Information Technology Reference
In-Depth Information
Table 2. Standardized loadings and reliabilities across gender
Total (N = 339)
Male (n
1
= 209)
Female (n
2
= 130)
Indicators retained
in the model
Construct
Standardized loading
(
t
value)
Standardized loading
(
t
value)
Standardized loading
(
t
value)
1
0.88 (19.85)
0.89 (16.10)
0.84 (11.47)
2
0.86 (19.12)
0.85 (14.96)
0.86 (11.81)
Loyalty
(F1)
4
0.85 (19.00)
0.86 (15.23)
0.84 (11.48)
7
0.66 (13.24)
0.63 (
0
9.88)
0.73 (
0
9.34)
Cronbach's alpha
0.88
0.88
0.89
8
0.90 (20.90)
0.89 (16.39)
0.90 (12.92)
Relationship quality
(F2)
9
0.89 (20.46)
0.92 (17.22)
0.83 (11.32)
11
0.83 (18.51)
0.84 (14.80)
0.82 (11.19)
Cronbach's alpha
0.90
0.92
0.89
19
0.79 (16.85)
0.82 (13.95)
0.73 (
0
9.31)
20
0.85 (18.83)
0.87 (15.42)
0.81 (10.71)
Expertise
(F3)
21
0.80 (17.17)
0.77 (12.68)
0.88 (12.21)
22
0.78 (16.39)
0.80 (13.53)
0.74 (
0
9.50)
Cronbach's alpha
0.88
0.89
0.87
23
0.80 (16.62)
0.84 (14.14)
0.70 (
0
8.39)
Relational selling behavior (F4)
24
0.75 (15.14)
0.77 (12.51)
0.71 (
0
8.58)
25
0.63 (12.09)
0.62 (
0
9.46)
0.65 (
0
7.70)
Cronbach's alpha
0.77
0.78
0.73
30
0.78 (15.26)
0.80 (12.35)
0.76 (
0
9.06)
Perceived network quality (F5)
32
0.78 (15.19)
0.78 (12.01)
0.77 (
0
9.25)
Cronbach's alpha
0.76
0.77
0.74
35
0.79 (16.73)
0.79 (13.26)
0.77 (
0
9.95)
Service recovery
(F6)
36
0.81 (17.42)
0.81 (13.79)
0.80 (10.57)
37
0.84 (18.22)
0.82 (13.87)
0.88 (12.18)
Cronbach's alpha
0.85
0.85
0.86
item: my agent is highly skilled. Relational sell-
ing behavior (F4) with five items also is modified
from Crosby et al. (1990); a sample item: my agent
offers appropriate information according to my
situation. Moreover, perceived network quality
(F5) with four items is modified from the ideas of
Jun and Cai (2001); a sample item: ADSL connec-
tion quality is reliable. Finally, service recovery
(F6) is measured using four items modified from
Andreassen (2000) and Boshoff and Leong (1998);
a sample item: my agent has tried to solve my
problems promptly.
measurement model and reliability
Following data collection, SEM (structural
equation modeling) is applied to conduct data
analysis. SEM is a multivariate statistical tech-
nique used to confirm the causal relations among
latent variables. This study follows a two-step
Search WWH ::
Custom Search