Quantifying the impact on distrust of e-commerce trust factors: A non-parametric study

Distrust, besides its current use as the antonymous of trust, has also been proposed as a separate and relevant aspect in the adoption of business to consumer e-commerce. This work empirically addresses this issue, analyzing and comparing probabilistic dependencies between consumers' perceptions of distrust and trust, site characteristics and opinions of respondents about the site or the vendor. Bi-variate mutual information is used to associate not only ordinal data as usually, but also ordinal (opinions) with nominal data (identified characteristics). The latter reveals to be a more effective approach since it allows to discriminate between mixed, pure trust and pure distrust factors. Tri-variate mutual information further identifies synergies through which trust factors might influence distrust. In addition to mutual information, Kendall's tau correlation identifies weak inverse relations in ordinal information. Focusing on the consumer without previous contact with a certain seller or site, the findings help to explain why trust was viewed as opposed to distrust and why they can instead be considered separate constructs.