Understanding consumer online shopping behaviour from the perspective of transaction costs

Table 5.32 Comparison of Path Coefficients via T-tests 
Paths Search 
Products 
Paths St Error 
Experience Products 
Paths St Error 
t value 
Results 
H1a 
-.056(n.s.) 
.028 
-.070 
.024 
8.348 
Significantly stronger for experience products 
H1b 
-.108 
.028 
-.167 
.022 
36.520 
Significantly stronger for experience products* 
H1c 
-.244 
.021 
-.219 
.021 
18.483 
Significantly stronger for search products* 
H2a 
-.012(n.s.) 
.032 .088 .023 56.057 
Significantly 
stronger 
for 
experience 
products 
H2c 
-.578 
.105 
-.407 
.043 
33.661 
Significantly stronger for search products* 
H2d 
-.128 
.041 
-.063(n.s.) 
.026 
29.665 
Significantly stronger for search products 
H3a 
-.083 
.036 
-.106 
.024 
11.765 
Significantly stronger for experience products* 
H3b 
.036(n.s.) 
.032 .077 .025 22.257 
Significantly 
stronger 
for 
experience 
products 
H4a 
-.690 
.238 
-.777 
.179 
6.448 
Significantly stronger for experience products* 
H4b 
-.387 
.095 
-.607 
.089 
37.076 
Significantly stronger for experience products* 
H4c 
-.800 
.084 
-.821 
.081 
3.945 
Significantly stronger for experience products* 
H5 
.391 
.055 
.283 
.044 
33.781 
Significantly stronger for search products* 
Note:*Indicates support for both paths being significant 

---

287
Furthermore, R-square values for the endogenous variables, including consumer TCs, online 
purchase behaviour, loyalty and satisfaction, were also compared across the modes, with 
some differences found, for instance, in the search product model: TCs = .92, online purchase 
behaviour = .48, customer loyalty = .55, satisfaction = .64; whereas in the experience product 
model: TCs = .96, online purchase behaviour = .60, customer loyalty = .73, satisfaction = .68. 
The average variance accounted for in experience products model (AVA= .74) was greater 
than search products model (AVA= .65).
In terms of the significance of the corresponding paths in two models, H9a was partially 
supported while H9b was rejected. Results shown in Table 5.29 exhibit bigger structural 
change in which four paths show differences. Analysis undertaken in Table 5.30 indicates 
that the model nearly remains unchanged structurally, except one path from reputation of 
online store to consumer TCs. On the other hand, in terms of the strength of the 
corresponding paths in two models, the proposed model differs across the two product 
categories, thereby supporting H9a and H9b.