Understanding consumer online shopping behaviour from the perspective of transaction costs

 
Step Two – Confirmatory
 
Constructs/Indicators CFA 
 
Stand. Loadings 
Stand. 
Error 
Critical 
Ratio
AVE CR 
TCS – Second-order construct 
PRTC 
.85 
-- -- .66 
.854 
COTC .78 
059 
16.87
*
POTC .80 
059 
17.08
*
PRTC – first-order construct 
PRTC2 
n/a 
n/a n/a .70 
.874 
PRTC3 .77 
-- 
-- 
PRTC4 .83 
.030 
36.49
*
PRTC5 .90 
.044 
25.40
*
COTC – first-order construct 
COTC1 
.85 
-- -- .77 
.911 
COTC2 .85 
.038 
26.83
*
COTC3 .94 
.040 
28.05
*
POTC – first-order construct 
POTC2 
n/a 
n/a n/a .76 
.925 
POTC3 .85 
-- 
-- 
POTC4 .92 
.027 
39.78
*
POTC5 .88 
.036 
29.73
*
POTC6 .82 
.036 
27.18
*
POTC7 n/a 
n/a 
n/a 
POTC8 n/a 
n/a 
n/a 
Model Fit 
Statistic 
Model Fit 
Statistic 
χ
2
104.41 
TLI 
.984 
df
28 
CFI 
.990 
P .000 
RMSEA 
.053 
GFI .979 
RMR 
.075 
* 
p 
< .001
5.5.11 Preliminary Analysis – Customer Satisfaction 
 
The customer satisfaction construct was measured by seven items (CSAT1 to CSAT3). The 
two-step preliminary data analysis process shown in Figure 5.1 was employed as a screening 
mechanism for data measured by these items (refer Table 5.12 for results). Evaluation of the 
correlation matrix through the KMO and Bartlett’s Test results in a high KMO statistic (.916) 
and a significant probability level (p< .001) for the Bartlett’s test. These results indicate that 
sufficient correlations were found within the correlation matrix for factor analysis to proceed. 
In addition, bivariate correlations were inspected and all coefficients fell within the 

---

238
acceptable range for factor analysis of .30 and .90. EFA was then conducted which produced 
a single factor structure with strong item loading ranging from .91 to .92 and the variance 
explained was 85.40%. Cronbach’s alpha of .91 was computed indicating good reliability of 
the scale. At this point, as all the items met the criteria of Step One, they were retained for 
CFA analysis in Step Two. 
In Step Two (refer Figure 5.1), CFA was conducted on the three items retained from Step 
One. Statistics indicating a good fit of the measurement model included 
χ
2
/df (.622) which 
was less than 5, GFI
(.974
), TLI (.983) and CFI (.977), which were above .90. RMSEA (.004) 
was less than .08 and RMR (.028) was less than .05. As a result, the CFA results indicated 
that this measurement model provided a very good fit to the data. Moreover, table 5.12 lists 
that the standardized factor loadings of the three items ranged from .85 to .93, all exceeding 
the preferable cutoff of .70 (Hair
 et al.
2006). The AVE for this construct was .78, exceeding 
the recommended criterion of .50 (Hair
 et al.
2006). The CR was .916, exceeding the 
threshold of .70 (Nunnally 1978). Therefore, items CSAT1 to CSAT3 were considered to 
provide a reliable and valid measurement scale for the construct, customer satisfaction. All 
five items were retained for constructing the overall measurement model. 

---

239