Travel Reference
In-Depth Information
superior CRM capability were developed by Day
and Van den Bulte (2002). To estimate CRM profit-
ability, we adopted the scale of CRM profitability
developed by Roh et al. (2005). The five-items with
five point likert scale focus on how the implementa-
tion of CRM could profit the organization.
significant factor loading (p < .01), providing
evidence for convergent validity. Table 1 presents
descriptive statistics and variable correlations.
Results of the CFA showed good model-to-data
fit (χ
2
(224) = GFI = 0.86, AGFI = 0.83, CFI = 0.91,
IFI = 0.91, RMSEA = .07). The study examined
alternative models to confirm discriminant validity
as show in Table 2. The fit indices for the four-fac-
tor model (baseline model) were superior to those
for the three-factor model (Δχ
3.2.1
Control variables
We controlled five important demographic factors
that include: gender, age, level of education, work-
ing experience and company size (Amburgey &
Rao, 1996).
2
= 78.44, p < .001),
the two-factor model (Δχ
2
= 89.75, p < .001), and
one- factor model (Δχ
2
= 195.25, p > .001). The
baseline model is the best fit among those models.
The results provided evidence for the discriminant
validity among these core constructs.
4 RESULTS
4.1
Confirmatory factor analyses
Cronbachs alpha's result for the scales are 0.91 for
CRM training orientation, 0.88 for CRM organi-
zation-wide, 0.90 for CRM capability and 0.89 for
CRM profitability. All results are greater than 0.70
which is above the acceptable threshold. Composite
reliability and average variance extracted are
higher than the evaluation criteria, with compos-
ite reliabilities larger than 0.70 and average vari-
ance extracted are larger than 0.50 with highly
4.2
Hypothesis tests
The first, we examined the effects of control vari-
ables to dependent variable. Education level and size
of firms had a positive impact to CRM profitability.
Other control variables were insignificant (R
2
= .05,
p < .001). Model 1 measured the effect of independ-
ent variable, CRM training orientation and CRM
organization-wide to mediator, superior CRM
capabilities (R
2
= 0.11, p < .001). The model meets
the first condition of Baron and Kenny's (1986)
Table 1.
Variable descriptive statistical and correlations (
N
= 314).
Variables
Mean
Sd
1
2
3
4
5
6
7
8
9
1. Gender
.45
.50
-
2. Age
1.90
.89
.05
-
3. Education level
1.85
.90
.05
(.10)
-
4. Experience of working
1.69
1.01
(.02)
(.04)
(.03)
-
5. Size of company
3.24
1.18
(.04)
(.05)
.00
(.09)
-
6. CRM training orientation
4.82
.97
.03
.07
.06
(.07)
(.04)
-
7. CRM organization-wide
4.01
.69
.02
.08
.07
(.09)
(.08)
(.06)
-
8. Superior CRM capability
4.34
.99
.09
.16**
.03
(.01)
.10
.20**
.17**
-
9. CRM profitability
3.86
.72
.06
.07
.14*
(.04)
.14*
.14*
.15*
.24**
-
Note:
*p < .01; *p < .05.
Table 2.
Results of the CFA.
Model
d.f
GFI
AGFI
CFI
IFI
RMSEA
χ
2
Δχ
2
Four-factor Model (baseline model)
592.25
224
.86
.83
.91
.91
.07
Three-factor Model (Model 1)
1389.40
227
797.15***
.67
.60
.72
.72
.13
Two-factor Model (Model 2)
1916.60
229
1324.30***
.61
.53
.60
.60
.15
One-factor Model (Model 3)
2701.04
230
2108.79***
.50
.40
.40
.40
.19
Note
: In the four-factor model, four constructs were treated as four independent factors. In the three-factor model
(Model 1), CRM training orientation and CRM organization-wide were combined into one factor. In the two-factor
model (Model 2), CRM training orientation and CRM organization-wide were combined into one factor; CRM capa-
bilities and CRM profitability were combined into one factor. In the one-factor model (Model 3), all constructs were
combined into one factor; **
p
< .01*
p
< .05.
Search WWH ::
Custom Search