To Trust or Distrust Online Recommendation Agents? An Experimental Investigation into the Effect Mechanisms of Recommendation Neutrality and Transparency Strategies

Project: Research

View graph of relations


As Internet fraud is on the rise, consumers have become increasingly skeptical towards online recommendation services which are likelybiasedtowards the sponsors. Therefore, userdistrustin these services has emerged as a critical factor in influencing the effectiveness of the recommendation services.Distrusthas been generally defined as confidentnegativeexpectations regarding another’s conduct (e.g., act harmfully), whereastrusthas been generally regarded as confidentpositiveexpectations regarding another’s conduct (e.g., act benevolently). Although distrust and trust have been widely considered as distinctive constructs, the existing literature primarily examines thebeneficialimpacts of recommendation services (e.g., user trust). Consequently, this project intends to investigate the effects ofrecommendation neutrality(whether the product recommendations provided are neutral or biased towards the sponsors of the recommended products) on both trust anddistrustin online recommendation agents (RAs) that elicit users’ preferences and facilitate their decision making by providing shopping advice on what product to buy.Furthermore, the researcher will investigate the strategies that can mitigate userdistrustand compensate user trust in RAs when these RAs provide product recommendationsbiasedtowards sponsors. In particular, the researcher propose to examine the effects of two transparency strategies (explanation facilities and a declaration of sponsorship of the RA), and the effects of theinteractionof these two transparency strategies with each other and with recommendation neutrality. Theasymmetriceffects of trust and distrust in RAs on user’s behavioral intentions towards the RAs will also be tested.The researcher draw the intellectual foundations from Luhmann’s (1979) articulation of trust and distrust as distinct but potentially coexistent mechanisms for relationship building. Recent Information Systems research has advanced the understanding of trust in e-commerce. Nevertheless, research ondistrustis still in the preliminary stage. This research will fill the theoretical gap in the development of and in mitigating userdistrustin an emerging domain of online recommendation technologies and RAs in particular. It will contribute to research by revealing the effect mechanisms of recommendation neutrality and two transparency strategies on user trust anddistrustin RAs and by examining the differential consequences of user trust and distrust. As many providers of recommendation technologies profit from sponsors, the results of this study will provide practitioners with guidelines for designing trustworthy and distrust-free recommendation technologies. With the provision of effective recommendation services, companies can build stronger relationships with online consumers via such services, leading to more successful e-businesses.


Project number9041465
Grant typeGRF
Effective start/end date1/12/0925/02/13