Journal of Electronic Commerce Research, VOL 8, NO.1, 2007
The first hypothesis (a consumer’s trust in an Internet shop is positively related to the perceived size of its physical store network) is supported by the significant path between perceived size and trust in the first sub-sample and the overall sample. The second hypothesis (a consumer’s trust in an Internet shop is positively related to the perceived reputation of its physical store network) receives strong support from the significant path between perceived reputation and trust in all three samples. In the same way, the third hypothesis (a consumer’s trust in an e- shop of a multi-channel retailer is positively related to the perceived privacy at the e-shop) is supported by the significant path between perceived privacy and trust in all three samples. Thus, our results confirm that a consumer’s trust in an Internet shop is positively related to the perceived size and perceived reputation of its store network. The influence of perceived online privacy on trust is even stronger than that of perceived reputation and size of the store network.
Hypothesis 4 assuming a negative influence of trust on risk perception and hypothesis 5 assuming an influence of risk perception on willingness to buy have not been confirmed with the conservative methodical approach presented above. Figure 3 summarizes our findings with regard to the original hypotheses (path coefficients from the total sample). 5.3 Familiarity Index and Trust
The surveyed data allowed for a more in-depth analysis of the factor influences of perceived size, reputation and privacy on trust. As the survey respondents reported whether or not they previously visited the e-shop and/or the physical store and whether they bought there, it is possible to model sub-groups of the sample, which differed in the degree of familiarity with the retailer. This approach is based on Bhattacherjee’s  attempt to model trust in relation to familiarity. Gefen  defines familiarity as an “activity-based cognizance based on previous experience or learning” (p. 727).
Willingness to Buy
Figure 3: Summary of findings with regard to original hypotheses
The reason for this explorative analysis is to find out whether different levels of familiarity lead to different factor scores in our model. In order to find a measure of familiarity, subgroups were identified that differed in the characteristics of previous visits and/or purchases to either store or site. For each of the nine possible groups, we defined a familiarity index. Our index values are based on the assumption that familiarity increases with the events "visit" and "purchase" at an e-shop and/or store. First, we assumed that users who know both e-shops and stores have a higher familiarity index than users of just one channel because they have the opportunity to experience the retailer on a broader basis. Second, it has been assumed that people who purchased at the retailer have a higher familiarity index than people who just visited store or site. A reason for this assumption is that a purchase is a trust- critical event that requires the users’ willingness to engage in a financial transaction and to bear the associated risks.
Therefore, individuals who had previously shopped at both the retailer’s site and store receive the highest familiarity index, while subjects with no previous visits to either site or store (first-time visitors) received the lowest index. The assignments of familiarity indices to the groups along with their appropriate group sizes are given in Table 5. The non-dichotomized numbers of previous purchases at and visits to an e-shop or store have not been further considered in the familiarity index because the group sizes would have been too small.