|
Table 5.16 Results of the Measurement Model of Latent Variables
Construct/Indicator Stand.
Loadings
Stand.
Error
Critical
Ratio
AVE CR
Internet Access Availability (IAA)
.65 .849
IAA1 .76
--
--
IAA2 .86
.044
26.36
*
IAA3 .80
.054
20.38
*
Perceived Internet Expertise (EXP)
.71 .906
EXP1 .74
--
--
EXP2 .87
.041
26.73
*
EXP3 .88
.041
27.13
*
EXP4 .87
.035
31.59
*
Product Quality Concern (PQC)
.75 .938
PQC1
.80
-- --
PQC2 .83
.032
34.93
*
PQC3 .90
.039
32.71
*
PQC4 .94
.043
32.71
*
PQC5 .87
.042
31.05
*
Site Design (DES)
.65 .916
DES1
.71
-- --
DES3 .79
.046
27.21
*
DES4 .81
.055
23.67
*
DES5 .88
.060
23.56
*
DES6 .83
.055
24.38
*
DES8 .80
.056
23.30
*
E-Service Quality (ESQ)
.63
.838
REL
.81
-- --
RESP .79
.043
21.46
*
PERS .78
.041
23.41
*
Reputation of Online Store (REP)
.66 .886
REP1
.80
-- --
REP2 .82
.032
31.40
*
REP3 .84
.038
27.28
*
REP4 .80
.037
25.53
*
Perceived Convenience (CONV)
.61 .887
CONV1 .73
--
--
CONV2 .86
.041
29.46
*
CONV3 .88
.047
25.73
*
CONV4 .73
.051
21.43
*
CONV6 .69
.047
20.44
*
Privacy and Security Concerns (PSC)
.73 .931
PSC1 .77
--
--
PSC2 .91
.033
37.21
*
PSC3 .93
.040
31.68
*
PSC4 .87
.041
29.22
*
PSC5 .79
.038
27.94
*
Environmental Uncertainty (EUN)
.66 .884
EUN1 .70
--
--
EUN2 .78
.036
29.53
*
EUN3 .88
.050
24.52
*
EUN4 .88
.051
24.54
*
Perceived Consumer TCs (TCS)
.67 .856
PRTC .86
--
--
COTC .78
.047
19.91
*
246
POTC .80
.047
20.14
*
Customer Satisfaction (CSAT)
.78 .916
CSAT1 .89
--
--
CSAT2 .91
.024
42.86
*
CSAT3 .86
.026
37.87
*
Customer Loyalty (LOY)
.71 .923
LOY1 .83
--
--
LOY2 .85
.025
40.97
*
LOY3 .86
.031
33.40
*
LOY4 .85
.033
32.46
*
LOY5 .80
.035
29.82
*
Risk-Bearing Propensity (RISK)
.52 .843
RISK1
.75
-- --
RISK2 .78
.043
23.12
*
RISK3 .76
.043
22.54
*
RISK4 .61
.049
16.72
*
RISK5 .71
.045
20.52
*
Perceived Enjoyment (ENJ)
.67 .910
ENJ1
.71
-- --
ENJ2 .81
.039
28.43
*
ENJ3 .86
.048
24.75
*
ENJ4 .88
.051
25.07
*
ENJ5
.83 .050
23.48
*
Model Fit
χ
2
5609.781
TLI
.946
df 2415 CFI
.950
P .000 RMSEA
.037
GFI .914
RMR
.116
*
p
< .001
5.6.1 Overall Model Fit
The key overall model fit statistics were:
χ
2
=
5609.781 with 2415 degrees of freedom (p =
0.000).
The
χ
2
/df (
1.081
)
was less than 5. As suggested by Hair et al. (2006), the model fit
should rely on at least one absolute fit index and one incremental fit index, in addition to the
χ
2
goodness-of-fit test statistic. The value for RMSEA, an absolute fit index, was .035 which
fell into acceptable value range (< .08) (Hair
et al.
2006). CFI, an incremental fit index,
was .950 which exceeded the CFI guideline of .90 for a model of this complexity and sample
size (Hair
et al.
2006). Thus, the results indicate a good fit for the overall measurement model.
Further, using the RMSEA and CFI satisfied the “rule of thumb” of Hair et al. (2006) that
both a badness-of-fit index and a goodness-of-fit index be evaluated. In addition, the other
index values were also supportive. For example, the GFI was .914, the TLI was .946,
exceeding the fit criteria of .90 (Hair
et al.
2006).
247
5.6.2 Convergent Validity
Convergent validity detects whether the measures of a construct are more correlated with one
another than with measures of other constructs (Balabanis
et al.
2006). It can be evaluated in
three ways i.e., by inspecting the AVE for each construct, by evaluating the strength and
significance of the factor loadings, and by examining the CRs.
As suggested by Fornell and Larcker (1981), convergent validity is achieved if the AVE in
items by their respective construct is greater than the variance unexplained (AVE > .50). The
results in the table 5.16 indicated that the AVEs for all constructs exceeded .50. Specifically,
the AVE for these constructs are as follows -- Internet access availability (.65), perceived
Internet expertise (.70), product quality concern (.75), site design (.65), e-service quality (.63),
reputation of online store (.66), perceived convenience (.61), privacy and security concerns
(.73), environmental uncertainty (.66), consumer TCs (.67), customer satisfaction (.78),
customer loyalty (.71), risk-bearing propensity (.52) and perceived enjoyment (.67).
In addition, convergent validity can also be assessed by factor loadings. Following the
recommendations of Hair et al. (2006), factor loadings should be greater than .50 and critical
ratios should be greater than 1.96. As shown in the table 5.16, all of the standardized item
loadings were larger than 0.5 and critical ratio values exceeded 1.96, indicating that all
loadings are significant at 0.001. Therefore the results provided another support for a high
degree of convergence.
The reliability of the constructs was estimated by CR. The CRs for all constructs in table 5.16
was above the recommended .70 level (Bagozzi
et al.
1991). Thus, the reliability was
confirmed for these constructs. The CR of each of the constructs is as follows --
Internet
248
access availability (.849), perceived Internet expertise (.906), product quality concern (.938),
site design (.916), e-service quality (.838), reputation of online store (.886), perceived
convenience (.887), privacy and security concerns (.931), environmental uncertainty (.884),
consumer TCs (.856), customer satisfaction (.916), customer loyalty (.923), consumer’s risk-
bearing propensity (.843) and perceived enjoyment (.910). In summary, the reliability and
convergent validity of the constructs were satisfactory.
Do'stlaringiz bilan baham: |
