,
ar
br
Li
d
za
A
na
Fig. 5.21: Goodness of fit indices and Customer Satisfaction impact on Cross
la
Before proceeding to the causal relationship of customer loyalty with cross buying the
goodness of fit was measured. In the measurement of structural model of the GOF
indices the model was found to be perfectly good fit. The statistics for structural
model of GOF are as- (CMIN/DF = .147, Chi- Square = 1.29, DF = 2, GFI = 0.999, NFI
=0.998, CFI = 0.997, RMSEA = 0.000, P-Value =0.006) given in table 5.56.
227
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
highlighted by the analysis of proposed hypotheses. Hence, several inferences were
i
rs
drawn from the results. Initially customer satisfaction effect on cross buying was
ve
found relatively positively significant explaining the variance with regression weight
ni
which is mentioned below in the table 5.56.
U
TABLE 5.56
lim
Goodness of fit indices and Impact of Customer Loyalty on Cross Buying
us
Allahabad Bank Data
Fit Indices
M
Statistics for structured model
h
CMIN/DF 0.147
ar
DF 2
A
GFI 0.999
y,
NFI 0.998
ar
CFI 0.997
br
RMSEA 0.000
Li
p-Value 0.000
d
228
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
CEX <--- IF -.020 .648 Not Supported
i
rs
ELC <--- IF .072 .094 Not Supported
ve
ni
PP <--- IF .000 .995 Not Supported
U
AB <--- IF .002 .960 Not Supported
lim
AB <--- SRB .013 .783 Not Supported
us
M
h
ar
lig
A
y,
ar
br
Li
d
za
A
na
la
au
Fig. 5.22: Goodness of fit indices and Impact of Customer Loyalty on Cross
M
229
CHAPTER–5: ANALYSIS AND INTERPRETATION
The structured equation model was run on the two different data of the same
ty
constructs and variables on the equal data set. The structured model was used to find
i
rs
out the complex causal effect of the multiple relationships. As the objective of the
ve
study, the comparison was made after calculation of impact of CRME constructs on
ni
the customer satisfaction constructs.
U
Initially, the goodness of fit model was calculated on the data of HDFC bank and
lim
Allahabad Bank individually. Subsequently, it was compared, It explained that the
us
data represented good fit model. After the GOF Model of the two different data, the
M
effect of CRME on Customer satisfaction was compared, which explained the mixture
of positive and negative relationships, indicating that some of the hypothesised
h
ar
relationships were supported and some were not supported. Further, the separate
lig
analysis of Allahabad bank data was carried out and it was found that most of the
A
items were deleted as those constructs and indicators were failed to attain the standard
y,
threshold limits.
ar
So, in comparison of the cause and effect relationship the data of HDFC Bank was
br
found more suitable than that of the Allahabad Bank on the front of practical
Li
terms of customer retention, customer services, trust maintained by the bank and the
za
commitment they have given to their customers in the case of HDFC Bank than that
A
of Allahabad Bank.
na
la
au
M
230
CHAPTER–5: ANALYSIS AND INTERPRETATION
TABLE 5.57
Comparison of Goodness of fit indices and CRME impact on Customer
satisfaction of HDFC Bank and Allahabad Bank
i ty
CMIN/DF 1.84 0.501
rs
Chi- Square 27.60 1.03
ve
DF 15 2
ni
GFI 0.992 0.991
U
NFI 0.978 0.999
lim
CFI 0.990 0.999
RMSEA 0.039 0.000
us
p-Value 0.000 0.000
M
h
Estimate
ar
lig
231
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
TRU <--- PDA .030 .158
i
rs
COMM <--- PDA .084 .317
ve
ni
SR <--- PDA .131 .039
U
CH <--- PDA .165 .163
lim
CO <--- REL -.221 -.040
us
PQ <--- REL -.113 .090
The structured equation model was run on the two different data of the same
constructs and variables on the equal data set. The structured model was used to find
out the complex causal effect of the multiple relationships. As the objective of the
232
CHAPTER–5: ANALYSIS AND INTERPRETATION
study, the comparison was made after calculation of impact of CRME constructs on
the customer loyalty constructs.
Initially, the goodness of fit model was calculated on the data of HDFC Bank and
Allahabad Bank individually. Subsequently, it was compared It explained that the
data represented good fit model. After the GOF Model of the two different data, the
ty
effect of CRME on Customer loyalty was compared, which explained the mixture of
i
rs
positive and negative relationships, indicating that some of the hypothesized
ve
relationships were supported and some were not supported. Further, the separate
ni
analysis of Allahabad bank data was carried out and it was found that most of the
U
items were deleted as those constructs and indicators were failed to attain the standard
lim
threshold limits.
us
So, in comparison of the cause and effect relationship the data of HDFC Bank was
M
found more suitable than that of the Allahabad Bank on the front of practical
h
implications. It means the Customer relation management effectiveness is appealing
ar
in terms of customer retention, customer services, trust maintained by the bank and
lig
the commitment they have given to their customers in the case of HDFC Bank than
A
TABLE 5.58
ar
DF 4 44
la
HDFC ALLAHABAD
233
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
SRB <--- CE -.001 DELETED
i
rs
CI <--- CE -.118 DELETED
ve
PCD <--- CE -.015 DELETED
ni
U
IF <--- CE .132 DELETED
lim
CC <--- CE .089 DELETED
us
CI <--- PDA
M.094 .373
h
PCD <--- PDA .094 -.021
ar
234
CHAPTER–5: ANALYSIS AND INTERPRETATION
The structured equation model was run on the two different data of the same
constructs and variables on the equal data set. The structured model was used to find
out the complex causal effect of the multiple relationships. As the objective of the
ty
study, the comparison was made after calculation of impact of Customer Satisfaction
i
rs
constructs on the customer loyalty constructs.
ve
Initially, the goodness of fit model was calculated on the data of HDFC Bank and
ni
Allahabad Bank individually. Subsequently, it was compared It explained that the
U
data represented good fit model. After the GOF Model of the two different data, the
lim
effect of CRME on customer loyalty was compared, which explained the mixture of
us
positive and negative relationships, indicating that some of the hypothesized
M
relationships were supported and some were not supported. Further, the separate
h
analysis of Allahabad Bank data was carried out and it was found that most of the
ar
items were deleted as those constructs and indicators were failed to attain the standard
lig
threshold limits.
A
So, in comparison of the cause and effect relationship the data of HDFC Bank was
y,
found more suitable than that of the Allahabad Bank on the front of practical
ar
terms of customer retention, customer services, trust maintained by the bank and the
Li
commitment they have given to their customers in the case of HDFC Bank than that
d
of Allahabad Bank.
za
A
TABLE 5.59
Comparison of Goodness of fit indices and Customer satisfaction on Customer
na
235
CHAPTER–5: ANALYSIS AND INTERPRETATION
HDFC ALLAHABAD
ty
PCD <--- CO .221 DELETED
i
rs
IF <--- CO .569 DELETED
ve
SRB <--- PQ -.003 DELETED
ni
CI <--- PQ .146 DELETED
U
lim
PCD <--- PQ .060 DELETED
us
CC <--- PQ .015 DELETED
236
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
SRB <--- CH -.003 DELETED
i
rs
CI <--- CH .146 DELETED
ve
PCD <--- CH .060 DELETED
ni
U
IF <--- CH -.009 DELETED
lim
CC <--- CH .015 DELETED
us
M
d) Impact of Customer satisfaction on Cross Buying- A Comparison
h
The structured equation model was run on the two different data of the same
ar
constructs and variables on the equal data set. The structured model was used to find
lig
out the complex causal effect of the multiple relationships. As the objective of the
A
study, the comparison was made after calculation of impact of CRME constructs on
y,
Initially, the goodness of fit model was calculated on the data of HDFC Bank and
Li
data represented good fit model. After the GOF Model of the two different data, the
za
effect of CRME on cross buying was compared, which explained the mixture of
A
relationships were supported and some were not supported. Further, the separate
la
analysis of Allahabad Bank data was carried out and it was found that most of the
au
items were deleted as those constructs and indicators were failed to attain the standard
M
threshold limits.
So, in comparison of the cause and effect relationship the data of HDFC Bank was
found more suitable than that of the Allahabad Bank on the front of practical
implications. It means the customer relation management effectiveness is appealing in
terms of customer retention, customer services, trust maintained by the bank and the
237
CHAPTER–5: ANALYSIS AND INTERPRETATION
commitment they have given to their customers in the case of HDFC Bank than that
of Allahabad Bank.
TABLE 5.60
Comparison of Goodness of fit indices and Customer satisfaction on Cross
ty
Buying of HDFC Bank and Allahabad Bank
i
rs
Fit Indices HDFC Bank Allahabad Bank
ve
CMIN/DF 2.81 0.147
ni
Chi- Square 47.70 1.29
U
DF 17 2
lim
GFI 0.986 0.999
NFI 0.943 0.998
us
CFI 0.960 0.997
RMSEA 0.057
M 0.000
h
p-Value 0.000 0.000
ar
lig
HDFC ALLAHABAD
A
238
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
PP <--- COMM .108 -.062
i
rs
ELC <--- COMM .069 -.032
ve
APP <--- COMM -.086 .106
ni
AB <--- COMM .054 -.043
U
CEX <--- SR .109 .033
lim
PP <--- SR -.054 .000
us
ELC <--- SR -.104 -.031
APP <---
M
SR .154 .033
h
AB <--- SR -.135 .098
ar
The structured equation model was run on the two different data of the same
na
constructs and variables on the equal data set. The structured model was used to find
la
out the complex causal effect of the multiple relationships. As the objective of the
au
study, the comparison was made after calculation of impact of Customer Loyalty
M
Initially, the goodness of fit model was calculated on the data of HDFC Bank and
Allahabad Bank individually. Subsequently, it was compared. It explained that the
data represented good fit model. After the GOF Model of the two different data, the
effect of customer loyalty was compared, which explained the mixture of positive and
239
CHAPTER–5: ANALYSIS AND INTERPRETATION
So, in comparison of the cause and effect relationship the data of HDFC Bank was
ty
found more suitable than that of the Allahabad Bank on the front of practical
i
rs
implications. It means the customer relation management effectiveness is appealing in
ve
terms of customer retention, customer services, trust maintained by the bank and the
ni
commitment they have given to their customers in the case of HDFC Bank than that
U
of Allahabad Bank.
lim
TABLE 5.61
us
Comparison of Goodness of fit indices and Customer Loyalty on Cross Buying of
HDFC Bank and Allahabad Bank
M
h
ar
lig
DF 8 2
GFI 0.993 0.999
br
HDFC ALLAHABAD
la
au
240
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
CEX <--- PCD -.068 .134
i
rs
PP <--- PCD .113 -.021
ve
ELC <--- PCD .135 -.020
ni
U
APP <--- PCD .071 .072
lim
AB <--- PCD -.005 .000
us
PP <---
M
IF .099 .013
h
ELC <--- IF .060 DELETED
ar
241
CHAPTER–5: ANALYSIS AND INTERPRETATION
REFERENCES
Abraham S.J., E. and Rao T.V., (1991), ‘Human Resource Development; Practices in
Indian Industries’ – A Trend Report, Oxford & IBH Publishing Co., Pvt.
ty
Limited, New Delhi, P.262 – 268.
i
rs
Bless C., Higson-Smith, C., &Kagee, A. (2006). Fundamentals of social research
ve
methods: An African perspective. Juta and Company Ltd.
ni
Bless, C., Higson-Smith, C., &Kagee, A. (2006). Fundamentals of social research
U
methods: An African perspective. Juta and Company Ltd.
lim
Cronbach, L. J. (1951). Coefficient alpha and the internal structure of
us
tests. Psychometrika, 16(3), 297-334.
M
Fornell, C., &Larcker, D. F. (1981). Structural equation models with unobservable
h
variables and measurement error: Algebra and statistics.Journal of marketing
ar
research, 382-388.
lig
George, D., &Mallery, P. (2003).SPSS for windows step by step: A simple guide and
A
Hair, J. F., Anderson, R. E., Babin, B. J., & Black, W. C. (2010). Multivariate data
la
analysis: A global perspective (Vol. 7). Upper Saddle River, NJ: Pearson.
au
Hair, J. F., Black, W. C., Babin, B. J., Anderson, R. E., &Tatham, R. L. (2006).
M
242
CHAPTER–5: ANALYSIS AND INTERPRETATION
ty
Kaiser, H. F. (1974). An index of factorial simplicity.Psychometrika, 39, 31-36.
i
rs
Kothari, C. R. (2012). Research methodology methods and techniques.New Delhi:
ve
Malhotra, N. K. &Dash, S. (2010). Marketing Research: An Applied
ni
Orientation.Prentice Hall, Delhi.
U
Malik., A (2016), Sports Celebrity Endorsement in Positioning of Brand for
lim
Measuring Purchase Intention of Customers Unpublished Thesis, Submitted to
us
Pondicherry University, 19 September, 2016
M
Neuman, Lawrence W. (2000) Social Research Methods; Qualitative and Quantitative
h
Approaches. Boston, MA: Allyn and Bacon. New Age International.
ar
243