Trbovich, P.L., & Patrick, A.S.  (2004).  The impact of context upon trust formation in ambient societies.  Position paper presented at the CHI 2004 Workshop on Considering Trust in Ambient Societies, April 26, Vienna, Austria.

The Impact of Context Upon Trust Formation in Ambient Societies

Patricia L. Trbovich
HOTLab, Carleton University
Ottawa, ON KIS 5B6 Canada
ptrbovich@rogers.com

Andrew S. Patrick
National Research Council of Canada
1200 Montreal Rd. Ottawa, ON Canada  K2H 8S2
Andrew.Patrick@nrc-cnrc.gc.ca       

Introduction

With the advent of ubiquitous computing, people can access information and goods in many different forms (Web pages, e-commerce sites, online auctions). With these new sources for gathering goods and information comes increasing concern about whether to trust the information and goods that are being exchanged. Our research examines the primary antecedents for development of trust in an online vendor, and subsequent intentions to purchase. Specifically, our research examines the impact of contextual factors upon formation of trust and intentions to purchase online. We introduce the notion of context to combine, for the first time, models of trust formation with contextual manipulations, and to address how variations in context influence the development of customer trust in online vendors.

The concept of ambient Intelligence (AmI) provides a vision as to how information technology, in which users are empowered through a digital environment, will develop. Amongst the challenges presented by ambient intelligence are the integration of technology into our society, and the facilitation of human interaction with these technologies. A vital component of AmI is user context (Dogac, Laleci & Kabak, 2003). Dogac, Laleci and Kabak (2003) define user context as any information that characterizes the user and his/her situation. They note that their definition includes information such as current temporal and spatial location, user preferences, and profile. They note that integration of user context information, with user profile and preferences, can facilitate the transmission and marketing of goods and information to the appropriate e-commerce market. E-commerce is an increasingly pervasive element of ambient intelligence.

Electronic-commerce involves purchase of products and services, payment, and/or exchange of information over the Internet or other electronic means (Chen & Dhillon, 2003, McKnight, Choudhury & Kacmar, 2002). There is a growing reliance upon Electronic Commerce (E-Commerce) as a medium for exchanging goods and information. This trend represents a challenge for the HCI community, to ensure that the exchange of such goods and information integrates trust and security. Trust may be defined as "users' thoughts, feelings, emotions, or behaviors that occur when they feel that an entity can be relied upon to act in their best interest when they give up direct control" (Patrick, 2002). Building customer trust is essential for electronic-Commerce vendors, because trust in a vendor affects whether a customer decides to buy online (McKnight, Choudhury & Kacmar, 2002). There are a variety of antecedents involved in the formation of trust between an e-commerce consumer and vendor. Perceived vendor competence, benevolence and integrity are amongst the more important trust formation antecedents. Although the importance of trust has been demonstrated in the literature, the antecedents of consumer online trust and how they influence overall trust formations warrant further study (Bhattacherjee, 2002). Our research focuses on the primary antecedents involved in the formation of trust between a consumer and an e-commerce vendor, with particular attention to the impact of context upon this process.

The context of an e-commerce purchase may dramatically impact formation of trust between an E-commerce consumer and vendor. For example, urgency may require a consumer to accelerate their formation of trust in a vendor, more rapidly than they might otherwise proceed. Similarly, an expensive or career critical purchase decision may represent contextual factors that could significantly impact consumer trust formation. Accordingly, the inter-relationship between trust formation and context represents a significant dynamic when designing e-commerce interfaces. With respect to the workshop on "Considering Trust in Ambient Societies", we propose to discuss the following issues: (a) the principal antecedents to formation of online trust, (b) the degree and manner in which these antecedents foster trust, (c) the impact of context manipulation upon the effectiveness of these antecedents in fostering trust, (d) the relationship between online trust and intentions to purchase, and (e) the role of contextual factors in moderating the relationship between trust and the intention to purchase. Answers to these questions can contribute to a better understanding of how trust, context and the inter-relationship between trust and context will impact e-commerce and other ambient societies.

Trust Building Model

Following a review of models of trust in e-commerce, we selected the trust component of the McKnight, Choudhury & Kacmar (2002) model as the most appropriate model for purposes of our proposed research. The McKnight et al. model was selected because it provides a framework for evaluating and quantifying the impact of various trust antecedents in the design of e-commerce human-computer interfaces. While other trust models identify such trust building antecedents, the McKnight et al. model further affords a concrete and testable approach for evaluating the impact of such antecedents in the design of interfaces. The McKnight et al. model achieves this through the use of Likert scale questionnaires.

McKnight et al. (2002) developed a model for establishment of customer trust in an E-commerce vendor. In their model, trust is defined as a multi-dimensional construct based on two inter-dependent elements. Specifically, these two elements are Trusting beliefs and Trusting intentions/willingness to depend.  They define Trusting beliefs as the customerís perception of the vendorís competence, benevolence and integrity. Trusting intentions/willingness to depend are defined as the customerís decision to render themselves vulnerable to the vendors. If customers develop positive trusting beliefs and positive trusting intentions/willingness to depend, the model contends that trust is established and customers decide to engage in transactions leading to and/or resulting in e-commerce (i.e., following advice offered by the vendor, sharing personal information with web vendor, purchasing from vendorís website). Conversely, negative impressions hinder trust and obstruct e-commerce. Development of positive or negative impressions of Trusting beliefs and Trusting intentions/willingness to depend are subject to three determinants. These determining factors are Web vendor reputation, perceived web site quality and structural assurance (i.e., customerís perception of web environment safety).

Context

Upon completion of our review of various definitions of context, we selected the Dey, Abowd and Salber (2001) definition of context as the most applicable to our research. It is important to integrate context into the development of interactive computer applications where the userís context may vary (Dey, Abowd & Salber, 2001). To understand the implications of differing contexts in the establishment of trust in E-commerce, it is essential to understand and define what we mean by context. Context is defined in a number of ways by various researchers, however, these definitions are not readily applicable in a research setting. To address this shortcoming, Dey, Abowd and Salber (2001) defined context in a manner that operationalizes the concept in terms of the entities and sources of information that generate "context". Their definition permits researchers to differentiate whether or not information can be considered "context". Given this operational definition, we are able to assess the impact of context upon formation of trust.

Dey, Abowd & Salber (2001) define context as: "Context is any information that can be used to characterize the situation of an entity. An entity is a person, place, or object that is considered relevant to the interaction between a user and an application, including the user and application themselves." Dey et al. (2001) argue that this definition of context facilitates formulation of different contexts for development of software applications. They maintain that information used to describe a personís situation when interacting with an application is context. Researchers have identified numerous categories of context (e.g., Dey, Abowd & Salber, 2001; Ryan, Pascoe & Morse, 1997; Schilit & Theimer, 1994). Dey et al. (2001), however, selected 4 categories of context that they view as most important when developing interactive applications. These four categories are: identity, location, status/activity and time. Identity denotes the unique assignment of an identifier (e.g., login and password) to an entity (e.g., customer). Location encompasses three-dimensional spatial position, orientation, and spatial relationship with other entities such as proximity and containment. Status/activity describes inherent features of the entity that can be perceived (e.g., temperature or lighting of a room, action of a human(s), upload time of a computer). Time characterizes a situation as to chronology, duration and frequency. Dey et al. (2001) also note that their four selected categories may afford inference or derivation of additional contextual information. For example, a personís phone number and address may be derived from a directory, by knowing their name.

Combining Trust and Context

Building consumer trust in online vendors is critical to the success of every online vendor.  We are engaged in a study that will allow us to explore the effects of building trust in e-commerce vendors when factoring context into the design and use of e-commerce sites. To this end, we have adopted McKnightís model of trust formation and the four categories of context identified by Dey et al. (2001) to develop an evaluation framework for assessing the impact of context upon the formation of trust. Specifically, McKnight et al.ís (2002) questionnaire solicits user responses concerning each of the critical trust formation antecedents. Differing context scenarios were developed to evaluate the impact of context upon user responses to the aforementioned questionnaires. A Likert scale was used to quantify questionnaire responses. Results of this study will be available for discussion at the workshop.

Conclusion

Building trust into the design of interfaces is amongst the primary objectives of future HCI innovations. Taking the complexities of context into account during the design and use of electronic commerce applications may have important implications for understanding how and why individuals trust the information that is given to them. Developers could potentially benefit from understanding how people interact with their surroundings. Applications in Ambient Societies that account for relevant context characteristics may benefit from increased functionality and productivity.

References

Bhattacherjee, A. (2002). Individual Trust in Online Firms: Scale Development and Initial Test. Journal of Management Information Systems, 19(1), 211-241.

Chen, S.C., & Dhillon, G.S. (2003). Interpreting Dimensions of Consumer Trust in E-Commerce. Information Technology and Management, 4, 303-318.

Dey,A.K., Abowd, G.D.,& Salber, D. (2001). A conceptual framework and a toolkit for supporting the rapid prototyping of context-aware applications. Human-Computer Interaction, 16, 97-166.

Dogac, A., Laleci, G.B., & Kabak, Y. (2003). A Context Framework for Ambient Intelligence. Technical Report, NO: EEEAG 102E035, Middle East Technical University, http://www.srdc.metu.edu.tr/webpage/publications/2003/context.pdf.

McKnight, D.H., Choudhury, V., & Kacmar, C. (2002). The impact of initial consumer trust on intentions to transact with a web site: a trust building model. Journal of Strategic Information System, 11, 297-323. 

Patrick, A.S. (2002) Building trustworthy software agents. IEEE Internet Computing, 6(6), 46-53.

Ryan, N., Pascoe, J., Morse, D. (1997). Enhanced Reality Fieldwork: the Context-Aware Archeological Assistant. In Gaffney, V., van Leusen, M., Exxon, S (eds.). Computer Applications in Archaeology.

Schilit, B., Theimer, M. (1994). Disseminating Active Map Information to mobile Hosts. IEEE Network, 8(5), 22-32.