Start Date

16-8-2018 12:00 AM

Description

Prediction tools as one type of online decision aid (DA) has gained popularity worldwide due to their facilitating role in consumers’ online decision-making. However, consumers still have concerns about trusting its predictions because of agency problem. One solution is to provide explanation facilities on its interface. In IS literature, more types of explanations other than well studied “how explanation” and “why explanation” have been called to study. Thus, we propose a new type of explanation facilities (i.e., “if not explanations”). Building on prior works in this vein and a classic expectancy view of trust, we develop a theoretical model, which delineates the impacts of three types of explanation facilities (i.e., how explanations, why explanations and if not explanations) on users’ trusting beliefs via two assessments of DAs (i.e., perceived prediction process transparency and perceived advice quality). An experimental study is designed to test the hypothesized model.

Share

COinS
 
Aug 16th, 12:00 AM

Prediction Tool for Consumer Decision Making in E-Commerce: Exploring “If not” type of Explanation Facilities on Trust

Prediction tools as one type of online decision aid (DA) has gained popularity worldwide due to their facilitating role in consumers’ online decision-making. However, consumers still have concerns about trusting its predictions because of agency problem. One solution is to provide explanation facilities on its interface. In IS literature, more types of explanations other than well studied “how explanation” and “why explanation” have been called to study. Thus, we propose a new type of explanation facilities (i.e., “if not explanations”). Building on prior works in this vein and a classic expectancy view of trust, we develop a theoretical model, which delineates the impacts of three types of explanation facilities (i.e., how explanations, why explanations and if not explanations) on users’ trusting beliefs via two assessments of DAs (i.e., perceived prediction process transparency and perceived advice quality). An experimental study is designed to test the hypothesized model.