|
5.6.3 Discriminant Validity
Discriminant validity refers to the degree to which a construct differs from other constructs
(Hair
et al.
2006). An approach for establishing discriminant validity is to compare the
χ
2
difference which involved 91 comparisons of the constrained and unconstrained measurement
models of all construct pairs. For example, the phi value for each pair of the fourteen latent
variables in the unconstrained measurement models were constrained to 1, one at a time, and
then
χ
2
difference was calculated for each pair of unconstrained and constrained models (Li
and Dant 1997). The results of these comparisons appear in Table 5.17. Of the 91
comparisons conducted, 5 comparisons did not pass the
χ
2
difference test (i.e.,
∆χ
2
> 3.841,
p
> .01), thus indicating that the constructs did not discriminate. It can be found that all of the
five failed comparisons (EXP-DES, DES-ESQ, DES-LOY, PSC-EUN, and EUN-TCS)
involved the site design (DES) and environmental uncertainty (EUN) constructs, clearly
indicating that these two measures were problematic. On this basis, site design (DES) and
environmental uncertainty (EUN) were eliminated from all further analyses.
249
Table 5.17 Results of Discriminant Analysis
Comparisons
Constrained model
Unconstrained model
Chi-square difference
Discriminant
validity
χ
2
df
χ
2
df
∆χ
2
∆
df
EXP IAA
90.63 14 47.11
13
43.52
1
Yes
PQC
614.17
23
141.60
22
472.57
1 Yes
DES
151.52
29
149.14
28 2.38 1
No
ESQ
169.10
58
157.59
57
11.51 1 Yes
REP 60.11
18
47.12
17
12.99 1 Yes
CONV
126.59
22
113.98
21
12.61 1 Yes
PSC 771.83
22 81.28 21 690.55 1
Yes
EUN 633.46
18 61.85 17 571.61 1
Yes
TCS
1130.27
69
287.11
68
843.16
1 Yes
CSAT
76.59
14
71.89
13 4.70 1 Yes
LOY
74.21
23
68.35
22 5.86 1 Yes
RISK
336.07
23
254.34
22
81.73 1 Yes
ENJ
138.96
23
110.39
22
28.57 1 Yes
IAA
PQC 700.81
17 68.67 16 632.14 1
Yes
DES 204.16
23 95.18 22 108.98 1
Yes
ESQ
112.69
48
100.75
47
11.94 1 Yes
REP
115.82
14
45.45
13
70.37 1 Yes
CONV
183.24
17
97.61
16
85.63 1 Yes
PSC 896.16
17 79.39 16 816.77 1
Yes
EUN
858.35
14
69.35
13 789 1 Yes
TCS
997.29
58
175.39
57
821.90
1 Yes
CSAT
136.77
12
108.65
11
28.12 1 Yes
LOY
81.89
18
61.87
17
20.02 1 Yes
RISK
325.38
18
132.72
17
192.66
1 Yes
ENJ
218.11
18
107.28
17
110.83
1 Yes
PQC DES 1651.52 35 145.22 34 1506.30 1
Yes
ESQ
1188.28
67
138.26
66
1050.02
1 Yes
REP 884.35
22 99.91 21 784.44 1
Yes
CONV
667.03
27 97.44 26 569.59 1
Yes
PSC
181.04
27
171.78
26 9.26 1 Yes
EUN
106.49
22
87.01
21
19.48 1 Yes
TCS
300.54
79
294.48
78 6.06 1 Yes
CSAT
949.54
17 49.47 16 900.07 1
Yes
LOY
1006.02
28
187.82
27
818.20
1 Yes
RISK
1048.82
28
132.46
27
916.36
1 Yes
ENJ 802.52
28 72.97 27 729.55 1
Yes
DES ESQ 296.82 79 296.43 78
0.39
1
No
REP
175.11
29
167.92
28 7.19 1 Yes
CONV
194.26
35
168.09
34
26.17 1 Yes
PSC
1481.24
35
162.03
34
1319.21
1 Yes
EUN 1013.75
29 96.03 28 917.72 1
Yes
TCS
1555.67
92
312.36
91
1243.31
1 Yes
CSAT
125.48
23
117.98
22 7.50 1 Yes
LOY
225.15
36
224.11
35 1.04 1
No
RISK
334.40
36
283.37
35
51.03 1 Yes
ENJ
180.10
36
143.60
35
36.50 1 Yes
250
ESQ REP
340.64 58 335.20 57
5.44
1
Yes
CONV
204.68
67
199.62
66 5.06 1 Yes
PSC
1435.61
67
179.08
66
1256.53
1 Yes
EUN
996.49
58
158.14
57
838.35
1 Yes
TCS
1646.66
139
348.78
138
1297.88
1 Yes
CSAT
173.44
49
99.01
48
74.43 1 Yes
LOY
276.68
68
231.69
67
44.99 1 Yes
RISK
311.20
68
302.06
67 9.14 1 Yes
ENJ
194.45
68
189.04
67 5.41 1 Yes
REP CONV 136.78 22 115.31 21
21.47
1
Yes
PSC
1129.43
22
102.80
21
1026.63
1 Yes
EUN 813.93
18 96.81 17 717.12 1
Yes
TCS
1259.91
69
202.48
68
1057.43
1 Yes
CSAT
83.91
14
59.44
13
24.47 1 Yes
LOY
165.29
23
156.98
22 8.31 1 Yes
RISK
241.09
23
199.05
22
42.04 1 Yes
ENJ
123.50
23
100.96
22
22.54 1 Yes
CONV PSC
940.32
27
100.49
26
839.83
1
Yes
EUN
945.67
22
262.09
21
683.58
1 Yes
TCS
1263.51
79
272.07
78
991.44
1 Yes
CSAT
120.48
17
115.86
16 4.62 1 Yes
LOY
115.80
28
111.69
27 4.11 1 Yes
RISK
345.74
28
272.65
27
73.09 1 Yes
ENJ
322.86
28
299.08
27
23.78 1 Yes
PSC EUN 139.47 22 136.01 21
3.46
1
No
TCS
287.98
79
283.94
78 4.04 1 Yes
CSAT
1205.44
17
83.02
16
1122.42
1 Yes
LOY
1307.47
28
151.68
27
1155.79
1 Yes
RISK
1178.15
28
207.34
27
970.81
1 Yes
ENJ
1034.89
28
131.95
27
902.94
1 Yes
EUN TCS
249.75 69 246.91 68
2.84
1
No
CSAT
1431.70
14
442.92
13
988.78
1 Yes
LOY
826.84
23
127.28
22
699.56
1 Yes
RISK
1565.81
23
460.46
22
1105.35
1 Yes
ENJ
941.32
23
116.25
22
825.07
1 Yes
TCS CSAT 1448.55 59 174.66 58 1273.89 1
Yes
LOY
1498.68
80
267.80
79
1230.88
1 Yes
RISK
1340.42
80
446.54
79
893.88
1 Yes
ENJ
1206.74
80
220.67
79
986.07
1 Yes
CSAT LOY
357.56
18
217.95
17
139.61
1
Yes
RISK
186.71
18
181.75
17 4.96 1 Yes
ENJ 92.64
18
86.25
17 6.39 1 Yes
LOY RISK 262.37 29 248.79 28
13.58
1
Yes
ENJ
156.14
29
151.40
28 4.74 1 Yes
RISK ENJ
247.89 29 190.12
28
57.77
1
Yes
5.6.4 Model Fit for New Measurement Model
251
Due to unsatisfactory discriminant validity, DES and EUN were removed from the initial
overall measurement model. The modified overall measurement model consisted of twelve
constructs. Re-estimation of the model was conducted to check whether the new
measurement model provides a good fit to the data and the measures have sufficient
convergent and discriminant validity.
As shown in Table 5.18, the key overall model fit statistics were:
χ
2
=3839.698 with 1779
degrees of freedom (p<0.001).
The
χ
2
/df (
2.158
)
falls within the suggested value of 5 or below.
In addition, the other indices satisfied the recommended value (CFI= 0.932, GFI = 0.963, TLI
= 0.959, RMSEA = 0.035). Therefore, there was a reasonable overall fit between the model
and the observed data.
252
Table 5.18 Results of the New Measurement Model of Latent Variables
Construct/Indicator Stand.
Loadings
Stand.
Error
Critical
Ratio
AVE CR
Internet Access Availability (IAA)
.65 .848
IAA1 .76
--
--
IAA2 .86
.044
26.27
*
IAA3 .81
.054
20.31
*
Perceived Internet Expertise (EXP)
.71 .906
EXP1 .74
--
--
EXP2 .87
.041
26.71
*
EXP3 .88
.041
27.14
*
EXP4 .87
.035
31.57
*
Product Quality Concern (PQC)
.75 .938
PQC1
.80
-- --
PQC2 .83
.032
34.89
*
PQC3 .90
.039
32.67
*
PQC4 .94
.043
32.69
*
PQC5 .87
.042
31.03
*
E-Service Quality (ESQ)
.64 .839
REL
.81
-- --
RESP .79
.043
21.30
*
PERS .79
.041
23.16
*
Reputation of Online Store (REP)
.66 .886
REP1
.80
-- --
REP2 .82
.032
31.38
*
REP3 .84
.038
27.25
*
REP4 .80
.037
25.50
*
Perceived Convenience (CONV)
.59 .876
CONV1 .75
--
--
CONV2 .78
.038
27.51
*
CONV3 .82
.044
24.58
*
CONV4 .72
.049
21.56
*
CONV6 .75
.048
20.99
*
Privacy and Security Concerns (PSC)
.73 .930
PSC1 .77
--
--
PSC2 .91
.033
37.15
*
PSC3 .93
.041
31.61
*
PSC4 .87
.041
29.13
*
PSC5 .79
.038
27.87
*
Perceived Consumer Transaction Costs (TCS)
.67 .856
PRTC .86
--
--
COTC .78
.047
19.93
*
POTC .80
.047
20.15
*
Customer Satisfaction (CSAT)
.78 .916
CSAT1 .89
--
--
CSAT2 .91
.024
42.84
*
CSAT3 .86
.026
37.83
*
Customer Loyalty (LOY)
.71 .923
LOY1 .83
--
--
LOY2 .85
.025
40.88
*
LOY3 .86
.031
33.28
*
LOY4 .85
.033
32.46
*
LOY5 .81
.035
29.81
*
Risk-Bearing Propensity (RISK)
.51 .839
RISK1
.73
-- --
RISK2 .73
.050
19.35
*
RISK3 .71
.050
18.72
*
RISK4 .69
.057
16.74
*
RISK5 .73
.051
19.50
*
Perceived Enjoyment (ENJ)
.66 .906
ENJ1
.73
-- --
ENJ2 .81
.039
28.24
*
ENJ3 .83
.046
23.98
*
ENJ4 .85
.049
24.17
*
ENJ5 .84
.048
23.85
*
Do'stlaringiz bilan baham: |
