One major obstacle to the widespread diffusion of e-commerce is consumer distrust.9 In the highly uncertain virtual environment, people have a more prevalent tendency to distrust than to trust, so as to avoid potentially negative consequences. Although mechanisms to promote trust have been extensively researched, it is not clear whether these recommended trust-building mechanisms would be equally effective in reducing distrust. More fundamentally, are trust and distrust merely opposite sides of the same coin, or are they in fact two independent concepts that can coexist and should be managed differently? In this article, we investigate this ambiguous phenomenon, which we characterize as the 'trust-distrust' paradox. The paradox raises the question of whether trust and distrust could indeed be two separate concepts which exist simultaneously in online consumers' minds. If trust and distrust are indeed two separate, simultaneously operating concepts, the mechanisms for building trust and eliminating distrust could be different: "it would be misleading to assume either that the positive predictors of trust would necessarily be negative predictors of distrust, or that the positive consequences of trust would necessarily be influenced negatively by increased distrust."7 To trust or to distrustthat is the question to consider: the paradox warrants a thorough examination.
We use two hypothetical Web site examples of Internet vendors: the provision of value-adding information and the assurance that all hyperlinks are valid, to illustrate the possibility that trust and distrust could be two separate operating concepts.
In the first example, a transactional Web site selling digital cameras provides value-adding tips and advice to help visitors enhance their photographic skills. The provision of value-adding information may be useful in building consumer trust,2 as it signals the knowledge and customer-oriented services that the Web site provides. However, does such value-adding information also help to reduce consumer distrust? Customers would probably not feel suspicious about a Web site that lacked information about photographic skills, since this is not a necessary requirement for buying digital cameras online. In the second example, when confronted with broken links in a transactional Web site (especially in the payment pages), online customers would be understandably wary and distrustful of the Web site; they might seek out other sites that do not have broken links.5 Yet it is unlikely that customer trust would be enhanced if a Web site had no dead links, since customers would typically take this for granted.
When surfing a Web site, consumers undergo many positive experiences (such as the first example above) and many negative experiences (such as the second example here), in which both trust and distrust could be reinforced. It appears that trust and distrust could simultaneously coexist and develop in different ways. Given the lack of a clear understanding and the dearth of empirical evidence as to whether trust and distrust are indeed separate constructs or not, we conducted a field study to further investigate the trust-distrust paradox in the online transaction context.
In this paper, trust is defined as the belief that an Internet shopper has in an Internet vendor and is willing to engage in an Internet shopping transaction, even with the possibility of loss (i.e. risk), based on the expectation that the vendor will engage in generally acceptable practices, and will be able to deliver the promised products or services.8 We adapt Lewicki and Tomlinson's definition of distrusta in this paper: the suspicion that the Internet vendor's motives, intentions and behaviors are sinister and harmful to the Internet shopper's interests. With distrust, an Internet shopper would fear a realization of loss if he or she is to engage in an Internet shopping transaction with the Internet vendor.
Based on a review of 39 studiesb of Web-based consumer trust/distrust, we summarize the mechanisms (see Table 1) that are potentially related to consumer trust and distrust. An assumption that underlies most of these studies is that trust and distrust are polar opposites of a single conceptual dimension. If this assumption is correct, then mechanisms that build trust should also exert the same influence on eliminating distrust.
Existing Web site interface studies also suggest that Web site features could be broadly categorized as either hygiene or motivating attributes.11 This categorization scheme helps to explain why trust and distrust deserve strategically different management. We consider hygiene attributes to be the minimal set of essential, function-related Web site attributes provided to perform an online transaction. Motivating attributes are the extra value-added Web site attributes provided to motivate surfing and buying. These two sets of attributes combine in consumers' aggregate, global evaluation of service performance. For example, specific design factors listed in Table 1 (such as security assurance and customization) are believed to influence global feelings about hygiene attributes (for example, whether the Web site has provided the basic necessary functions before they are willing to consider buying) and motivating elements (for example, whether the site has provided extra value-added design features to motivate surfing and buying). By including consumers' global evaluations of these two aspects of Web site performance in the research model, a more complete and generalizable picture of consumers' cognitive processes can be achieved. In order to verify that consumer trust and distrust have influencing power on their consequential outcomes, the extent that trust and distrust affect consumers' buying intentions is also worthy of investigation.
Following a preliminary confirmation of the above list of Web site design factors through interviews with a consumer panel, a field study was conducted in Hong Kong. Two retailing Web sites selling digital products, such as cameras, PDAs and MP3 players, were used to evaluate consumers' trust and distrust formation based on their first visiting experience. A total of 324 participants filled out an online questionnairec regarding their perceptions about these two Web sites immediately after visiting them. The purpose of this questionnaire is to test these propositions about the trust-distrust paradox, that is, their potential distinction and their possible determinants and outcomes.
Before testing the propositions, we evaluated the validity of the measurements. Each construct's item loadings are well above 0.80. The square root of the average variance extracted (AVE) figures are all above 0.80 and most of them are above 0.90, and the AVE for each construct is greater than the squared correlations between constructs.1,4 All these conditions provide support to indicate that constructs used in the research model had satisfactory convergent and discriminant validities. Furthermore, the result of the principal components factor analysis indicated that common method bias is not a problem with this data.6
Focusing on the trust and distrust measurements, the strong convergent and discriminant validities of trust and distrust lend empirical support to the conjecture that trust and distrust are two distinct and separate constructs. In addition, we contrast the combinations of high and low trust versus distrust, using an SPSS 11.5 cross-table procedure. Table 2 shows that 63.9% (207) of the total cases (324) fit in quadrants 2 and 3, which indicates the predominant existence of either trust or distrust. Interestingly, 36.1% (112) of the cases fit in quadrants 1 and 4, suggesting the coexistence of high trust and high distrust, as well as low trust and low distrust. This provides empirical support for the key proposition in this study that trust and distrust could simultaneously coexist in individuals.
The determinants and outcomes of trust and distrust are also examined in this study. The structural model analysis indicates that the consumers' evaluation of hygiene attributes forms a good predictor of distrust (b=-0.484, p<0.01); while consumers' evaluation of motivating attributes lead more to trust (b=0.571, p<0.01) than to distrust (see Table 3). Regarding the outcomes of trust and distrust, we found an asymmetric influence of trust (b=0.176, p<0.05) and distrust (b=-0.391, p<0.01) on shaping consumers' buying intentions. Distrust has a substantially greater impact on buying intention than trust. This verifies the differential roles that trust and distrust play in influencing behavior outcomes. In order to stimulate buying behavior among customers, Web sites need to pay significantly more attention to incorporating Web site features that alleviate distrust, thereby removing barriers to buying behaviors, rather than simply to incorporate trust-building features.
Accordingly, we also found that the set of Web site design attributes in Table 1 can be classified into four categorieshygiene attributes, motivating attributes, bivalent (hygiene and motivating) attributes and insignificant attributes, as shown in Table 3. Hygiene attributes (labeled with 'h' in Table 3) include Technical Functionality, Problem-Solving Orientation, Situational Normality, and Information Quality. Motivating attributes (labeled with 'm' in Table 3) include Willingness to Customize, Enjoyment Perception, and Knowledge and Skills Provided. Bivalent attributes (labeled with 'b' in Table 3) are those that exert bidirectional effects on the evaluations of hygiene and motivating attributes. They include Structural Assurance, Ease of Use, and Perceived Usefulness. Finally, a number of attributes (labeled with 'i' in Table 3) were found insignificant in this study, including Consumer Control, Overall Appearance, Consumer Feedback, Third-Party Recognition, and Perceived Completion Success. Figure 1 summarizes the data analysis and models the different determinants'd and outcomes of trust and distrust. The structural model analysis implies that an attribute (such as knowledge and skill, customization design, enjoyment design) that strongly promotes trust does not necessarily reduce distrust; likewise, an attribute that attenuates distrust (such as technical functionality, problem-solving orientation, and situation normality) may not be effective in improving trust.
This study suggests that current wisdom on Web site design that promotes trust needs rethinking so as to recognize the importance of alleviating consumer distrust. Our study suggests that trust and distrust are two distinct and separate constructs, not two ends of a single continuum, both are thus worthy of attention. Meanwhile, the basic set of Web site functionalities should no longer be considered as sufficient conditions to establish trust. Rather, the effects of the factors that can eliminate consumer distrust deserve more attention. In order to retain new comers on their first visit, or for new Web sites, Internet vendors must design their Web-stores in a manner that matches consumers' mental models of a "typical Web site". It is also crucial to provide consumers with more relevant and updated information. If such basic requirements are not met, consumers are likely to assume that the Web site does not function well, leading to less interaction activity. Especially for first-time visitors or new Web sites, these are considered to be essential components of a workable transactional Web site (see Table 4). Thus, designers should be aware of the hygiene role of situational normality and information quality, deploying them as a necessary prerequisite for other design factors.
However, it is important to stress that after maintaining the above hygiene attributes, more resources should be dedicated to the other three aspects, namely customization, enjoyment design, and knowledge and skills provision. This is because these three aspects add extra value to consumers, engender trust, and directly favor the execution of transactions. As online consumers become more demanding for customized services and products, it is worth informing them that firms can offer individually customized products, using Internet-integrated manufacturing and distribution technology. At the same time, incorporating appropriate multimedia and 3D presentations inspires consumers' intrinsic interest in a Web site and consequently motivates purchases. Furthermore, providing consumers with value-adding expertise in the form of knowledge and skills differentiates the vendor from its competitors. This can be achieved by offering professional advice that improves consumers' skills in using the products. The findings from this study indicate that such designs will build customer trust and directly favor selling online. For repeat visitors who may have taken hygiene features for granted, more motivating features, such as customization, enjoyment, and knowledge and skills provision, will increase the likelihood that they will develop a long-term relationship with the Web site.
Two surprising results were the non-significance of consumer feedback mechanisms and third-party recognition, contradicting prior trust research.8 Our empirical findings indicate that managerial efforts invested in these two features did not invoke perceptions either of a hazard-free environment or of additional value for consumers. This could be because consumers in this study view these features are marketing tools by the Internet vendors. It could also be because Web users rely more on their own experience to judge sites rather than other external cues that signal a firm's trustworthiness or untrustworthiness. It is also understandable that at first glance, consumer feedback mechanisms and third-party recognition would both work. However, once consumers gain surfing experience, they could place less emphasis on such external signals about the company, choosing instead to rely on their practical understanding (experience). Therefore, in designing websites, designers should actively evaluate the usefulness of these two kinds of Web site features with respect to their effectiveness on the targeting consumers. Nevertheless, consistent with our previous suggestions, firms should always take experiential design such as enjoyment, customization, and knowledge and skills provided into account. The above design features reward the firms with a competitive advantage in the increasingly competitive Web environment.
Additional findings from this study are the strong relationship between the evaluation of motivating attributes and consumers' buying intention. Although it is not the planned focus of this study, the evaluation of motivating attributes might act as an "impulse enactment". The provision of enjoyment, new knowledge and skills, and customization can make consumers feel more motivated, which can trigger a buying intention. This additional finding can be applied to research about Web site design for online impulse buying.
8. Lim, K.H., Sia, C., L, Lee, M. K. O., and Benbasat, I. How do I trust you online, and if so, will I buy?: An Empirical Study of Two Trust Building Strategies. Journal of Management Information Systems 23, 2, (2006) 233266.
9. McKnight, D.H. and Chervany, N. L. Trust and distrust definitions: One bite at a time, in Trust in Cyber-Societies: Integrating the Human and Artificial Perspectives. R. Falcone, M. Singh, and Y.H. Tan, Eds. 2001, Springer, Berlin. 2754.
10. Podsakoff, P.M., MacKenzie, S. B., Lee, J. Y., and Podsakoff, N. P. Common method biases in behavioral research: A critical review of the literature and recommended remedies. Journal of Applied Psychology 88, 5, (2003) 879903.
b. Studies were published in: Communications of the ACM, MIS Quarterly, Information Systems Research, Journal of Management Information Systems, International Journal of Electronic Commerce, Information & Management, and Decision Support Systems
d. Causal relationships could exist between individual determinants. For instance, Overall Appearance could have impacts on Ease of Use, Enjoyment Perception and Situational Normality. As this is an exploratory research that seeks to focus on whether trust and distrust are indeed separate concepts, and that aims to segregate website features that influence these perceptions, we did not explicitly model their causal relationships. Nevertheless, this is a very interesting issue that could provide significant insights on the formation of trust and distrust, and would be investigated in our future research.
The work described in this paper was partially supported by research grants from City University of Hong Kong (9040956) and Research Grants Council of Hong Kong (CityU 1303/04H).
©2009 ACM 0001-0782/09/0500 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2009 ACM, Inc.
No entries found