The Mediating Role of Consumer Trust in an Online Merchant in Predicting Purchase Intention

International Journal of Information Management 33 (2013) 927–939

Contents lists available at ScienceDirect

International Journal of Information Management journal homepage: www.elsevier.com/locate/ijinfomgt

The mediating role of consumer trust in an online merchant in predicting purchase intention Ilyoo B. Hong ∗ , Hoon S. Cha College of Business and Economics, Chung-Ang University, Bubhakkwan Bldg., Rm. 1404, 221 Heuksuk-dong, Dongjak-ku, Seoul 156-756, Republic of Korea

a r t i c l e

i n f o

Article history: Available online 26 September 2013 Keywords: Online shopping e-Commerce Perceived risk Trust Purchase intention

a b s t r a c t It is widely known in related literature that trust in a merchant reduces the perceived risk of an online transaction. However, there are theoretical reasons to postulate that the perceived risk acts as a barrier to consumer trust. Furthermore, existing studies suggest that trust is an important predictor of purchase intention. Thus, this research aims at investigating the mediating role of consumer trust in an online merchant in the relationships between components of perceived risk and purchase intention: (1) examining the total effect without mediation, and (2) examining the mediation effect. When we probed the total effect, the findings revealed that performance, psychological, financial, and online payment risks have a significant negative influence on purchase intention. On the other hand, an examination of the mediation effect indicated that trust in an online merchant completely mediates the effect of performance risk, but partially mediates that of the psychological risk. Given the mixture of unmediated as well as mediated effect of perceived risks on purchase intention, the paper concludes that efforts, made by online merchants, to lessen certain types of risk will first improve consumer trust, and then ultimately, increase consumer’s intention to buy online. © 2013 Elsevier Ltd. All rights reserved.

1. Introduction Internet-based commerce has undergone explosive growth over the past decade as consumers today find it more economical as well as more convenient to shop online. Nevertheless, the shift in the common mode of shopping from offline to online commerce has caused consumers to have worries over issues, such as private information leakage, online fraud, discrepancy in product quality and grade, unsuccessful delivery, and so forth. Unfortunately, there has been a steady increase in the number of incidents that cause consumers to have such worries; for example, the Internet Crime Complaint Center reports that it has received 303,809 Internet fraud complaints in 2010, up from 95,064 complaints in 2003 (Center, 2011a). Meanwhile, the number of privacy breaches reported in the U.S. has increased from 157 in 2005 to 662 in 2010 (Center, 2011b). Therefore, today’s consumers feel unsafe about making purchases online. The concerns that consumers have over online buying are collectively termed as consumers’ perceived risk. Numerous studies have been conducted to examine the role of perceived risk as a chief barrier to online purchases and to understand the theoretical relationships among perceived risk, trust, and purchase intentions. However, most studies (for example, Cheung

∗ Corresponding author. Tel.: +82 2 820 5549; fax: +82 2 813 8910. E-mail address: [email protected] (I.B. Hong). 0268-4012/$ – see front matter © 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.ijinfomgt.2013.08.007

& Lee, 2001; Corbitt, Thanasankit, & Yi, 2003; Flavian, Guinaliu, & Gurrea, 2005; Gefen, 2002; Gefen, Karahanna, & Straub, 2003; Jarvenpaa, Tractinsky, & Vitale, 2000; Pavlou, 2003; Salam, Iyer, Palvia, & Singh, 2005) focus on empirically investigating the effects of trust on perceived risk with little attention devoted to the effects of perceived risk on trust. While the influence of trust on perceived risk is worth studying, the influence in the opposite direction is equally important, enabling insights into the potential of perceived risk as an inhibitor of trust. For example, a consumer who perceives huge risk concerning an online transaction is likely to foresee a great potential of loss and thus, places little trust in the merchant. According to Pavlou (2003), the primary source of the perceived risk is either the technological uncertainty of the Internet environment or the behavioral uncertainty of the transaction partner. Due to such types of uncertainty, the increase in worries over the perceived risk may negatively affect trust. For example, if a consumer who sends sensitive transaction data over the Internet is concerned that his or her private information may leak out due to a lack of security, trust may decrease (Olivero & Lunt, 2004). By the same token, if the consumer feels that the online merchant has the potential to profit by behaving in an opportunistic manner by taking advantage of the remote, impersonal nature of online commerce, then it is unlikely that the merchant will be trusted. That is, the more likely it is for a danger to occur, the lesser is the trust and the greater is the need to control the transaction (Olivero & Lunt, 2004). Thus, the related studies as a whole indicate that while some researchers noted the influence of the overall perceived risk on the trust level,

928

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

not much attention has been given to the effects of different types of perceived risk. Meanwhile, the related literature suggests that consumer trust in an online merchant is a key predictor of purchase intention (Hong & Cho, 2011; Pavlou, 2003; Pavlou & Gefen, 2004; Verhagen, Meents, & Tan, 2006). Then, we are led to believe that the relationship between perceived risk and purchase intention can be indirect as well as direct; moreover, it can be mediated by consumer trust. However, little attention has been given to this research issue to date. The present research is a step toward closing that gap in extant research. It aims at addressing the need to study the intriguing relationships between perceived risk and purchase intention in an e-commerce setting. To accomplish the research purpose, we established two research questions. First, does perceived risk act as an inhibitor of purchase intention? Second, does trust in an online merchant mediate the relationship between perceived risk and purchase intention? We classified perceived risk into six different types based on literature, and empirically analyzed both the direct and the mediation effects of each dimension of perceived risk upon purchase intention. The contribution of our research has both theoretical and practical dimensions. Theoretically, it will contribute to the existing body of knowledge by providing new insights into the mediating role of consumer trust in an online merchant in the relationships between dimensions of perceived risk and purchase intention. Practically, the research will help e-businesses develop strategies to reduce the specific types of perceived risk found to negatively influence trust, thereby engendering consumer trust in an online merchant and ultimately increasing online sales.

types of risk including financial, performance, psychological, and social risks. Meanwhile, risks faced by online consumers are those engendered by the Internet as a sales channel, in addition to the traditional consumer risks. The use of the Internet, as a mode of purchase, creates risks for online transactions with the merchant since transactions are remote, involving no face-to-face contact between the merchant and the consumer (Cases, 2002). For example, Internet-based shopping requires a delivery process, unless an order is placed for a digital product that can be delivered online via the Internet; therefore, there is a risk for inconsistency between the ordered product and the delivered product (Ward & Lee, 2000). In addition, consumers may perceive a payment risk because they are likely to pay by a credit card, and thus, important personal information needs to be transmitted when the payment transaction is executed. Although security measures, such as encryption and authentication, are in place, consumers feel insecure about the possibility of personal information leakage that may result from hacking during the course of an online transaction. Jarvenpaa and Todd (1997) pointed out personal and privacy risks as well as economic, social, and performance risks in Internet-based transactions. Personal (or payment) risk refers to the fear of giving one’s creditcard number online, and privacy risk is associated with the buyer’s fear that personal information will be collected without authorization. Based on the above theoretical evidence, it is inferred that Internet transactions can introduce delivery and payment risks in addition to the common risks inherent in traditional commerce. Additionally, we propose an integrative model of risk dimensions: performance, psychological, social, financial, online payment, and delivery risks.

2. Literature review

2.2. Trust

2.1. Perceived risk

Trust has been widely studied over the years, as it is recognized as a key element in relationships between individuals, between organizations, and between an individual and an organization. Nevertheless, trust is perhaps one of the most highly challenging notions in which concepts are hardly agreed upon by researchers (Hong & Cho, 2011). As Lee and Turban (2001) noted, trust has been examined in various contexts including buyer–seller relationships, strategic alliances, and labor–management negotiations. In general, trust is defined as the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party (Mayer, Davis, & Schoorman, 1995). Morgan and Hunt (1994) defined trust as the belief that the trustee will behave in a favorable manner. Further, they state that trust is critical in successful alliances between firms. As such, trust refers to believing that the trustee will not do harm to the trustor and that negative consequences will not occur. In the context of electronic commerce, trust becomes an even more important issue since exchange relationships are based on the impersonal nature of the Internet infrastructure. In particular, consumers face the challenge of buying a product or service online from an unfamiliar merchant; moreover, they cannot actually see or touch the product. Trust plays a central role in helping consumers overcome the perceptions of risk and insecurity (McKnight, Choudhury, & Kacmar, 2002). Since privacy and security concerns are major barriers to the Internet channel, without trust, customers will not give vendors their personal information, including credit card information (Hoffman, Novak, & Peralta, 1999). Therefore, online trust is formed slowly over time as a consumer gains experience through repeated transactions (Cheskin-Research, 1999). For the purpose of the present research, trust is defined as the consumer’s belief that the online merchant will not behave in an

Bauer (1960) proposed that consumer behavior could be viewed as an instance of risk taking. He maintained that consumer behavior involves risk in the sense that any action of a consumer will produce consequences that one cannot anticipate and of which at least some are likely to be unpleasant. An individual perceives a situation as bearing risk if entering this situation might lead to negative consequences, and also if the individual is not able to control the occurrence of these consequences (Koller, 1988). Thus, the more negative are the consequences and the less the individual can control the consequences, the higher is the level of the perceived risk. Bauer (1960) emphasized that it is not a “real world” (or objective) risk but a perceived (or subjective) risk that influences consumer behavior. In the context of electronic commerce, Cox and Rich (1964) defined “perceived risk” as the nature and amount of risk perceived by a consumer in contemplating a particular purchase decision. A consumer perceives risk because prior to making a purchase, she cannot always be certain that the planned purchase will allow her to achieve her goals of purchasing. The uncertainty perceived by the consumer with regard to the choice of a product, brand, retailer, or channel determines the nature of the risk. Meanwhile, the amount of risk perceived by the consumer is a function of two general factors: the amount at stake in the purchase decision, and the individual’s feeling of a subjective certainty that she will “win” or “lose” all or some of the amount at stake (Cox & Rich, 1964). The risks perceived by consumers in traditional commerce are classified from various perspectives in the literature. While they each exhibit unique classification schemes, these studies (for example, Jacoby and Kaplan, 1972; Kurtz & Clow, 1997; Peter & Ryan, 1976; Schiffman & Kanuk, 1994; Stone & Gronhaug, 1993; Taylor, 1974; Zikmund & Scott, 1977) have focused on four essential

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

opportunistic manner and that the e-commerce environment is secure enough to provide risk-free transactions. 2.3. Perceived risk and trust Trust and perceived risk are in a very close and inseparable relationship. Most existing studies predominantly focus on the effects of trust on the perceived risk (Cheung & Lee, 2001; Corbitt et al., 2003; Flavian et al., 2005; Gefen, 2002; Gefen et al., 2003; Jarvenpaa et al., 2000; Pavlou, 2003; Salam et al., 2005). For example, Pavlou (2003) and Jarvenpaa et al. (2000) reported that an increase in consumer trust in an online merchant lessens the perceived risk. However, Mayer et al. (1995, p. 711) noted that “it is unclear whether risk is an antecedent to trust, is trust, or is an outcome of trust.” Johnson-George and Swap (1982) state that the “willingness to take risks may be one of the few characteristics common to all trust situations.” The presence of trust implies the acceptance of a certain degree of risk toward the loss when the expected outcome is positive. While the trustor chooses to trust if the risk that she has to take is within an acceptable range, the trustor has no choice but to give up making the trusting choice in case the risk is likely to go beyond the limit. Therefore, the perceived risk can be an important predictor of the trusting decision. Further, Deutsch (1973) postulated that a trusting choice will be made if the subjective probability of an event of positive valence is higher than the subjective probability of an event of negative valence. That is, a trustor will not choose to trust in case risks are expected to be greater than benefits. This theory is applied to the electronic commerce setting. If a consumer associates high risk with an online transaction, then the level of trust in the online merchant decreases and the need to control the transaction increases (Olivero & Lunt, 2004). Studies of perceived risk suggest that a key source of perceived risk is uncertainty. Ring and Van de Ven (1994) found that the risks inherent in a transaction increase in proportion to reductions in time, information, or controllability. Thus, under circumstances where there are time pressures or a lack of information or difficulties in controlling the trustee’s behavior, uncertainty will be present. Pavlou (2003) suggested that risks in electronic commerce are introduced by both the impersonal nature of the online environment and the uncertainty of using the Internet for transactions. Then, such uncertainty has two components: behavioral uncertainty from the transaction partner and environmental uncertainty from the technical environment of online transactions (Pavlou, 2003; Ring & Van de Ven, 1994). Behavioral uncertainty exists because the Web vendor has the chance to behave in an opportunistic manner, whereas environmental uncertainty exists due to the unpredictable characteristics of the Internet infrastructure. When behavioral uncertainty is high, a consumer is likely to feel that the transaction partner may potentially bring about a loss upon him or her by taking advantage of the remote, impersonal nature of the online transaction. On the other hand, in situations where environmental uncertainty is high, the consumer is most likely to fear that an unauthorized person may take his or her personal information, even if the merchant’s server is equipped with protective technologies such as encryption. Some empirical studies (for example, Jarvenpaa & Leidner, 1999; Pavlou, 2003) found that perceived risk has a direct negative influence on transaction intentions. They suggest that consumers perceiving a great risk are motivated to avoid engagement in the transaction since they are not sure they can expect a positive payoff. However, in the present research, we will examine the potential role of trust as a mediator between risk and transaction intentions. It is important to note that a consumer may not wish to participate in an online transaction because s/he is not quite sure that the online merchant will act favorably in the interest of the consumer,

929

not merely because a risk is present. That is, the consumer may choose not to shop online because the transaction partner cannot be reasonably trusted. Since a consumer as a whole cannot accurately predict the likelihood that the partner will behave in an opportunistic manner and thus, can only guess the degree of risk under uncertainty, the actual risk perception will be developed based on the exposure to media concerning related incidents or on past shopping experiences. Moreover, if the risk perceived over time goes beyond the level that s/he can tolerate, then the consumer may choose to abandon the trusting choice. Therefore, we will develop a research model centered on the role of trust as a variable that mediates the relationship between perceived risk and purchase intention. 3. Conceptual model and hypotheses 3.1. Conceptual model development The purpose of this paper is to examine the relationships between dimensions of perceived risk and purchase intention. In particular, we will explore the mediating role of consumer trust in such relationships. Earlier in the literature review, we provided the theoretical grounds for the impact that perceived risk has on trust. A close examination of the relationship between risk and trust indicates that the influence of risk is valid for only some specific types of risk rather than the overall perceived risk. For example, if a consumer cannot trust the merchant because s/he may behave in an opportunistic manner, then what makes the consumer unable to trust the merchant is most likely to be either a financial or a performance risk. Moreover, it would be possible that certain types of risk may have a direct influence on purchase intention without the mediating role of consumer trust (Jarvenpaa & Leidner, 1999; Pavlou, 2003). For that reason, in order to correctly understand the causal relationship between perceived risk and trust, we need to focus on the types of perceived risk as independent variables and their differential impacts on trust through an empirical analysis. As we observed in the literature review, the related studies suggest that the types of risk perceived by a consumer in an electronic commerce setting include performance, psychological, social, financial, online payment, and delivery risks. Thus, we will use this taxonomy in order to classify the perceived risk in our research. Our conceptual model is presented in Fig. 1. The first six hypotheses, namely H1-1, H1-2, H1-3, H1-4, H1-5, and H1-6, focus on the “total” effects of the dimensions of perceived risk on purchase intention. The next six hypotheses, namely, H2-1, H2-2, H2-3, H2-4, H2-5, and H2-6, are designed to explore the mediation effects of consumer trust in the relationships between the individual types of perceived risk and purchase intention. We will look at the theoretical background for each of the hypotheses in the following subsection. 3.2. Hypothesis development Fig. 2 describes the unmediated and mediated models following Baron and Kenny’s notation (Baron & Kenny, 1986; Frazier, Tix, & Barron, 2004; Shrout & Bolger, 2002), where the types of risk are not specified for simplicity. In the unmediated model as shown in the figure, we posited that perceived risk negatively influence purchase intention. Path c in this model is called the total effect. On the other hand, in the mediated model also shown in the figure, we hypothesized that the effect of perceived risk on purchase intention was mediated by trust. Path c is called the direct effect, while paths a and b are called the indirect effect. If perceived risk no longer directly affects purchase intention (i.e., path c = 0) after trust has been controlled, complete mediation exists. When path c

930

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

Perceived risk

Performance risk

H2-1

Trust

H2-2 H2-3 H2-4

Psychological risk H1-1

Social risk

H2-5 H2-6

H2-1,2,3,4,5,6

H1-2 H1-3

Financial risk

H1-4

Purchase intention

H1-5

Online payment risk H1-6 Control variables: • • • •

Delivery risk

Age Gender Internet usage Internet shopping frequency

Fig. 1. The conceptual model.

is reduced in absolute size but is greater than zero, partial mediation exists. As we can see in Fig. 1, we established one hypothesis for each type of risk in the unmediated model for a total of six hypotheses, whereas we formulated six additional hypotheses in order to examine the mediating role of consumer trust in the relationships between each dimension of perceived risk and purchase intention. 3.2.1. Examining the total effect: Perceived risk and purchase intention Related studies have in general found a negative relationship between the overall perceived risk and purchase intention. For example, according to Jarvenpaa et al. (2000), the theory of planned behavior predicts that a consumer is likely to buy from an online store, which is perceived to be low in risk, although the consumer’s attitudes toward the merchant are not positive. In the context of Internet shopping, perceived risk may reduce consumer’s perception of behavioral control that refers to the extent to which a consumer feels that engaging in a behavior is completely up to him or her (Jarvenpaa et al., 2000, p. 50). Pavlou (2003) also suggests that perceived risk is negatively related to purchase intention. He suggests that transaction intentions are influenced by beliefs about online retailers that are partly determined by the behavioral and environmental factors that may lead to risk perceptions. Given that losses are likely, a consumer will have no reason to engage in a transaction. The negative relationship found by the existing related studies between perceived risk and intention to buy is likely to hold true for the individual dimensions of perceived risk, although the differential impact of each dimension of risk may vary with product classes or consumers.

Unmediated Model

First, compared to perceived risk in traditional shopping, the risk associated with the product performance in online shopping is especially significant because of consumers’ limited ability to communicate through the Internet and to accurately judge the quality of the product. For example, when consumers have difficulty grasping the features of products such as clothes, shoes, and furniture solely from Website pictures, they could be easily concerned that the product ordered might not be exactly as it appeared on the Website or might not perform up to their expectations (Hassan, Kunz, Pearson, & Mohamed, 2006). Indeed, many online stores have witnessed the negative impact of performance risk perceptions on actual sales and thus have been trying to come up with various mechanisms to lower consumers’ perceptions of performance risk. For instance, instead of simply displaying pictures for the product features, some online stores host a discussion forum to allow consumers to freely exchange their comments, opinions, or recommendations about the products (Garbarino & Strahilevitz, 2004), which provide useful purchase guidelines for online consumers. Given the discussion above, we propose the following hypothesis: H1-1. tion.

Performance risk is negatively related to purchase inten-

Second, an online consumer could experience psychological discomfort due to personal ego in making purchase decision (Jacoby & Kaplan, 1972). This type of psychological loss may result from consumers’ lack of experience in buying products or services. In general, consumers with less online shopping experience may feel more mental discomfort from potentially making the wrong product choice than those with more experience of shopping online.

c

Perceived risk

Purchase intention

Total effect

Trust b

a

Indirect effect

Indirect effect

Mediated Model

Perceived risk

c’ Direct effect

Fig. 2. The total effect vs. direct effect vs. indirect effect.

Purchase intention

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

For example, consumers with prior experience would feel less concerned as they know how to choose products best aligned with their expectations and how to return products that they do not like. Therefore, as consumers perceive more psychological risk, they may experience greater level of anxieties and be less willing to buy online. Therefore, we propose the following hypothesis: H1-2. tion.

Psychological risk is negatively related to purchase inten-

Third, Hassan et al. (2006) states that online shoppers are concerned about the reaction of others who think of the online prospective shopper as being foolish or showy. Cases (2002) also defines social risk as the fear of the reaction of friends and family concerning the Internet as a mode of purchase. The rise of Internet has increased the convenience of shopping in any place and at any time; however, at the same time, consumers can get caught up by the sheer variety of items and an illusion that they have not spent too much money. As the online shopping addiction and similar compulsive online behaviors on the Internet become important social issues and problems, online consumers become more afraid of their acquaintances’ view about online shopping. When consumers’ perceived benefits of online shopping are outweighed by perceived social risk, the purchase will likely be avoided. Therefore, the following hypothesis can be formulated: H1-3.

Social risk is negatively related to purchase intention.

Fourth, financial risk is defined as the probability of monetary loss associated with purchasing a product. Thus, in the online environment, purchasing by consumers has been dominated by products that carry lower levels of financial risk such as books, clothes, and music files. Although the online purchase has been gradually expanding its area over more expensive products, like a laptop computer and even an automobile, many online consumers still perceive relatively high financial risk with those expensive products. Thus, consumers may be more hesitant when purchasing a product or a service likely to have potentially high economic loss. Based on the above grounds, we suggest the following hypothesis: H1-4.

Financial risk is negatively related to purchase intention.

Fifth, a risk dimension that can become a key consideration in online shopping is the perceived risk associated with online payment. Various surveys have shown that Internet users are increasingly concerned about the possibility that their private and credit card information may be captured, collected, and misused by a hacker or even by online marketers without permission. These concerns will cause the consumer to look for an alternative mode of shopping (e.g., making purchases at a department store). As a result, we propose the following hypotheses. H1-5. Online payment risk is negatively related to purchase intention. Finally, when purchasing online, a consumer needs to wait for an order to arrive. The shipment containing the ordered product could be lost or delivered to a wrong address if there is a lack of business experience on the part of the delivery company. In addition, it is possible that the order arrives later than expected, provided that there is a backorder on the ordered product (Cases, 2002). A consumer who has strong perception of delivery risk will most likely lose interest in the online purchase. Therefore, we propose the following hypothesis: H1-6.

Delivery risk is negatively related to purchase intention.

3.2.2. Examining the mediation effect: Perceived risk, trust, and purchase intention Prior studies on the relationship between risk and trust focus on examining the causal relationship between the two constructs

931

where trust is viewed as an antecedent of the perceived risk. However, as we have seen in the literature review, the risks that result from behavioral and environmental uncertainties are what make an online merchant untrustworthy to consumers. Behavioral uncertainty is associated with consumers’ concerns that the online vendor may not behave in a socially responsible manner based on opportunistic calculation. Environmental uncertainty has to do with the possibility that consumers’ private information may leak out as transactional data are transmitted over the Internet (Pavlou, 2003). Pavlou (2003) suggested that behavioral uncertainty may lead to economic risk (i.e., financial loss), personal risk (i.e., the likelihood that the consumer may be a victim due to the use of unsafe products or services), seller risk (i.e., negative consequences that may result because of the inability to monitor the seller’s transactions), and online payment risk (i.e., the danger that exists because the consumer’s private information is given to a third party). On the other hand, environmental uncertainty that surrounds the online transactional infrastructure may result in economic risk (i.e., concerns over financial loss) and privacy risk (i.e., the likelihood that the consumer’s private information may be stolen or illegally disclosed). Likewise, when a consumer perceives risks due to the uncertainty associated with the transaction partner or with the online transactional infrastructure, the consumer would find it difficult to trust the transaction mechanism and furthermore, to participate in the online transaction. This is particularly true when we consider Mayer et al.’s (1995) definition of trust as “the willingness of a party to be vulnerable to the actions of another party based on the expectation that the other will perform a particular action important to the trustor, irrespective of the ability to monitor or control that other party.” While a consumer can trust the other party despite the presence of some degree of risk, once the amount and probability of the risk goes beyond an acceptable range, the consumer is most likely to give up that trust (Gefen, 2000). Meanwhile, the causal relationship between trust and purchase intention has been also noted by researchers. Jarvenpaa et al. (2000) applied the theory of reasoned action (TRA) to Webbased shopping, and concluded that a consumer’s online purchase intentions are influenced by attitude, and attitude is affected by consumer trust. Heijden, Verhagen, and Creemers (2003) conducted an empirical study based on TRA, and reported a similar finding; trust has an indirect effect on transaction intentions through the attitude as a mediator. In online commerce, trust in a transaction partner represents behavioral beliefs about the partner, and these beliefs can change the consumer’s behavioral intentions for online transactions. However, the majority of other related studies (for example, Gefen et al., 2003; Salam et al., 2005) provide contrary research findings indicating that trust has a direct impact on purchase intentions. To cite one example, Shankar, Urban, and Sultan (2002) found that online trust has a significant influence on purchase intention and customer loyalty, and confirmed a direct relationship between trust and purchase intention. The above line of reasoning leads us to believe that trust is most likely to play a mediating role between perceived risk and purchase intention. At the level of individual components of perceived risk, the mediation effect is likely to hold true. That is, the relationship between each type of perceived risk and purchase intention is likely to be indirect and to be mediated by consumer trust in an online merchant. Each of these risk dimensions will act as an antecedent to trust that in turn will become an antecedent to purchase intention. First, strong perception of product performance risk will result in minimal trust in the online merchant. For example, an online consumer considering purchasing such products as fresh fruits, a fabric detergent, or a computer often takes precautions to ensure that the product under consideration for online purchase meets his or her performance expectations. Are the fruits really fresh? Will

932

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

the fabric detergent surely function to remove all the dirts? Will the computer run fast and store large amounts of data? If the consumer has some doubts concerning these performance questions, he will put little trust in the merchant. And low consumer trust is likely to lead to little or no intention to make a purchase. Based on this line of reasoning, we propose the following hypothesis: H2-1. Trust in an online merchant mediates the relationship between performance risk and purchase intention. Second, a consumer who is much worried about the potential psychological discomfort that may result from the incorrect choice of a product will first lower her trust in the online merchant, which will, in turn, act to negatively influence her intention to buy online. For example, an online consumer considering the purchase of a fashion clothing product may be most likely serious about the visual match between her image and the product. Since the online store does not provide a way of fitting the clothing on, the consumer will have intense psychological concerns. Under this circumstance, she will begin to distrust the online transaction system as far as the fit of the clothing is concerned, and will probably choose not to buy online. Thus, we propose the following hypothesis: H2-2. Trust in an online merchant mediates the relationship between psychological risk and purchase intention. Third, if a consumer is much concerned that his acquaintances may feel that the purchased product does not appear suitable for her, then she will be reluctant to put trust in the online store. Typical examples of products of this category include laptops, smartphones, wrist watches, and a mink robe, which all tend to easily grab the attention of fellow workers, friends, and family members. Although the consumer perceives no psychological risk (i.e., there is a good fit between her self-image and the product), if she feels that other people is likely to find it unsuitable for her, then the social risk may grow large. A substantial amount of social risk will render the online merchant less trustworthy. As a result, the consumer will no longer trust the online purchase system, thereby abandoning the purchase intention. Therefore, we propose the following hypothesis: H2-3. Trust in an online merchant mediates the relationship between social risk and purchase intention. Fourth, given that a consumer has unusual worries over big financial loss associated with the opportunistic behavior of the seller, his trust in that seller will diminish. For example, a computer user considering the online purchase of a $ 4000 Apple Mac Pro on eBay may be concerned about the possibility that the unknown eBay seller might take his payment without shipping the ordered product. If the records show that the seller has a minimal number of positive ratings on eBay, then the risk perception for that seller will be quite strong. As a result, the consumer will have no intention to buy online from that merchant. Therefore, we propose the following hypothesis: H2-4. Trust in an online merchant mediates the relationship between financial risk and purchase intention. Fifth, provided that a consumer is unusually concerned that his private and financial information may leak out due to the possibility of a hacking incident, he is likely to lose trust in the online environment, in which case there will be no further desire to buy. For example, when customers find that an online shopping site does not provide minimum security protection (e.g., security protocol, keyboard encrypting, electronic certificates, etc.), they are likely to doubt the reliability of the Website and even to choose not to buy online. Thus, we propose the following hypothesis: H2-5. Trust in an online merchant mediates the relationship between online payment risk and purchase intention.

Finally, an online shopper who experienced several incidents of wrong delivery and thus perceives strong delivery risk will no longer trust the online merchant, and probably intend not to buy. It is also applicable when the online stores outsource their delivery process to the third party service providers. We can take Amazon.com and eBay.com for example. Amazon uses UPS as their major delivery service provider, but sellers in eBay often ship via smaller, less reliable delivery companies. With the higher level of perceived delivery risk, consumers may put less trust in the online store, and thus, may look for other alternatives to buy online from. Based on the theoretical grounds, we propose the following hypothesis: H2-6. Trust in an online merchant mediates the relationship between delivery risk and purchase intention. Overall, whether a consumer considering an online purchase perceives risk with regards to product performance, psychological/social damage, monetary loss, online payment, or delivery, that dimension of risk intensely perceived by the consumer will first lower consumer trust in a merchant, thereby eventually making the consumer reluctant to buy online from that merchant. 4. Research methodology To test the research model, we employed an empirical study using data from online survey responses. The survey participants were undergraduate students at a large university who voluntarily participated in the survey for extra credit. Although student participants may not fully represent the online shopper population, many previous studies (Bhatnagar, Misra, & Rao, 2000; Featherman & Pavlou, 2003; Gefen, 2000; Jarvenpaa et al., 2000; Jarvenpaa & Tractinsky, 1999; Lee & Turban, 2001; Pavlou, 2003) showed that college students are a good surrogate for online consumers. Indeed, the data collected in the present study also indicated that the participants are active online consumers: over 90% of the respondents reported that had shopped online least once in the last six months. 4.1. Measures A survey questionnaire was designed to measure the research constructs under consideration in this study. Above all, the perceived risk construct was not measured as an overall perceived risk, but as individual dimensions or components of the perceived risk. We considered a total of six types of perceived risk: performance, psychological, social, financial, online payment, and delivery risk (Cases, 2002; Jacoby & Kaplan, 1972). Performance risk was defined as the likelihood of problems associated with purchasing unfamiliar brands or defective products. Psychological risk was defined as the likelihood of an insufficient fit between the purchased product and the consumer’s self-image or self-concept. Social risk was defined as the likelihood of the purchased product influencing others’ view of the consumer. Financial risk was defined as the likelihood of some financial loss resulting from overpriced products, online fraud, or from unexpected expenses (e.g., a 15% restocking fee). Online payment risk refers to the likelihood that a consumer’s private information, including personal and credit card information, may be exposed to potential threats, and that such private information may be misused. Finally, delivery risk was defined as the likelihood of a delivery problem (e.g., late delivery of products, delivery to a wrong address, and delivery of a wrong product). Each type of risk was measured using three item variables. Many of these measurement items were adapted from existing consumer behavior and e-commerce literature (Featherman & Pavlou, 2003; Jarvenpaa & Todd, 1997; Pavlou, 2003; Schiffman & Kanuk, 1994; Stone & Gronhaug, 1993).

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

933

Table 1 Profile of the respondents (n = 206). Attribute

Value

Frequency

Percentage (%)

Gender

Male Female 20s 30s Never Less than once every six months At least once every six months At least once every three months At least once a month At least once a week Less than 10 h Between 10 h and 30 h More than 30 h

141 65 202 4 4 20 17 63 83 19 73 111 22

68.4 31.6 98.0 2.0 9.2 40.3 30.6 8.3 9.7 1.9 35.4 53.9 10.7

Age Internet shopping frequency

Weekly Internet usage

On the other hand, trust was defined as the extent to which a consumer believes that the merchant will behave in the interest of the consumer in purchasing a product online. Three items were used to measure this construct, drawn from Pavlou (2003) and Hong and Cho (2011). Finally, purchase intention was defined as the consumer’s inclination to purchase online. This construct was measured by three items designed to rate the extent to which a respondent chooses to buy from an online merchant and recommend the merchant to acquaintances. The items used were drawn from Jarvenpaa et al. (2000). All the above five-point Likert scales ranged from 1 strongly disagree to 5 strongly agree. The resulting twenty-four (24) items – and the list of the survey questions making up each measure – are summarized in Table A1. 4.2. Survey procedure and data analysis Student participants were asked not only about their overall shopping experience but also about their risk perceptions, trusting beliefs, and purchase intention with regards to a popular online retailer that sells a wide variety of products to consumers. While we provided the store name “Interpark.com” in the questionnaire as a representative online storefront, students were instructed to consider other familiar storefronts (e.g., Lotte.com or Samsungmall.com) as well in answering the questionnaire items in order to avoid store-dependent responses. Interpark.com, founded in 1997, is Korea’s first Internet-based shopping mall selling a broad range of goods and services including collectibles, appliances, computers, equipment, vehicles, food, tickets, clothes, jewelry, and tour packages. In 2011, Interpark.com has recorded market shares of 27% in books, 70% in entertainment tickets, 50% in tour products, and 8% in general merchandise in Korea. Although it recently introduced an online marketplace within the same Website (just as Amazon.com storefront and marketplace coexist on a single Website), the sales are predominantly generated by the digital storefront portion of the business. Prior to the main survey, we conducted a pilot test using 25 students in order to make sure that the questionnaire items were properly developed to meet the research objectives. We examined the responses to the preliminary instrument for consistency and revised the items in the questionnaire, such that there are no redundant items; all items are phrased clearly and concisely. Then, we surveyed a total of 214 students in order to access a suitable sample of consumers who experienced B2C online shopping. After eliminating observations with missing and unusable data, we used 206 observations to test the model and hypotheses. The participants were 68% male and 32% female, and most of the respondents were aged between their 20s and 30s. The respondents’ profile is summarized in Table 1. We used structural equation modeling (SEM) in order to analyze the data collected and test the research model. SEM is a statistical

technique that incorporates factor analysis (using a measurement model) and path analysis (using a structural model) (Qureshi & Compeau, 2009; Wetzels, Odekerken-Schroder, & Oppean, 2009). The advantages of SEM compared to other statistical techniques include more flexible assumptions (e.g., partial allowance of multicollinearity) and less measurement error with confirmatory factor analysis (CFA) enabled by multiple indicators per construct. In particular, we tested the model through partial least squares (PLS) using SmartPLS 2.0 with bootstrapping (Wetzels et al., 2009). 5. Results 5.1. Measurement model assessment The internal consistency (reliability) statistics were assessed by Cronbach’s alpha and composite reliability (Dillon Goldstein’s Rho), and the results are summarized in Table 2. All Cronbach’s Alpha and composite reliability values exceeded the recommended reliability threshold of 0.7 (Fornell & Larcker, 1981). Therefore, all of the questionnaire items were deemed reliable. In addition, we tested the convergent validity by examining the average variance extracted (AVE), which measures the percentage of the variance of the measurement items that can be accounted for by the constructs relative to the measurement error. Table 2 illustrates that for each construct, the AVE value was greater than the cut-off value of 0.5 (Yoo & Alavi, 2001). Further, we tested the discriminant validity by examining whether a latent variable better explains the variance of its own indicators than the variance of other latent variables. To validate this, we compared the square root of AVE for each construct with its cross-correlation with other constructs. The results supported the discriminant validity of our constructs in that in all cases, the diagonal elements in the matrix (i.e., the square root of AVE) were higher than the off-diagonal elements in the corresponding rows and columns, as shown in Table 2. Lastly, we tested the convergent validity using the factor and cross loadings of all indicator items in relation to their respective latent constructs. The results are summarized in Table 3, which indicate that all items loaded (i) on their respective constructs with a factor between 0.70 and 0.95 and (ii) more highly on their respective constructs than on any other construct. Further, these entire factor loadings were highly significant (t-statistics > 11.377, p < 0.001) based on the SmartPLS output. Therefore, we can confirm that these indicator items accurately represent distinct latent constructs. 5.2. Structural model assessment The assessment of the structural model includes estimation of the path coefficients and R2 values. In particular, to measure the

1.00

Note: 1. PER: Performance risk; 2. PSR: Psychological risk; 3. SOR: Social risk; 4. FIR: Financial risk; 5. OPR: Online payment risk; 6. DER: Delivery risk; 7. TR: Trust; 8. PI: Purchase intention; 9. GEN: Gender; 10. AGE: Age; 11. IUH: Internet usage hour; 12. ISF: Internet shopping frequency. The principal diagonal (in boldface) of the inter-correlation matrix represents the square root of the average variance extracted (AVE) per construct. Control variables include gender, age, Internet usage hour (IUH), and Internet shopping frequency (ISF).

1.00 −0.09 1.00 −0.09 0.10 1.00 −0.36 −0.06 0.02 0.86 −0.02 0.04 0.06 0.17 0.87 0.57 −0.03 0.03 0.06 0.06 0.86 −0.24 −0.17 0.04 −0.12 −0.10 −0.04 0.87 0.31 −0.25 −0.30 0.09 −0.06 −0.05 −0.06 0.80 0.18 0.30 −0.25 −0.35 0.14 −0.14 −0.03 −0.06 0.86 0.31 0.07 0.26 −0.32 −0.33 −0.08 −0.04 −0.12 −0.06

12 10 9 8 7 6 5 4 3 2

0.85 0.56 0.35 0.12 0.28 −0.35 −0.36 0.00 −0.07 −0.12 −0.02 0.84 0.27 0.27 0.37 0.33 0.26 −0.41 −0.36 0.04 −0.14 −0.03 −0.07

1 AVE

0.71 0.72 0.74 0.64 0.75 0.73 0.77 0.73 1.00 1.00 1.00 1.00 0.88 0.89 0.90 0.84 0.90 0.89 0.91 0.89 1.00 1.00 1.00 1.00

Reliability Alpha

0.80 0.80 0.83 0.74 0.84 0.82 0.85 0.82 1.00 1.00 1.00 1.00 0.85 0.89 0.91 0.77 1.00 0.77 0.71 0.77 0.47 2.19 14.60 1.17

SD Mean

Table 2 Reliability and convergent validity assessment of the measurement model.

3.14 2.44 2.12 2.68 3.13 3.07 3.27 3.07 0.32 24.48 18.25 4.25

14

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

1. PER 2. PSR 3. SOR 4. FIR 5. OPR 6. DER 7. TR 8. PI 9. GEN 10. AGE 11. IUH 12. ISF

934

effect of mediation in the research model, we sequentially assessed two separate structural models: the unmediated model and the mediated model. Fig. 3 and Table 4 show the unmediated structural model results with the ˇ values of all path coefficients. We found that performance risk (ˇ = −0.160, t-statistic = 2.328, p < 0.05), psychological risk (ˇ = −0.177, t-statistic = 2.643, p < 0.01), financial risk (ˇ = −0.167, t-statistic = 2.270, p < 0.05), and online payment risk (ˇ = −0.201, t-statistic = 2.967, p < 0.01) negatively affect purchase intention. However, we could not find a significant influence either from the social risk (ˇ = −0.121, t-statistic = 1.638, n.s.) or delivery (ˇ = −0.063, t-statistic = 0.905, n.s.) risk. The R2 for purchase intention was 0.29, reflecting that the variation in the given risk factors explains 29% of the total variance of consumer purchase intention. Fig. 4 and Table 5 show the mediated structural model results with the ˇ values of all path coefficients. Consistent with the unmediated model, social risk and delivery risk did not show any significant influence either in the direct or indirect path. We found that performance risk (ˇ = −0.273, t-statistic = 3.887, p < 0.01) and psychological risk (ˇ = −0.174, t-statistic = 2.114, p < 0.05) have a negative and significant impact on trust. Note that, after controlling trust, psychological risk still kept its direct impact on purchase intention (ˇ = −0.101, t-statistic = 1.836, p < 0.1); however, performance risk no longer showed a direct influence on purchase intention (ˇ = −0.038, t-statistic = 0.584, n.s.). Financial risk and online payment risk did not affect trust; yet, it only presented a direct impact on purchase intention (ˇ = −0.161, t-statistic = 2.384, p < 0.05 and ˇ = −0.154, t-statistic = 2.394, p < 0.01, respectively). Lastly, we found a significant positive impact of trust on purchase intention (ˇ = 0.428, t-statistic = 7.700, p < 0.01), which is necessary to support the hypotheses regarding the indirect impact of perceived risk on purchase intention by means of trust. R2 for purchase intention was 0.43, which is far greater than 0.291 found in the unmediated model. In terms of R2 , we found that R2 increased greatly from 0.291 in the unmediated model to 0.428 in the mediated model, which implies that the mediated model has a better fit than the original model. Given the results of the mediated model, we further examined the mediation effect of trust following the Baron and Kenny (1986) steps. Using the same notations shown in Fig. 2 (c, a, b, and c ) in the previous section, Table 6 presents the outcomes of the analysis in order to examine the mediational hypotheses. There are several ways to assess whether the mediated effect is significant or not. In particular, we tested the significance of the indirect effect (product of paths a and b) using the Sobel test (Sobel, 1982). The test statistic1 for both performance risk (z = −3.47, p < 0.01) and psychological risk (z = −2.04, p < 0.05) showed that trust was a significant mediator. The amount of mediation is often defined as the reduction of the effect of the initial variable on the outcome or the difference between the total effect and direct effect (i.e., |c−c |). Theoretically, this is same as the indirect effect or product of paths a and b (i.e., |c − c | ≈ |ab|). Baron and Kenny (1986) suggested that a small effect size would be |ab| = 0.01, medium size would be |ab| = 0.09, and large size would be |ab| = 0.25. In our results, performance risk showed that |c − c | = 0.160 (where c = 0 since the path coefficient is not statistically significant), which is slightly greater than |ab| = 0.117. On the other hand, psychological risk showed that |c − c | = 0.076, which is almost the same as the value of |ab| = 0.075. As a result, we concluded that medium to large size mediated effects for performance risk, and small to medium size mediated effects for psychological risk.

1

z=

ab . (b2 SEa2 )+(a2 SE 2 ) b

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

935

Table 3 The cross-loading matrix.

PER1 PER2 PER3 PSR1 PSR2 PSR3 SOR1 SOR2 SOR3 FIR1 FIR2 FIR3 PRR1 PRR2 PRR3 DER1 DER2 DER3 TR1 TR2 TR3 PI1 PI2 PI3 GEN AGE IUH ISF

PER

PSR

SOR

FIR

PRR

DER

TR

PI

GEN

AGE

IUH

ISF

0.84 0.87 0.82 0.20 0.19 0.33 0.30 0.28 0.14 0.22 0.31 0.34 0.27 0.22 0.33 0.15 0.18 0.34 −0.35 −0.34 −0.39 −0.29 −0.37 −0.25 0.04 −0.14 −0.03 −0.07

0.28 0.22 0.19 0.91 0.92 0.71 0.43 0.48 0.53 0.29 0.22 0.32 0.07 0.06 0.16 0.24 0.29 0.19 −0.32 −0.31 −0.30 −0.32 −0.32 −0.28 0.00 −0.07 −0.12 −0.02

0.29 0.17 0.20 0.58 0.55 0.26 0.84 0.91 0.82 0.17 0.08 0.37 −0.02 −0.02 0.16 0.24 0.24 0.17 −0.30 −0.23 −0.31 −0.27 −0.23 −0.35 −0.08 −0.04 −0.12 −0.06

0.33 0.26 0.32 0.32 0.36 0.20 0.24 0.36 0.20 0.77 0.72 0.90 0.06 0.06 0.28 0.32 0.27 0.20 −0.22 −0.25 −0.19 −0.33 −0.31 −0.25 0.14 −0.14 −0.03 −0.06

0.27 0.32 0.25 0.07 0.03 0.27 0.05 0.02 0.10 0.10 0.13 0.19 0.87 0.85 0.87 0.27 0.27 0.25 −0.16 −0.25 −0.25 −0.24 −0.31 −0.22 0.09 −0.06 −0.05 −0.06

0.25 0.27 0.14 0.26 0.20 0.27 0.18 0.19 0.28 0.21 0.20 0.30 0.26 0.22 0.30 0.86 0.90 0.81 −0.19 −0.16 −0.28 −0.20 −0.08 −0.17 0.04 −0.12 −0.10 −0.04

−0.38 −0.27 −0.38 −0.34 −0.32 −0.23 −0.19 −0.27 −0.35 −0.10 −0.12 −0.30 −0.17 −0.11 −0.31 −0.17 −0.24 −0.21 0.86 0.88 0.88 0.47 0.50 0.49 −0.03 0.03 0.06 0.06

−0.36 −0.29 −0.25 −0.34 −0.31 −0.26 −0.26 −0.29 −0.29 −0.21 −0.18 −0.37 −0.22 −0.23 −0.31 −0.17 −0.13 −0.15 0.46 0.53 0.50 0.89 0.90 0.77 −0.02 0.04 0.06 0.17

0.00 0.05 0.05 0.02 −0.01 −0.01 −0.06 −0.09 −0.05 0.13 0.09 0.13 0.08 0.09 0.07 0.04 −0.01 0.07 −0.03 −0.04 0.00 0.00 −0.06 0.01 1.00 −0.36 −0.06 0.02

−0.16 −0.11 −0.09 −0.06 −0.07 −0.03 −0.06 −0.01 −0.04 −0.10 −0.07 −0.14 −0.02 −0.03 −0.09 −0.09 −0.06 −0.16 0.04 0.03 −0.01 0.00 0.01 0.10 −0.36 1.00 −0.09 0.10

0.02 −0.11 0.00 −0.10 −0.07 −0.16 −0.13 −0.12 −0.06 0.02 −0.02 −0.04 −0.03 −0.02 −0.07 −0.03 −0.09 −0.14 0.02 0.03 0.10 0.02 0.07 0.06 −0.06 −0.09 1.00 −0.09

0.13 0.06 −0.01 0.05 0.06 −0.07 0.09 0.04 0.03 0.06 0.01 0.06 −0.01 0.00 0.12 0.07 0.02 0.01 −0.08 0.02 −0.10 −0.13 −0.04 −0.28 −0.02 −0.10 0.09 1.00

Perceived risk

Performance risk

- 0.160 (2.328)**

Psychological risk - 0.177 (2.643)***

Social risk

Financial risk

- 0.121(1.638)

Purchase intention R2 = 0.291

- 0.167 (2.270)** - 0.201 (2.967)***

Online payment risk

Control Variables: • Age: -0.039 (0.625) • Gender: 0.002 (0.027) • Internet usage: 0.016 (0.296) • Internet shopping frequency:0.132 (1.887)*

0.063 (0.905)

Delivery risk

Fig. 3. The results of the unmediated research model.

Table 4 Summary of the results of the unmediated model. Hypothesis

Effect

Coefficient

H1-1 H1-2 H1-3 H1-4 H1-5 H1-6 Control

Performance risk → purchase intention Psychological risk → purchase intention Social risk → purchase intention Financial risk → purchase intention Online payment risk → purchase intention Delivery risk → purchase intention Age Gender Internet usage Internet shopping frequency

−0.160 −0.177 −0.121 −0.167 −0.201 −0.063 −0.039 0.002 0.016 0.132

* ** ***

p < 0.1. p < 0.05. p < 0.01.

S.E. 0.069 0.067 0.074 0.073 0.068 0.069 0.060 0.063 0.055 0.070

t-Statistics **

2.328 2.643*** 1.638 2.270** 2.967*** 0.905 0.625 0.027 0.296 1.887*

Conclusion Supported Supported Not supported Supported Supported Not supported

936

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

Perceived risk

Performance risk

-0.273 (3.887)***

Trust R2 = 0.258

-0.174 (2.114)** -0.126 (1.580) -0.015 (0.242) -0.115 (1.632) -0.038 (0.584) -0.046 (0.605)

Psychological risk

Social risk

0.428(7.700)***

-0.101 (1.836)* 0.077(1.053)

Financial risk

Purchase intention R2 = 0.428

-0.161 (2.384)** -0.154 (2.394)***

Online payment risk 0.088 (1.345) Control Variables: • Age: -0.012 (0.217) •Gender: 0.016 (0.291) •Internet Usage: 0.019 (0.361) • Internet Shopping Frequency:0.124 (1.875)**

Delivery risk

Fig. 4. The results of the mediated research model.

A different way to describe the amount of mediation is in terms of the proportion of the total effect that is mediated, which is defined by ab/c (Frazier et al., 2004; Shrout & Bolger, 2002). Given the path coefficients, we obtain 0.117/0.160 = 0.73 for performance risk and 0.075/0.177 = 0.42 for psychological risk. Thus, about 73% of the total effect of performance risk on purchase intention is mediated by trust, and about 42% of the total effect of psychological risk on purchase intention is mediated by trust. Similar to this approach, we examined the types of mediation as well. The relationship between performance risk and purchase intention was

completely mediated by trust (i.e., c = 0, n.s). In contrast, the relationship between psychological risk and purchase intention was partially mediated by trust, where the absolute size of the direct path coefficient was reduced by |c − c | = 0.160, while c is still not zero. 6. Discussion The findings of the present research point to a set of implications for the academics. Most of all, our analysis of the cause-and-effect

Table 5 Summary of the results of the mediated model. Hypothesis

Effect

Coefficient

S.E.

t-Statistics

Conclusion

H2-1

Performance risk → purchase intention Performance risk → trust Psychological risk → purchase intention Psychological risk → trust Social risk → purchase intention Social risk → trust Financial risk → purchase intention Financial risk → trust Online payment risk → purchase intention Online payment risk → trust Delivery risk → purchase intention Delivery risk → trust Trust → purchase intention Gender Age Internet usage Internet shopping frequency

−0.038 −0.273 −0.101 −0.174 −0.077 −0.126 −0.161 −0.015 −0.154 −0.115 0.088 −0.046 0.428 0.017 −0.013 0.020 0.124

0.066 0.070 0.055 0.082 0.073 0.080 0.068 0.064 0.064 0.071 0.066 0.077 0.056 0.057 0.057 0.054 0.066

0.584 3.887*** 1.836* 2.114** 1.053 1.580 2.384** 0.242 2.394** 1.632 1.345 0.605 7.700*** 0.291 0.217 0.361 1.875*

Supported

H2-2 H2-3 H2-4 H2-5 H2-6 H2-1–6 Control

* ** ***

Supported Not supported Not supported Not supported Not supported – – – – –

p < 0.1. p < 0.05. p < 0.01.

Table 6 Summary of the results for mediation effect. Risk type

Path

Path coefficient

S.E.

t-test

Sobel test

Mediation type

Performance risk

c a b c c a b c

−0.160 −0.273 0.428 −0.038 −0.177 −0.174 0.428 −0.101

0.069 0.070 0.056 0.066 0.067 0.082 0.056 0.055

2.328** 3.887*** 7.700*** 0.584 2.643*** 2.114** 7.700*** 1.836*

z = −3.47 (p < 0.01)

Complete mediation

z = −2.04 (p < 0.05)

Partial mediation

Psychological risk

* ** ***

p < 0.1. p < 0.05. p < 0.01.

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

relationship between the perceived risk and purchase intention was conducted by separating the risk into six dimensions. Existing research often considers perceived risk as a unidimensional construct, and thus, focuses on demonstrating that the perceived risk as a whole tends to inhibit consumer attitude and transaction intentions. On the contrary, our results revealed that the impact of perceived risk on its consequences were different depending on the dimensions of the perceived risk. The results provide substantial support for the research model as shown in Fig. 1. Four (H1-1, H1-2, H1-4, and H1-5) out of six hypotheses were supported, regarding the unmediated influence of perceived risk. We also found two significant mediation effects supporting H2-1 and H2-2. In the unmediated model, while we found that the consumer’s perceived risk mostly has a negative influence on purchase intention (i.e., performance risk, psychological risk, financial risk, and online payment risk), two types of perceived risk (i.e., social risk and delivery risk) turned out not to influence purchase intention; thus, H1-3 and H1-6 were rejected. The lack of support for these two hypotheses may be the outcome of the changes in the recent online shopping environment. A likely reason why there was no significant influence of social risk was that Internet shopping has become so popular amongst users that nobody considers purchasing a product online as unusual. In particular, with the advance of Web 2.0 tools, such as social network systems, consumers who have purchased products are increasingly sharing their buying experience online. Furthermore, potential customers can also have access to useful suggestions from others and anticipate what responses their friends and family will give concerning their online purchase in advance. As a result, other people’s views on online shopping are no longer a concern. With regards to the delivery risk, most online merchants outsource their delivery function to well-known specialized companies, such as UPS, FedEx, and DHL. With advances in new technologies, such as RFID and wireless barcode reading devices, these companies provide real-time tracking information. In particular, in a metropolitan area, same-day delivery service is very common and reliable. In addition, the consumer knows that in the event of incorrect delivery he can always call the customer service to identify the potential problem and request that the order should be reshipped to the correct address. Hence, a consumer who is willing to buy a product online is not likely to abandon his intention, even if he has some worries over correct delivery of the order, because he will assure himself that any potential issue with delivery can be properly addressed by the vendor. In the mediated model, it was confirmed that performance and psychological dimensions of perceived risk and trust are in a very close, inseparable relationship, and this finding is consistent with other studies (Johnson-George & Swap, 1982; Olivero and Lunt, 2004). From a managerial perspective, to reduce the perceived performance risk, a firm may consider ways to reduce discrepancies in product appearance, specification, and quality as advertized in the online Website. For example, Matsuhita Electric Works has decided to allow consumers to design their kitchen in virtual reality and choose matching appliances (Haag & Cummings, 2009). Likewise, a CAVE (cave automatic virtual environment) provides a 3D virtual reality room where one can even talk with a remotely located sales person, feeling that she is in the same room. Similar technologies, such as hepatic interfaces and custom-fit clothes through biometrics, are evolving to overcome the limitations of online shopping. As another way to reduce performance risk, firms may utilize active marketing, online advertising, and promotional activities in order to attract consumers by emphasizing that the quality and performance of the products purchased online is as good as that of those purchased offline. For example, firms can promote the active participation of existing shoppers through a discussion board where they can post reviews indicating that their purchased products met

937

their expectations. This may reduce the risk associated with product discrepancy. To overcome the psychological risks for consumers, it is suggested that online merchants focus on identifying target customers and offering products that best meet the psychological needs of those customers. In addition, it may be necessary to improve the process related to returning and exchanging products purchased online. When consumers know that they can easily return or exchange any product with which they do not feel quite comfortable, much of their psychological concerns will be relieved. For example, Amazon.com provides an automated process of enabling a customer to request a return and to print a return address label, and thus, customers feel that the cost of resolving the psychological discomfort resulting from the wrong choice of a product is minimal. Online merchants must keep in mind that online trust is formed slowly over time as consumers gain experience through repeated transactions (Cheskin-Research, 1999). Meanwhile, it is interesting to note that the financial risk and online payment risk had a direct negative influence on purchase intention but not on the consumer’s trust in a merchant. In terms of financial risk, many price-comparison Websites (e.g., pricegrabber.com, nextag.com, and bizrate.com) are available to provide consumers with easily accessible and reliable price information. Moreover, consumers do not consider a merchant trustworthy merely based on its relatively low prices. Instead, in addition to pricing information, these Websites also provide sellers’ ratings based on existing shoppers’ reviews, which may be critical information for consumers in building their trust. Indeed, people often shop at highly reputable stores that they trust, such as Amazon.com, even though prices may be higher than those of competitors. Meanwhile, one possible reason why online payment risk was directly related to purchase intention may be that today more and more online stores tend to outsource the online payment function to a reliable third party payment solution provider in order to avoid risks associated with payment handling. For example, PayPal has been embedded in many online stores and has processed over $ 71 billion through 87 million registered users in 2009 (www.wikipedia.com, retrieved on Sep. 10, 2013). Although the relevant risks associated with using this kind of specialized payment service still exist and have an impact on purchase intention, they may not reduce the consumers’ trust in the store itself.

7. Conclusions Recently, the Internet is being widely used as an important vehicle to conduct business transactions online, as it removes time and space barriers to enable convenient 24/7 shopping for customers. We have seen steady growth in electronic commerce sales as well as the number of online consumers. Such changes have been driven in part by improvements in the Web-based ordering system and reduction in transaction costs. Nevertheless, consumer perceptions of risks associated with online purchases remain a great obstacle to the continued growth of electronic commerce. In this context, this paper focused on investigating the intriguing relationships among dimensions of perceived risk, consumer trust, and purchase intention. An empirical study was conducted in two phases: (1) examining the total effect without mediation, and (2) examining the mediation effect. When we probed the total effect under the unmediated model, the findings revealed that performance, psychological, financial, and online payment risks have significant negative influence on purchase intention. On the other hand, an examination of the mediation effect under the mediated model indicated that trust in an online merchant completely mediates the effect of performance risk but partially mediates that of psychological risk. Given the mixture

938

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939

of unmediated as well as mediated effects of perceived risks on purchase intention, the findings have confirmed that while there is a negative relationship between perceived risk and purchase intention, this relationship is also mediated by consumer trust in an online merchant. Therefore, it is reasonably conceivable that efforts made by online merchants to lessen certain types of risk will first improve consumer trust, and then ultimately increase the consumer’s intention to purchase online.

7.1. Implications

Appendix A. Table A1. List of item variables and survey questions. Item code

Questionnaire

PeR1

The product quality may be lower than that advertised in the online store The product appearance may be different from the product picture shown in the online store The product dimension may be different from that advertised in the online store If I bought a product from the online store, I would abase myself If I bought a product from the online store, it would not fit with my image The online store would not sell high-class products If I bought a product from the online store, I would be held in lower esteem by my friends and families If I bought a product from the online store, I would be negatively thought of by my friends and family If I bought a product from the online store, I would be demeaned by my friends and family I would be concerned that the product in the online store may be more expensive than products in a different place I would be concerned that I might be able to buy the same product at a different place at a lower price than in the online store If I bought a product from the online store, I may suffer monetary loss due to sales fraud I would be concerned as to whether the online store is equipped with a security monitoring tool I would be concerned as to whether the online store is equipped with a security-enabled log-in process I would be concerned as to whether the online store appropriately manages customers’ private information If I bought a product from the online store, I would be concerned as to whether the product would be delivered to a wrong address If I bought a product from the online store, I would be concerned as to whether the product would be lost during delivery If I bought a product from the online store, I would be concerned as to whether a wrong product would be delivered I trust the online store and would purchase products from this Website I believe that the online store is trustworthy. I believe the online store will keep its promises and commitments I would like to purchase a product from this online store I would like to recommend my friends and family to purchase a product from this online store If there is a product that I want to purchase, I would like to use the online store What is your age? What is your gender?

PeR2 PeR3 PsR1 PsR2

The present research offers academic as well as practical implications. First, it makes scholarly contributions by providing new insights into the theoretical relationships among perceived risk, consumer trust, and purchase intention. Unlike the existing research, our research has found that some, if not all, components of perceived risk are an inhibitor of consumer trust. It was revealed that performance and psychological risks have a negative effect on trust. Research findings further indicate that the relationships between these two risks and purchase intention are mediated by trust in an online merchant. In addition, some components of perceived risk (i.e., performance, psychological, financial, and online payment risks) were shown to have a negative influence on consumer’s intention to buy online. This finding provides new insights, since we addressed the effects of the individual components of perceived risk, rather than that of perceived risk as a whole. On the other hand, the study provides practical implications for managers of electronic commerce firms. First, the mediating role of consumer trust in the relationships between perceived risks and purchase intention suggests that an online merchant can increase sales by first lowering the perceived risks, thereby improving consumer trust that then will function to boost purchase intention. In order to make this vision a reality, electronic commerce firms will need to ensure that consumer trust can be enhanced by effectively reducing performance and psychological risks associated with online purchases. Furthermore, practitioners must keep in mind that purchase intention is directly affected by performance, psychological, financial, and online payment risks. It implies that they should make efforts to mitigate these risk perceptions if they are to increase revenues.

PsR3 SoR1 SoR2 SoR3 FiR1 FiR2

FiR3 OpR1 OpR2 OpR3 DeR1

DeR2

DeR3

TR1 TR2 TR3 PI1 PI2 PI3

7.2. Limitations Despite the potential contributions mentioned earlier, this research is subject to a few limitations. The first shortcoming is that the use of students as respondents in the survey makes the research results less realistic than when actual consumers were employed. Although an increasing number of college students today are online consumers themselves, the range of products they buy online is somewhat limited. Second, the research model may have overlooked other antecedents to consumer trust. While the estimation shows that purchase intention is significantly influenced by consumer trust, purchase intension could be affected by other factors, such as the reputation of the online merchant and the advertisement, which may be also correlated with consumer trust. Should this be the case, trust and purchase intention may have no direct causal connection. Third, the present research does not take into account the reputation of an online store. In general, consumer trust depends largely on the reputation of online stores. For example, many consumers have confidence in buying goods from Amazon, yet perceive a considerable amount of risk when buying from an unknown e-commerce Website.

Age Gender

References Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. J. Pers. Soc. Psychol., 51, 1173–1182. Bauer, R. A. (1960). Consumer behavior as risk taking. Chicago, IL: American Marketing Association. Bhatnagar, A., Misra, S., & Rao, H. R. (2000). On risk, convenience and Internet shopping behavior. Commun. ACM, 43(11), 98–108. Cases, A. S. (2002). Perceived risk and riskreduction strategies in Internet shopping. Int. Rev. Retail Distrib. Consum. Res., 12(4), 375–394. Center I.C.C. (2011). IC3 2010. In Center I.C.C. (Ed.), Annual report on Internet crime Center I.T.R. (2011). In Center I.T.R. (Ed.), 2011 data breach stats. Cheskin-Research. (1999 January). e-Commerce trust study: a joint research project, Cheskin and studio archetype/sapient. Cheung, C. M. K., & Lee, M. K. O. (2001). Trust in Internet shopping: Instrument development and validation through classical and modern approaches. J. Global Inf. Manage., 9(3), 23–32. Corbitt, B. J., Thanasankit, T., & Yi, H. (2003). Trust and e-commerce: A study of consumer perception. Electron. Commer. Res. Appl., 2, 203–215. Cox, D. F., & Rich, S. (1964). Perceived risk and consumer decision making. J. Marketing Res., 1, 32–39.

I.B. Hong, H.S. Cha / International Journal of Information Management 33 (2013) 927–939 Deutsch, M. (1973). The resolution of confiict: Constructive and destructive processes. New Haven, CT: Yale University Press. Featherman, M. S., & Pavlou, P. A. (2003). Predicting e-services adoption: A perceived risk facets perspective. Int. J. Hum. Comput. Stud., 59(4), 451–474. Flavian, C., Guinaliu, M., & Gurrea, R. (2005). The role played by perceived usability, satisfaction and consumer trust on website loyalty. Inf. Manage., 43(1), 1–14. Fornell, C. R., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and measurement error. J. Marketing Res., 18(1), 39–50. Frazier, P. A., Tix, A. P., & Barron, K. E. (2004). Testing moderator and mediator effects in counseling psychology. J. Counseling Psychol., 51(1), 115–134. Garbarino, E., & Strahilevitz, M. (2004). Gender differences in the perceived risk of buying online and the effects of receiving a site recommendation. Journal of Business Research, 57(7), 768–775. Gefen, D. (2000). e-Commerce: The role of familiarity and trust. Omega: Int. J. Manage. Sci., 28(6), 725–737. Gefen, D. (2002). Reflections on the dimensions of trust and trustworthiness among online consumers. ACM SIGMIS Database, 33(3), 38–53. Gefen, D., Karahanna, E., & Straub, D. W. (2003). Trust and TAM in online shopping: An integrated model. MIS Q., 27(1), 51–90. Haag, S., & Cummings, M. (2009). Management information systems for the information age (8th ed.). Irwin, ID: McGraw-Hill. Hassan, A. M., Kunz, M. B., Pearson, A. W., & Mohamed, F. A. (2006). Conceptualization and measurement of perceived risk in online shopping. Marketing Management Journal, 16(1), 138–147. Heijden, H. v., Verhagen, d., & Creemers, T. M. (2003). Understanding online purchase intentions: Contributions from technology and trust perspectives. Eur. J. Inf. Syst., 12, 41–48. Hoffman, D. L., Novak, T. P., & Peralta, M. (1999). Building consumer trust online. Commun. ACM, 42(4), 80–85. Hong, I. B., & Cho, H. (2011). The impact of consumer trust on attitudinal loyalty and purchase intentions in B2C e-marketplaces: Intermediary trust vs. seller trust. Int. J. Inf. Manage., 31(5), 469–479. Jacoby, J., & Kaplan, L. B. (1972). The components of perceived risk Paper presented at the third annual conference, Association for Consumer Research, Iowa City, IA. Jarvenpaa, S. L., & Leidner, D. E. (1999). Communication and Trust in Global Virtual Teams. Org. Sci., 10(6), 791–815. Jarvenpaa, S. L., & Todd, P. A. (1997). Consumer reactions to electronic shopping on the world wide Web. J. Electron. Commer., 1(2), 59–88. Jarvenpaa, S. L., Tractinsky, J., & Vitale, M. (2000). Consumer trust in and Internet store. Inf. Technol. Manage., 1, 45–71. Jarvenpaa, S. L., & Tractinsky, N. (1999). Consumer trust in an Internet store: A crosscultural validation. J. Comput. Mediated Commun., 5(2), 1–35. Johnson-George, C., & Swap, W. (1982). Measurement of specific interpersonal trust: Construction and validation of a scale to assess trust in a specific other. J. Pers. Soc. Psychol., 43(6), 1306–1317. Koller, M. (1988). Risk as a determinant of trust. Basic Appl. Soc. Psychol., 9(4), 265–276. Kurtz, D. L., & Clow, K. E. (1997). Services marketing. New York, NY: John Wiley & Sons. Lee, M., & Turban, E. (2001). A trust model for consumer Internet shopping. Int. J. Electron. Commer., 6(1), 75–91. Mayer, R. C., Davis, J. H., & Schoorman, F. D. (1995). An integrative model of organizational trust. Acad. Manage. Rev., 20(3), 709–734. McKnight, D. H., Choudhury, V., & Kacmar, C. (2002). Developing and validating trust measures for e-commerce: An integrative typology. Inf. Syst. Res., 13(3), 334–359. Morgan, R., & Hunt, S. (1994). The commitment-trust theory of relationship marketing. J. Marketing, 58(3), 20–38. Olivero, N., & Lunt, P. (2004). Privacy versus willingness to disclose in e-commerce exchanges: The effect of risk awareness on the relative role of trust and control. J. Econ. Psychol., 25(2), 243–262.

939

Pavlou, P. A. (2003). Consumer acceptance of electronic commerce: Integrating trust and risk with the technology acceptance model. Int. J. Electron. Commer., 7(3), 69–103. Pavlou, P. A., & Gefen, D. (2004). Building effective online marketplaces with institution-based trust. Inf. Syst. Res., 15(1), 37–59. Peter, J. P., & Ryan, M. J. (1976). An investigation of perceived risk at the brand level. J. Marketing Res., 13, 184–188. Qureshi, I., & Compeau, D. (2009). Assessing between-group differences in information systems research: A comparison of covariance-and componet-based SEM. MIS Q., 33(1), 197–214. Ring, P. S., & Van de Ven, A. H. (1994). Developing processes of cooperative interorganizational relationships. Acad. Manage. Rev., 19, 90–118. Salam, A. F., Iyer, L., Palvia, P., & Singh, R. (2005). Trust in e-commerce. Commun. ACM, 48(2). Schiffman, L. G., & Kanuk, L. L. (1994). Consumer behavior. Englewood Cliffs, NJ: Prentice Hall. Shankar, V., Urban, G. L., & Sultan, F. (2002). Online trust: A stakeholder perspective, concepts, implications, and future direction. J. Strategic Inf. Syst., 11, 325–344. Shrout, P. E., & Bolger, N. (2002). Mediation in experimental and non-experimental studies: New procedures and recommendations. Psychol. Methods, 7, 422–445. Sobel, M. E. (1982). Asymptotic confidence intervals for indirect effects in structural equation models. Sociol. Method, 13, 290–312. Stone, R. N., & Gronhaug, K. (1993). Percerived risk: Further considerations for the marketing discipline. Eur. J. Marketing, 27, 39–50. Taylor, J. W. (1974). The role of risk in consumer behavior. J. Marketing, 38, 54–60. Verhagen, T., Meents, S., & Tan, Y.-H. (2006). Perceived risk and trust associated with purchasing at electronic marketplaces. Eur. J. Inf. Syst., 15, 542–555. Ward, M. R., & Lee, M. J. (2000). Internet shopping, consumer search and product branding. J. Prod. Brand Manage., 9, 6–18. Wetzels, M., Odekerken-Schroder, G., & Oppean, C. V. (2009). Using PLS path modeling for assessing hierarchical construct models: Guidelines and empirical illustration. MIS Q., 33(1), 177–195. Yoo, Y., & Alavi, M. (2001). Media and group cohesion: Relative influences on social present, task participation, and group consensus. MIS Q., 25(3), 371–390. Zikmund, W. G., & Scott, J. E. (1977). An Investigation of the role of product characteristics in risk perception. Rev. Bus. Econ. Res., 13, 19–34. Ilyoo B. Hong is presently professor of management information systems at ChungAng University, Seoul, Korea. He earned his Ph.D. degree in MIS from the University of Arizona, his MS in Business from the University of Illinois at Urbana-Champaign, and his B.S. in Management from Indiana University. He was a visiting scholar at the UCLA Anderson School of Management. Ilyoo Hong has published in such journals as Decision Sciences, Information & Management, and International Journal of Information Management, among others. He also presented academic papers at numerous international conferences, including HICSS. His research interests include building online consumer trust, measuring the quality and impact of Web-based information systems, and information disclosure in social networking sites. Hoon S. Cha is associate professor of management information systems at ChungAng University. He holds an M.S. and Ph.D. in Management Information Systems from the University of Arizona and a B.S. in Material Sciences and Engineering from Seoul National University. He worked for Samsung for three years as an IT consultant. His research examines the impact of IT offshoring decisions on firm knowledge and costs and the allocation of IT investment among business functions and its effects on business value. His publications have appeared in MIS Quarterly, Journal of Management Information Systems, and Communications of the ACM, among others.