Understanding consumer online shopping behaviour from the perspective of transaction costs

5.9.3 Moderating Effect of Perceived Enjoyment of Online Shopping on Perceived TCs 

 Online Purchase Behaviour 
To test the moderating effect of perceived enjoyment of online shopping on the relationship 
between perceived TCs and online purchase behaviour (H8a), a two-group basic structural 
model was set up using the two groups of perceived enjoyment. Initially, the two-group basic 
structural model was tested allowing all the hypothesized relationships, including the path 
from TCs to online purchase behaviour, to be estimated freely in both groups. Next, a second 
model was tested adding a constraint fixing the path of TCs 

online purchase behaviour to 
be equal in both groups. The fit results for each model are presented in Table 5.27. 
The GFI, CFI and RMSEA of the totally free (TF) model were better than those of the 
constrained model. The 
∆χ
2 
= 9.729 with 1
 df
was significant (
∆χ
2 
> 3.841,
p < .05). Given 
model fit statistics and the significant 
∆χ
2
value, the conclusion is that constraining the path 
of TCs 

online purchase behaviour to be equal between the two enjoyment groups 

---

273
produced a worse fit. Thus, the TF model in which the perceived TCs 

online purchase 
behaviour relationship was freely estimated in both groups was supported. This result 
indicated that perceived enjoyment of online shopping moderated the relationship between 
TCs and actual online purchase. Therefore, H8a was supported. 
The unstandardized parameter estimates for the TF model results were examined and it was 
found that the TCs 

online purchase behaviour relationship was significant in both groups 
(p < .001). Specifically, the relationship was greater for the low enjoyment with a completely 
standardized estimate of -.82 as compared to a completely standardized estimate of -.65 for 
the high enjoyment group.