Content area
Full Text
Introduction
It has been a common practice for internet users to provide their personal information in exchange for website services access. Users become registered members and are allowed to use a website’s services that are either personalized (system-tailored) or customized (user-tailored) by offering limited or detailed demographics or personal preference information, limited (Sundar and Marathe, 2010). While users may continue to enjoy excellent website services, information privacy concerns are increasingly at stake (Liu et al., 2004). Information privacy concerns arise when a user subjectively perceives a threat resulting from his/her personal information being intruded upon in one or more of the following ways: improper access, unpermitted collection, unauthorized secondary use and incorrect capture (Smith et al., 1996).
After an interdisciplinary review of literature on privacy concerns, Smith et al. (2011, p. 989) proposed the antecedent–privacy concerns–outcome (APCO) model. Application of the APCO model revealed that once users perceive that their privacy is likely invaded, they might be unwilling to share their personal information with websites. However, this negative relationship may not always hold true; users may actually disclose personal information despite their privacy concerns (Lee and Cranage, 2011). Such a phenomenon is known as the privacy paradox (Norberg et al., 2007). The privacy paradox is relevant to the web environment because an increasing number of websites, such as Amazon, invite users to provide personal information and enjoy personalized services (Garfinkel et al., 2008). Researchers explained this paradoxical phenomenon with the privacy calculus theory: extrinsic benefits or rewards such as price discounts should additionally be taken into an account in reaching the decision to provide personal information (Wang and Wu, 2014). Likewise, privacy disclosure behavior is the function of privacy benefits and privacy costs (Dinev et al., 2008). Recent studies have proposed that different levels of information processing (i.e. central or peripheral) exist when users perform privacy calculus (Angst and Agarwal, 2009). When users perform central information processing, they are likely to focus on the interaction of benefits and costs; in contrast, when users perform peripheral information processing, they may only consider the main effects of benefits and costs or one of them. However, few studies have examined privacy calculus mechanisms while considering different levels of information processing. Hence, the first...