The Relationship between Customers Channel Usage and
Privacy Concerns: An Empirical Study in the Clothing
Industry
Student Name: Songjuan Li
Student Number: 11709693
June 22, 2018
Final Version
MSc. in Business Administration – Digital Business
University of Amsterdam
Content
1. Introduction ...5 2. Literature Review ...7 2.1 Privacy Concerns ...7 2.2 Perceived Risks ...8 2.3 Perceived Benefits ...92.4 Willingness to Share Personal Information ...10
2.5 Customer Purchasing Journey ...11
2.5.1 Pre-purchase Stage ...12 2.5.2 Purchase Stage ...13 2.6 Retailing Channels ...15 2.6.1 Physical Stores ...15 2.6.2 Website ...16 2.6.3 Mobile Channel ...17 2.6.4 Social Media ...18 2.7 Customer Segments ...20 2.8 Conceptual Model ...22
3. Research Design and Methodology ...25
3.1 Sampling and Data Collection ...25
3.2 Measurement ...26
4. Results and Discussion ...28
4.1 Sample Characteristics ...28
4.2 Segmentation ...29
4.3 Demographic Profile of the Segments ...32
4.4 Hypotheses Testing ...33
4.4.1 Correlation Analysis ...33
4.4.2 Regression Analysis ...35
4.5 Privacy Concerns of the Segments ...47
4.6 The Willingness to Share Personal Information of the Segments: One-way ANOVA ...48
4.7 The Willingness to Share Personal Information of the Segments: Regression Analysis ...57
5. Conclusions ...66
6. Limitations ...68
Statement of originality
This document is written by Student Songjuan Li who declares to take full responsibility for the contents of this document.
I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.
The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.
Abstract
As the retailing landscape is moving to a multi-channel or omni-channel era, customer purchasing journey goes beyond the online channel, much existing research on privacy issues is still limited to the online context. Although there is a significant amount of literature on consumer channel usage, little research has investigated customer channel usage from the perspective of privacy concerns at present. Therefore, this study was aimed at investigating the relationship between customer channel usage and privacy concerns. This study (1) proposes a conceptual model of customer privacy concerns; (2) segments customers on the basis of their channel usage in information search and purchase stages; (3) explores the differences between the segments in terms of privacy concerns and willingness to share four types of personal information: contact, general financial, demographic and lifestyle. Using survey data from 143 students who are still under higher education, three segemnts are identified: mobile-oriented shoppers, omnichannel shoppers and physical store-oriented shoppers. The results of analysis that focused on the privacy concerns and willingness to share personal information of the segments suggest that omnichannel shoppers are less concerned about their privacy compared to other two segments; compared to mobile-oriented shoppers, omnichannel shoppers are more likely to share their contact information, demographic information and lifestyle information, while physical store-oriented shoppers are more willing to share their general financial information.
Key words: perceived risks, perceived benefits, privacy concerns, willingness to share personal information, types of personal information, channel usage, , customer segments
1. Introduction
With the attempt to satisfy customers’ needs for ever more personalised products or service, companies for a long time have endeavored to improve their abilities to meet such needs. This inherently involves the demand by companies for customer information, both on a greater and more accurate scale. The rapid advances in information technology have enhanced data collection, data sharing, and data mining techniques (Gaurav et al. 2016), and companies today have more access to customers’ personal information than ever. However, the advent of same technology also brings concerns and threats which are related personal data. Customers feel uncomfortable with their personal information being collected because of potential risks such as unauthorized use of their data, data breaches, identity theft. Thus, Individuals become increasingly reluctant to share personal information and even resort to falsifying personal information to mitigate the risk of abuse (Xunet al. 2008).
Privacy concerns of personal information are recognized as a fundamental theme in marketing literature (Nuno et al. 2017). Much of the previous research has investigated privacy issues in the context of e-commerce and has generated many valuable findings. Some have investigated the factors that influence customers’ privacy concerns (Maurice et al. 1997; Gaurav et al. 2016), others have focused mainly on the direct impact of privacy concerns in online purchase intention (Gaurav et al. 2016;Budnitz et al. 1998) or willingness to disclosure information online (Xie et al. 2006; Mothersbaugh et al. 2012; Joseph et al. 2000). Much of the literature is limited to an online context. However, the world of retailing has changed
dramatically in the last two decades. The advent of the online channel and new additional digital channels such as mobile channels and social media have pushed the retail landscape continuing to change: moving from a single online or offline channel to a multi-channel to an omni-channel retailing (Peteret al. 2015). As the retailing is moving to the omnichannel era, customers no longer interact with retailers only online. Instead, different channels may be interchangeably and seamlessly used during each stage of the customer journey. Despite the fact that customers may no longer interact with brands only through the online channel because of the recent proliferation of channels, little research on customer privacy concerns and how such concerns affect customers’ privacy related behaviors goes beyond the online
context. Although many prior studies have segmented customers based on their’ channel choice or usage at different shopping stages, so far there is no such researchinvestigating customer segmentation incorperating privacy concerns. In order to fill such a gap, this study makes efforts following a three-step strategy. At first step, a research model in which perceived risks, perceived benefits and privacy concerns conceptualized as the factors that influence a customer’s willingness to share personal information is developed to understand the impact of perceived risks, perceived benefits and privacy concerns on willingness to share information; then this study will segment customers based on their channel usage at
information search and purchase stages of the customer journey; at the final step, this study will investigate whether each customer segment differs in privacy concerns and willingness to share personal information.
Furthermore, while little research has examined whether the effects of disclosure antecedents might differ depending on the type of personal information requested, this study incorporates the types of personal information into a broader framework of the willingness to disclose information. Failure to account for the types of personal information requested is important for three reasons. First, failure to account for the types of personal information leaves open questions about when and why disclosure antecedents affect customers’ willingness to share personal information with firms. Second, the frequent use of overall disclosure measures that do not account for the types of personal information requested leads to the mixed results of prior research if, as we expect, the effects of disclosure antecedents vary depending on which type of personal information. Last, failure to consider the different types of personal information hinders progress in developing strategies to obtain different types of personal information or to adapt information requests to match the situation (Mothersbaugh et al. 2012).
Thus, this study can be expected to contribute the existing literature in two ways. First, this study investigates what factors influence consumers’ willingness to disclose information as a function of the different types of the information requested. Second, it is the first empirical work exploring the relationship between channel usage, privacy concerns and willingness to share personal information. More specifically, this study aims to understand if
the willingness to disclose personal information varies among different customer channel segments (i.e. segments determined using customers’ channel usage at information search and purchase stages as segmentation bases).
2. Literature Review
2.1 Privacy Concerns
Privacy is not a new topic, and there is substantial literature having explored privacy issues from a wide range of perspective. Privacy in this research refers to the right of
individuals to determine how, when and to what extent information about them is revealed to others (Smith et al. 1996). Privacy issues are growing as the fact that the increasingly
sophisticated information technology is enabling companies in collecting quantitative and qualitative data to profile better their customers (Margherita et al. 2017), which leads to an increase in concerns among customers about how companies collect, use and store their personal information. Customers are vulnerable in that they have little control over
information collection and use by companies beyond the original purpose of the information collection. When a privacy intrusion happens, customers may experience associated adverse consequences, including a fear of being personally monitored, loss of anonymity,
unauthorised commercial use, identity theft, and fraud (Julia et al. 2014).
Privacy concerns may occur, depending on consumers' perception of potential adverse outcomes of giving away their personal information. Frequently, consumers perceive the disclosure of private data as a personal sacrifice (Milne and Mary 1993; Son and Sung 2008). Consumers with strong privacy concerns have a generally negative attitude toward all forms of offers by firms (Maurice et al. 1997). They have a strong fear that their personal
information is used wrongly and usually do not trust the good intentions of firms either. Given this, this study already assumed a direct, negative impact of privacy concerns on customers’ willingness to share personal information.
Privacy concerns are considered as a result of a situational and context-specific cost-benefit analysis of information disclosure (Acquisti and Grossklags 2005; Culnan and
Armstrong 1999; Dinevet al. 2006). The cost stands for disclosing personal information which may have potentially adverse outcomes on customers, includingcollection, errors, unauthorized sharing and access, and information use beyond its intended purpose (Metzger 2006; Smith et al. 1996). In contrast, the benefit is defined as what customer can gain from the same disclosure. Thus, in this study customers’ privacy concerns is considered as the result of privacy calculus, which constitutes a trade-off analysis of risks and benefits of sharing personal information with others.
2.2 Perceived Risks
Consumers make decisions regarding whether or not to share their personal data with firms. The outcomes (or consequences) of such decisions are often uncertain, and the consumer perceives some degree of risk in information disclosure. This perceived risk is defined as the uncertainty that consumers face when they cannot foresee the consequences of their information disclosure decisions. The prior study (Stuart et al. 2006) found a clear negative relationship exists between perceived privacy risk and willingness to provide personal information online. The degree of risk that consumers perceive and their own tolerance for risk taking are factors that influence their willingness to share personal
information. Disclosing personal information has been considered inherently risky, which is due to the uncertainty associated with how companies handle shared personal information and who has access to it. These negative consequences include selling to or sharing information with third parties, financial institutions (Budnitz 1998) or government agencies (Preston 2004; Wald 2004), and the misuse of personal information, such as insider disclosure or
unauthorized access and theft (Rindfleisch 1997). In online environments, people who perceive higher threats to privacy are willing to share information about the self because they perceive themselves as less able to control information and also protect themselves. In contrast, when people perceive lower privacy risks and higher control, such as when privacy policies are clearly exposed, they disclose more personal information (Stuart et al. 2006).
Studies have discovered that customers’ expectations of negative outcomes, such as online scams or online identity theft, were positively related to their privacy concerns
(LaRoseet al. 2007; Norberg et al. 2007). Dinev and Hart (2004) discovered that perceived vulnerability to privacy risks was positively related to privacy concerns. When customers provide their personal information to the brand, they give over something they feel belong to them and thus “they feel they should retain the right to control it and not be harmed, even after disclosure” (Metzger 2006; Metzger 2007). Sharing personal information renders people vulnerable to opportunistic exploitation because the shared private information becomes co-owned by other parties (Petronio 2002). As such, disclosure always involves some degree of risk (Metzger 2007). A number of e-commerce studies empirically verified the positive effect of perceived risk on customers’ privacy concerns (Fusilier and Hoyer 1980;Petronio 1991; Petronio 2002). Therefore, the higher the perceived risk customers have, the higher their privacy concerns are.
2.3 Perceived Benefits
Information disclosure has both benefits and risks and thus involves a complicated calculation and informed decision making about boundary opening or closure (Seounmi 2009). Therefore, one might care deeply about privacy in general but, depend on the costs and benefits prevailing in a specific situation, seek or not seek privacy protection (Timothy and Susan 2002). Rogers (1975, 1983) states that perceived benefits associated with risky
behavior weaken individuals’ intention to protect themselves from risks. Westin (1997) states that customers usually consider the nature of the benefit being offered in exchange for information. People are willing to share information if they receive some types of benefit from the disclosure (Milne and Mary 1993).
In addition, this study also assumes that perceived benefits reduce the impact of privacy concerns over disclosing information. Prior studies (Kishalay and Rajeev 2018; Stevenet al. 2005) indicated that there is a negative relationship between customers’ perceived benefits and their privacy concerns. Phelps et al. (2000) suggested that privacy concerns could be alleviated if companies offer benefits in exchange for personal details. Nowak and Phelps (1997) found that in most cases, personal information was willingly supplied in expectation of such future benefits as reduced prices, premiums, or other
incentives.
2.4 Willingness to Share Personal Information
The association between privacy concerns and protection behavior was first explored by Altman (1975, p. 50), suggesting that “People attempt to implement desired levels of privacy by using behavioral mechanisms”. A review of literature confirms a significant negative relationship between individuals’ privacy concerns and their willingness to share information (Steven et al. 2005;David et al. 2012; Gaurav et al. 2016; Xie et al. 2006; Sheehan and Mariea 1999; Lwin and Jerome 2003; Lwin et al. 2007). The stronger a customer’s concerns are about companies’ information collection practices, the more likely the customer to refuse to give information. Although prior research has found that most customers have refused to give firms some personal information at one time or another due to privacy concerns, few studies have delved deeper into customers’ willingness to provide different types of personal information to firms (Mothersbaugh et al. 2012; Joseph et al. 2000). Aside from assessing issues related to financial and credit data, much research has simply assessed customers’ overall willingness to share personal information (Wasserman et al. 1994).
The primary source of customer privacy concerns revolves around personal or individual-specific information which includes data such as names, addresses, demographic characteristics, lifestyle interests, shopping preferences, and purchase histories of identifiable individuals (Nowak and Joseph 1995). Previous research (Joseph et al. 2000) suggests that most of the customer personal information targeted by firms fall into the five broad categories of demographic characteristics, lifestyle characteristics (including media habits), financial information, personal identifiers (e.g., names, addresses) and shopping/purchasing habits. More important, the categories wary in the degree to which each evokes customer privacy concerns (Miller 1997; Nowak and Joseph 1992). Few studies that have considered
information categories suggest that customers are more reluctant to share financial data and personal identifiers and are more willing to share demographic and lifestyle information (Nowak and Joseph 1992; Wang andPetrison 1993; Vidmar and David 1985). Similarly, prior
study (Mothersbaugh et al. 2012) also demonstrated that online privacy concern had a
substantial negative effect on willingness to disclosure when contact and the general financial information is requested, but not whenmedia usage and website perceptions are requested, which canceled out the effect of online privacy concern when assessed relative to overall disclosure. After an overview of previous research on privacy issues (Mothersbaugh et al., 2012; Joseph et al. 2000), four types of personal information (showed in Table 1) are summarized and included in the conceptual model of this study in order to assess customers’ willingness to share personal information.
Table 1. Types of Personal Information Used to Assess Customer Willingness to Do Disclosure
Information Types Items
Contact email, phone number, address General financial home value, range yearly income
Demographic first/last name, birth year, education level, age
Lifestyle use of products/brands, freq online purchase, leisure time
Previous research on the same area point to a number of additional factors that should be included because of their potential influence on dependent and mediating variables in the research model (Kelly and Patrick 2017). Customers’ perceived risks, perceived benefits and privacy concerns may also depend on individual characteristics (Yang and Peterson 2014). So, with respect to individual characteristics, the model includes age, gender and education as covariates.
2.5 Customer Purchasing Journey
One of the most-cited and widely known consumer purchasing journey models is the five-stage purchasing journey (Julia and Nipawan 2014). During the purchasing journey, customers go through five stages, including need recognition, information searches,
called together the pre-purchase phase of the customer journey (Katherine andPeter 2016). This research aims to investigate the existence of customer segments incorporating multiple stages of the consumer journey—especially information search and purchase. Since the usage of postpurchase services appears to be rare in some of the chosen channels (Umut et al. 2008), the postpurchase stage is excludedin this research.
2.5.1 Pre-purchase Stage
The pre-purchase stage contains all aspects of the customer’s interaction with the retailer, category, and environment before a purchase transaction. The pre-purchase stage of customer buying journey is complex as it involves a composite set of factors and activities (Rodoula and Jochen 2015). At the pre-purchase stage, the consumer’s goal is not necessarily limited to finding information for a specific consumption (Milner and
Rosenstreich 2013). Instead, consumers may interact with brands from different channels based on their purpose and end with buying nothing. For example, in the absence of subsequent actual shopping behavior, customers may like to provide email to brands to be notified when the products which currently are out of stock become available or when brands have new in. Also, consumers may register membership to be able to enjoy purchase
discounts at some point in the future. Sign-up usually requires customers to provide personal information including name, email, gender, date of birth and a few other details, no matter they sign up online or offline. After gaining a membership, customers usually provide their membership cards or membership number at each purchase to get a discount on the current purchase or to accumulate spending so as to be able to enjoy the discount in the future. In this way, membership programs enable companies to capture data at the point of sale and build a picture of individual customers’ spending patterns and preferences.
Also, even without any purchase intention, consumers may follow the brand on social media and interact with the brand or other customers who also follow the same brand on a social media platform.After following the brand on social media, consumers can not only obtain new product information and new promotional activity information timely but also interact with others who also follow the same brand (Rubathee and Rashad 2014; Sonja et al.
2013Sonja et al. 2013). Unlike other channels, using social media channels requires
consumers first to have a social media account. To get such an account the customer needs to provide some personal information for registration (Alessandro and Ralph 2006). Information required for registering social media platforms typically includes email, name, mobile
number, age and so on.
When consumers have the desire to shop, they usually have multiple channels to use to search for product information. If consumers decide to go to physical stores to gather information on products, they can personally experience one product before a purchase decision is made. From the combination of touch, sight, smell, and hearing, consumers are allowed to fully understand the products’ appearances, properties, and characteristics. Especially for clothing selecting process, product selection usually needs a multi-angle evaluation (Xu and Chen 2017). Most important, searching product information in a physical store barely sacrifice customer personal information. However, this can barely be achieved if customers do pre-purchase information search through other channels. For example, searching product information on the retailer’s website and/or brand-owned mobile app does not require customers to provide personal information for registration. However, customer data can also be compromised because modern Web technologies have made it possible for advertisers to track individuals’ online activities. With third-party cookies, the retailer not only can identify individual customers' machines but also can collect information about customers’ online activities, including which product clicked and viewed for how long (Alessandro and Ralph 2006). In general, consumers will compromise their personal information to varying degrees when searching information on the nontraditional channel.
2.5.2 Purchase Stage
The purchase stage covers all customer interactions with the retailer and its environment during the purchase event itself ,and it is characterized by behaviors such as choice, ordering, and payment (Katherine and Peter 2016). Customers today have more channel choices at each purchasing stage than previous generations. What an omnichannel retailer brings to the customer is the opportunity for customers to undertake the purchasing
process on their terms and at their convenience.
Some research pointed out that online shores sell products generally at lower prices than physical stores do (Hsiao 2009). However, it is not the case in the omnichannel retailing context in which the retailer’s market approach is identical across different channels with regard to assortments, pricing, service policies, communication policies and store branding (Peter et al. 2015). In some cases, the cost of the same product bought in some channels will be higher than that in other channels. For example, website or mobile stores explicitly add the delivery cost onto the price; the ultimate selling price is then higher than in physical stores.
When shopping in nontraditional channels, there exists transaction risk, which is the uncertainty associated with giving information (such as credit card number, name, address, etc.) to the retailer during a transaction. Prior research shows the most worrying problem with online purchases is about transaction (Ming-Hsiung 2009) andtransaction security is one of the significant factors affecting customers’ willingness to shop online (Liao and Cheung 2001; Chen et al. 2004; Luarn and Lin 2004). Consumers mostly pay by credit cards for their online purchases, and this inevitably raises concern about transaction security problem (Ming-Hsiung 2009). However, this kind of perceived risk was never that acute in traditional store settings. The perceived transaction risk is higher in nontraditional channel shopping since the information is sent to a distant recipient over a public network, and the shopper is likely to perceive enhanced uncertainty regarding any possible misuse of the information given out (Steven et al. 2005). Although when shopping in a physical store, customers may also put their privacy at risk by paying with a credit card, and in this way customers also leave their credit card information to the brand they interact with. However, when shopping in the traditional channel customers can choose among various payment methods, such as cash, saving cards, credit cards, or e-commerce payment, and they can always choose the payment method that they consider to be the safest and worry least about their privacy. This is
impossible to achieve in nontraditional channels, in which customers always leave their payment information.
After paying for the product, customers can receive it immediately if they purchase it at a physical store. By shopping through nontraditional channels, however, customers will
have to wait for the product delivery. In such a case, customers have to provide some personal information including name, home address, phone number for the delivery such the package will be delivered to the right person and the right place.
2.6 Retailing Channels
In this study, channel is defined as a customer contact point or a medium through which the company and the customer interact (Neslin et al. 2006). Retailing has changed rapidly in the past two decades due to the advent of the online channel and new digital and specifically mobile channels (Peter et al. 2015; Rigby 2011). The use of mobile channels and specifically mobile apps has significantly changed the way how individuals to purchase (Xu et al. 2014). With the explosion of mobile technologies, customers today can choose where, when and how they interact with a retailer and expect the same retailer to deliver a relevant and seamless experience between channels (Avery et al. 2012; Balasubramanian et al. 2002). When shopping via multiple channels is a rapidly growing phenomenon among customers (Ansari et al. 2008; Coughlan et al. 2006; Geyskens et al. 2002), retailers are rapidly adapting omni-channel strategies to respond to continual changes in individual shopping behaviour in the wish to be present in as many channels as possible to give customers the ability to interact, socialise and purchase where and when they want.
Although in omnichannel retailing, retailers offer the customers all channels that currently widespread (Anderson et al. 2010), it is unrealistic to include all these channels in one research to investigate customer channel usage. From the practical consideration, this study will focus on four channels, which are physical stores, website shop, mobile channel and social media, to study customers’ channel usage in information search and purchase stages. These channels were selected for two main reasons:they are available to consumers for both search and/or purchase, and they are available in the context of clothing retailing that this study aims to focus.
2.6.1 Physical Stores
physical store or offline channel is firmly planted in the physical world. When compared to online shopping, physical store shopping is generally considered time-consuming and inconvenient, as individuals have to go in person to the shopping place and may risk time wasted in queuing or waiting to check out. However, due to privacy and security issues continue to be an important barrier to the adoption of e-shopping, many people today still prefer doing business with a physical store (Mokhtarian 2004). Compared to other forms of shopping, traditional store shopping still quite competitive in some ways. First, compared to shopping in other channels, customers can acquire a desired good more quickly (Rajammaet al. 2007; Cuneyt and Gautam 2004). In the context of clothing shopping, after paying for the product, consumers can receive it immediately if they purchase it at a physical store. By e-shopping, however, consumers will have to wait for the product delivery. Secondly, traditional store shopping is perceived as less risky. Mail order has been considered to be riskier than in-store purchasing and users of the internet encounter more risks than they do in face-to-face transactions (Sally 2006). Much research suggests that compared to offline shopping, there are a number of risksassociated with e-commerce such as credit-card fraud (Sally 2006), release of private information and lack of any guarantee of quality of goods (Cuneyt and Gautam 2004).These risks prevent some consumers from buying any items from the Internet. Thirdly, physical stores also provide consumers with the real products
experience, an intuitional shopping space and environment and the real service from the shop assistants (Xu and Chen 2017).
2.6.2 Website
With the aid of information and communication technologies (Mpofu and Staden 2015), customers can shop via the Internet, which has dramatically changed the way people shop. Online shopping, also names like e-shopping, network shopping, Internet shopping, or Web-based shopping, featuring in freeing customers from having to visit a physical store personally. Prior research suggested that the internet is a more efficient shopping channel than traditional offline channels (Xu and Chen 2017; Yoo-Kyoung and Marjorie 2007; Ming-Hsiung 2009). The internet offers diverse kinds of convenience to consumers for them to
search for information, evaluate and purchase products more efficiently and effectively than other channels to satisfy their needs (Byeong-Joon 2004).The internet can also enhance customer efficiency by facilitating customers’ access to consumption related information more quickly and save time, effort, and monetary cost for information collection (Couclelis, 2004; Farag et al. 2007). In addtion, by providing the mixture of sound, image, text, and visual tools, the internet can improve consumer effectiveness, enhance consumer learning and help consumers choose products that can best satisfy their needs (Byeong-Joon 2004).
However, online transactions also pose threats that consumers need to be vigilant of, such as the placement of cookies, hacking into hard drives, intercepting transactions, and observing online behavior via spyware (Cohen 2001). In some cases, online identity theft could gain access to individuals’ credit card data and steal thousands of dollars from customers (New York Times 2003). Although such risks could also occur in the physical world, it can be easier and more efficient for thieves to collect data online(Katyal 2001). Milne et al. (2004) suggested that individuals who interact with retailers online are vulnerable in three general ways: (1) the data transfer to an online seller may be compromised, (2) the data stored by the retailer may be compromised, (3) the data on their computer may be compromised. When consumers give credit card and personal information to a retailer’s website, such information can be intercepted if the transfer is not encrypted using SSL (secure socket layer) protocols (Milne et al., 2004). Another threat to consumer privacy occurs after a company obtains consumer data. In some cases, companies have not kept their promises not to share the data with third parties. When individuals are connected to the Internet,
information on their computers can be increasingly vulnerable to intrusions and theft if a firewall is not installed to protect them from hacking (Milne et al.2004).
2.6.3 Mobile Channel
Smartphones have penetrated rapidly, and mobile shopping provides promising market opportunities for retailers. With the prevalence of mobile phones, the mobile channel has become the third marketplace, following the offline and online channels (Kim et al. 2017); With the growth of mobile devices, brands participate in this rapidly evolving channel by
offering customers new applications (apps) designed explicitly for mobile devices. Mobile apps in many cases are the mobile versions of online sites and brands usually design and launch apps similar to their online sites when expanding their business to the mobile platform (Bang et al. 2013). People browser the web on their mobile devices and use various mobile applications (Dhruv et al. 2016), and this means customers who do business with companies through mobile channels may be also vulnerable in same ways as customers who interact with brands online. Internet-capable mobile devices have changed the way people communicate, interact and take advantage of the Internet, allowing them to access the Web whenever they want and wherever they are (Margherita 2017). Mobile devices are highly individualised, and important personal communication tools (Bacile, Ye, and Swilley 2014), and consumers use their smartphones to conduct a host of activities, incuding talking or texting (Dhruv et al. 2016). Mobile devices have truly enabled consumers' ubiquitous access to digital information, anytime and anywhere. However, the ubiquity of mobile devices and their always-on nature make customers more trackable.
An important feature that is unique to mobile devices is their ability to support location-based applications. Customers often use apps for quick access to location-based information, such as the nearest highly rated restaurant (Grewal and Levy 2016). Location-based apps enable a user to receive relevant information while away from home, such as movie times at the local theater, price comparisons across local stores, or directions to a particular local outlet. To be able to offer such location-based services, mobile devices include sensors that can identify each user's context, including her or his exact location (e.g., via GPS, WiFi, beacons) and viewing direction (e.g., compass, built-in camera) (Dhruv et al., 2016)). Such knowledge makes it possible to track consumers' location and movements and then forecast the movements of consumers (Bellovin et al., 2014).
2.6.4 Social Media
At the most basic level, social media is an Internet-based community where individuals interact, often through profiles that (re)present their public persona (and their networks of connections) to others (Alessandro and Ralph 2006). Existing research shows that
individuals are spending 25% of their Internet time on social media, up from 15% in 2009 (Nielsen 2010). Individuals use social media to interact with friends, view photos and videos, and find businesses and brands (Adam et al. 2013). As social media experienced exponential growth in users in recent years (Alessandro and Ralph 2006),brands start to leverage social media not only for digital advertising and promotions but also to handle customer service issues, mine innovation ideas and authentically engage with customers (Solis 2010). Social media is changing the retailing landscape and redefining how retailers communicate across their channels of distribution and with their customers (Adam et al. 2013).At the same time, customers are now using social media in most of the different stages of the purchasing
process: when searching product or service information, when making the decision to buy and when they are encountering any issues and are in need of customer service. Like the mobile channel, social media also offers users a dynamic, ubiquitous, and often real-time interaction with brands (Sonja et al. 2013). By following a brand on social media, users can be updated from time to time on any special events, contests or a new promotion organized by the brand or product (Rubathee and Rashad 2014). Consumers, in particular, are more empowered by social media, as these technologies enable consumers to share their brand stories easily and widely with peers. Consumer-generated brand stories shared through social media are much more influential than stories spread through traditional channels because they utilize social networks, are digital, visible, ubiquitous, available in real-time, and dynamic (Hennig-Thurau et al. 2010). Customers today have gained a more powerful voice that brands can no longer afford to ignore, while in the past such voices were not strong and could be safely ignored by brand managers (Sonja et al. 2013).
Though social media has enormous benefits for consumers, at the same time it also makes users vulnerable to privacy and data security issues. Users must register before using the site. In order to create such an account, individuals need to provide some information about themselves, such as name,phone number, gender and email (Alessandro and Ralph 2006). Although information requested by social media platforms may differ in some ways, essentially they work using the same principles. Social media platforms’ security and access controls are weak by design to leverage their value as network goods and enhance their
growth by making registration, access, and sharing of information uncomplicated. At the same time, the costs of mining and storing data continue to decline (Alessandro and Ralph 2006). Information provided by consumers on social media can be stored and searched for by an unlimited amount of people (Lis and Korchmar 2013), which multiplies the spread of information (Mangold and Faulds 2009). Social media also permit the brands to collect, register, analyse and use customer data and feedback for better targeting users and
customizing its messages (Gurau 2008). Due to its ease way of getting access to information, lack of monitoring and control, undoubtedly social media favors many risks and cybercrimes (Rubathee and Rashad 2013). In recent years, new social media applications designed explicitly for mobile devices, such as location-based social media applications, have been gaining notoriety. These applications combine location specificity and interactivity, allowing users to connect with others based on their current locations. When customers “check in” to businesses or addresses on social media apps, they instantly share the information with friends and connections also on social medias. So, whenever a user checks in with his/her application, he/she makes his/her position potentially identifiable to everyone (Xuet al. 2009).
2.7 Customer Segments
The retailing landscape has changed dramatically in the past decade. The advent of the online channel and new additional digital channels such as mobile channels and social media have transformed not only retail practices but also consumers’ shopping processes, (Peter et al. 2015). With the dawn of the mobile channel, tablets, social media, customers can choose where, when and how they interact with a brand and expect retailers to deliver a relevant and seamless experience between channels. In order to offer customers more choice and
flexibility, retailers are rapidly adapting their businesses and operational infrastructures to support these changing customer needs and shopping behaviors, particularly for the
significant proportion of customers using multiple channels to browse and buy. As a result, many retailers have initiated omnichannel strategies which are characterized by the integrated and seamless experience using all channels.
customers no longer interact with retailers in only one channel. Instead, the Internet, mobile devices, call centers, direct marketing, home shopping networks, and catalogs, as well as bricks-and-mortar stores, are now commonplace means by which consumers shop. More important, customers have multiple channel choices and can freely use each of them
throughout the buying journey. As a result, customers display complex shopping behaviors in the emerging omnichannel environment. Some consumers may use one channel to perform all shopping activities during the journey, while others may rely on different channels at different stages of shopping. Then, a single channel may reappear during the journey multiple times, while different channels may also be used simultaneously at one shopping stage (Julia and Nipawan 2014). For instance, in the omnichannel setting showrooming is becoming an important issue. Shoppers now frequently search for information in the store and
simultaneously search on their mobile device to get more information about offers and may find more attractive prices (Peter et al. 2015). The opposite of showrooming also occurs, which is now referred to as webrooming, where shoppers seek information online and buy offline (Verhoef et al. 2007).
So, in the omnichannel environment, several customer segments based on their channel usage at each stage of the buying journey. First, a single channel user group may exist. Single channel shoppers are customers who use only one channel throughout their purchasing journey. Behaviorally, customers in this segment use the same channel at each purchasing stage. Offline/in-store customers, web channel customers and mobile channel customers all belong to this segment. Second, both webrooming shopper segment and the showroom shopper segment can also be expected. Webrooming shoppers are customers who search product information on the website and/or the mobile device but purchase at a physical store. Showrooming shoppers are customers who are more likely to search product
information at a physical store but decide to purchase on the website or the mobile device. Last, there may be an omnichannel shopper segment characterized by frequently using all the channels included in this study in both information search and purchase stages. In their research, Kumar and Venkatesan (2005) suggested there is a customer segment that employs multiple channels for either search or purchase. Consumers in this segment represent an
increasingly large proportion of consumers (Rangaswamy and van Bruggen 2005; Verhoef et al. 2007).
2.8 Conceptual Model
Many theoretical models have been proposed and tested in the past to understand the privacy concerns in different contexts (Kim and Mariea 2000; Gaurav et al. 2016; Steven et al. 2005; Joseph et al. 2000). Drew on previous research, this study proposes a conceptual model including determinants of the willingness to share personal information with a seller, as represented in Figure 1. Based on literature overview, this study adopts a benefit–cost
framework. Specifically, following the approach of a consumer calculus, this study assumes that customers face a cost–benefit trade-off when they decide to engage in interactions which entail the disclosure of personal information (Dinev and Paul 2006; Smithet al. 2011; Xie et al. 2006; Zhao
et al. 2012
). The use of a “utility maximization framework” follows the idea of the privacy calculus by creating a function that confronts costs and benefits (Award and Krishnan 2006; Rustet al. 2002
).In this study, benefits refer to customers perceived benefits which relate to time saving, convenience, location-based service that customers may derive from using a specific channel which usually requires customers give up their privacy to some degree, while costs are considered as perceived risks, including the potential adverse outcomes associated with personal information requested. While it can be aware that the chosen factors do not cover the full range of costs and benefits, they represent the majority factors frequently investigated in prior research and, to the best of our knowledge, have never been studied together in the context of omnichannel clothing retailing.
In the conceptual model, perceived risks and perceived benefits are directly linked to willingness to share personal information, where this study assumes that these benefits and risks influence a latent underlying utility (Manfred et al. 2017). This approach is prevalent in marketing when studying purchase decisions (Umut et al. 2008; Rustet al. 2004). Moreover, an indirect effect of perceived risks and perceived benefits on the willingness to share personal information through privacy concerns are expected. Following this, the proposed
hypotheses aim to find the impact of perceived risks, perceived benefits and privacy concerns on the willingness to share information in omnichannel clothing retailing. According to the proposed hypotheses, perceived risks, perceivedbenefits and privacy concerns are the factors that influence a customer’s willingness to share personal information in the context of clothing retailing.
H1: Perceived risks are positively related to privacy concerns. H2: Perceived benefits are negatively related to privacy concerns.
H3: Privacy concerns are negatively related to willingness to share personal information. H4: Perceived risks are negatively related to willingness to share personal information. H5: Perceived benefits are positively related to willingness to share personal information.
In their study, Umut et al. (2008) summarized relevant research which investigated segmentation based on customers’ channel usage. Their study shows that there is a customer segment that employs multiple channels for either information search or purchase (Kumar and Venkatesan 2005) and customers in this segment represent an increasingly large proportion of customers (Rangaswamy and van Bruggen 2005; Verhoef et al., 2007). Customers who shop across multiple channels tend to transact in higher volumes (Blattberg, Kim, and Neslin 2008; Neslin et al. 2006), have higher past customer value, a higher share of wallet, and are more likely to be active than other customers (Kumar and Venkatesan 2005). Although the results of prior research are mixed but generally lean toward the finding that retailers offering an array of delivery channels can increase customer satisfaction and loyalty (Patrali 2010) because more channels suggest better service, which often leads to higher loyalty. Similarly, Wallaceet al. (2004) found that multichannel usage is associated with higher perceptions of the firm’s channel offerings, which in turn are associated with higher customer satisfaction and greater loyalty. Research on whether a multichannel strategy grows sales has shown that multichannel customers spend more than do single-channel customers. Kumar and
Venkatesan (2005) found that customers who use mutiple channels durinng their buying process buy more; and Myers et al., (2004) reported that “multichannel customers spend 20 to
30 percent more money, on average, than single-channel ones do.” Kushwaha and Shankar (2005) reported that multichannel shoppers buy more often, more items, and spend more than single channel shoppers. Kumar and Venkatesan (2005) indicated that multichannel
customers received more contact from the company through a variety of channels.
While the studies mentioned above have contributed significantly to our understanding of channel choice, there is no study investigating customer segmentation, incorporating customer privacy concerns and willingness to share personal information. Consumers who shop across multiple channels differ significantly from those who shop across some of the available channels (V. Kumar and Rajkumar 2005). Compared to customers who shop through a single channel, customers who frequently use multiple channels for their purchase may be familiar with the retailer and have deeper relationships with the seller and have greater trust and lower perceived risk in their interactions that could motivate them to spend more with the brand. Also, the depth of a relationship and trust seem to increase as the customers start increasing the number of channels through which they interact with a seller (V. Kumar and Rajkumar 2005). Hence, one can reasonably expect that compared to customers who shop through a single channel, customers who use multiple channels in information search and purchase stages may be less concerned about their privacy and more willing to share personal information with the retailer.
Figure 1. Proposed research model
3. Research Design and Methodology
SPSS 25.0 statistical software was used for data analysis to test the proposed assumptions.
3.1 Sampling and Data Collection
The data was collected by means of an online survey administered by Qualtrics, a digital data collection provider. The respondents of this study are students still under higher education, who are thought to have experience on omnichannel clothing shopping. Subject to their own experience, they freely fill in the questionnaire. The reasons for choosing higher education students as respondents include: first, they are familiar with the computer, mobile,
Customer Segments Archetypes: Mainly Offline Mainly Online/Mobile Webrooming/Showrooming Omnichannel Perceived risks Perceived benefits Privacy Concerns
The willingness to share personal information:
Contact information (e.g., email, phone number, address)
Financial information (e.g., home value, range yearly income)
Demographic information (e.g., first/last name, birth year, education level, age)
Lifestyle (e.g., use of products/brands, freq online purchase, leisure time) H1
H2
H3 H4
and social media knowledge, and it is easier for them to use these channels for
communication and transaction. Second, omnichannel shoppers are mainly young people with good education, and they will become the primary force of omnichannel retailing. Third, students under higher education are more adventurous, willing to accept new things, and more receptive to omnichannel shopping. Respondents were contacted via course websites, and the purpose of the survey was explained to them. Respondents were given a web link to the survey and asked to proceed to it to fill it out.
3.2 Measurement
The survey instrument was composed of items designed to learn about the respondents’ use of four channels for searching information and three channels for purchasing. Four items are used to collect frequencies of searching information in four different channels (physical stores, computer-based websites, mobile apps and social media), and three items are used for collect frequencies of purchasing product in three different channels (physical stores, computer-based websites, and mobile apps). In order to anchor the respondents in recalling their shopping behaviors, they will be asked to think about their shopping over the last 12 months, both information searching and actual purchasing. The respondents are asked to indicate how often they use each type of channels to search for information and purchase. Surveyee’s channel usage will be measured on a 7-point scale, with response options ranging from 1 (0% never) to 7 (100% every time).
The measurement scales intended to represent the constructs of perceived risks, perceived benefits and privacy concerns in the research model were derived from previously-validated scales. Where necessary, a few previous scales were adjusted upwards slightly to cope with the nature of the research. The three were measured on a 7-point scale, with response options ranging from 1 (0% Strongly disagree) to 7 (100% Strongly agree). The reliability of three variables in the conceptual model (Cronbach’s coefficient) and their sources are shown in Table 2.
Table 2. Measurement scale sources and Cronbach’s Alpha
Scale Source Cronbach’s Alpha
Perceived risks Dinev & Hart (2006a); Malhotra et al. (2004a) .925 Perceived
benefits
Ko et al., (2009); Lu & Yu-Jen Su, (2009);
Mathwick et al., 2001; Kleijnen et al., (2007); Lee & Jun, (2007)
.927
Privacy concerns Dinev & Hart (2004); Low (2005), Larose & Rifon (2007); Crossler (2010)
.904
Reliability was implemented to examine the consistency of measurements. Reliability checks were run for perceived risks, perceived benefits and customers’ privacy concerns. The Cronbach’s alpha, which represents the estimator of the internal consistency, has been tested to verify if all the items in one scale measure the same, or if some questions should not be used for analysis. As exhibited in Table 1, the results support the internal consistency of each scale, as all Cronbach’s Alpha is equal to or higher than .900. In this study, the gender, age, and education level of respondents were treated as control variables in order to control the impact of these variables on the results of this study.
The measure of willingness to share personal information draws on that used by David et al. (2000). To capture the willingness to disclose personal information, four types of information (and their underlying items) were measured on a 7-point scale, with response options ranging from 1 (0% likely to reveal) to 7 (100% likely to reveal). Also, respondents’ willingness to provide personal information was measured combined with the goal of providing such type of personal information. For example, the contact information question asked ‘‘How likely are you to provide your contact information (e.g., email, address, phone number) to brands or retailers that you frequently consider for your clothing purchases to purchase items from their website?’’ Note that the question itself provides the specific items in that category as examples of the type of information requested. This study summarized ten different goals (showed in Table 3) associated with customers’ personal information
representing the average of the four types of information.
Table 3. Ten different goals associated with individuals’ personal information disclosure
behaviors
1 to purchase items from clothing brands’ website 2 to purchase items from clothing brands’ store 3 to purchase items through clothing brands’ app 4 download clothing brands’ app
5 get access to discounts
6 participate in loyalty programs 7 to subscribe to the newsletter
8 to get access to a personal on clothing brands’ corporate website 9 to obtain gifts and samples
10 to be informed about exclusive events (e.g. private sales, new collection, etc.)
4. Results and Discussion
4.1 Sample Characteristics
A total of 203 questionnaires were collected for this study, and incomplete and unusable (n = 60) questionnaires were excluded from the analysis. As a result, 143 valid questionnaires (70.4% of the questionnaires were qualified) were obtained for data analysis. The demographic characteristics of the respondents ( showed in Table 4) were as follows: among 18-34-year-old 58.7 percent of men and 41.3 percent of women.; Bachelors and Masters accounted for 39.2% and 31.5%, followed by College/HBO (16.8%) and the rest were High school (7.7%) and Post Graduate (4.9%).
Table 4. Sample characteristics Characteristics Frequency % Gender Male 84 58.7 Female 59 41.3 Total 143 100.0 Education
High school degree 11 7.7
College/HBO degree 24 16.8
University Bachelor’s degree 57 39.2 University Master's degree 45 31.5
Post Graduate degree 7 4.9
Total 143 100.0
4.2 Segmentation
For customer segmentation, self-report channel usage at both information search stage and purchase stage was used as the subjective segmentation variable. More specifically, respondents were segmented based on their use of four selected channels at the information search stage and their use of three selected channels at the purchase stage. Therefore, the segmentation criteria include seven indicators. Two different clustering techniques were applied to classify survey participants into homogeneous groups according to their channel usage at two mentioned stages: hierarchical clustering and k-means. First, a hierarchical clustering (Ward’s method) was performed to define the number of clusters: apply Ward’s method on the seven dimensions, check the agglomeration schedule, identify the step where the “distance coefficients” makes a more significant jump and decide the number of clusters. Then the K-means procedure was used actually to form the clusters. For the metric response of privacy concerns, One-way ANOVA was used, and post-hoc effects (the Tukey test) were estimated, affording analysis of the significant differences among the segments, while both one-way ANOVA and the hierarchical multiple regression analysis were implemented to investigate the significant differences among the segments in terms of the willingness to share personal information.
to be the more reliable model. Also, the four-and five-cluster solutions were tested in detail; however in these cases, there was a big difference concerning the segment’s size as well as underrepresented channels in various groups.To prove the reliability of the three-cluster solution, a statistical test was conducted. Hierarchical cluster methods provide only
insufficient guidance for making a decision on the number of clusters. The only meaningful indicator relates to the distances at which the objects are combined; this is also known as elbow criterion (Mooi and Sarstedt, 2011). Using this method, which is based on the ratio of total within-group variance to between group variance plotted against the number of
clusters(Malhotra and Birks, 2007), it was found that the percentage of variance explained declines substantially after there are three clusters, hence suggests that the optimal number of groups should be three. The statistical test confirmed the practical consideration; therefore, the three-cluster solution was chosen. The clustering results are shown in Table 5A. As Table 5A shows, three segments could be identified.
To derive the channel usage scores of each segment (
U
icp), the value in Table 6A(
V
icp) were first averaged over both segments (Ū
i) and channel at each stage (Ū
cp). Then, thechannel usage scores for each segment (
U
icp) were calculated using the equation shownbelow.
U
icp =(
V
icp/Ū
i×
V
icp/Ū
cp)
×100
Table 5A. Customer segments: Final Cluster Centers
Information search Purchase
Website Mobile Social Media
Physical
Store Website Mobile
Physical Store Segment 1 3.98 4.17 3.83 3.95 3.67 4.71 3.89 Segment 2 5.79 5.34 5.07 5.69 5.45 5.45 5.55 Segment 3 3.10 2.00 2.63 4.27 2.86 2.33 4.86
Table 5B shows three segments’ channel usage scores. The segments were named according to the usage of the different channels. Respondents of the first segment, called “Mobile-oriented shoppers”, most likely search for information and buy clothes through the mobile channel. Similarly, physical store-oriented shoppers most likely search for information and buy products at physical stores. Different from respondents in these two segments, respondents identified as Omnichannel shoppers frequently use all selected channels at both information search and purchase stages.
The segments have different sizes: Mobile-oriented shoppers made up the largest part of the
U
icpSegment i’s channel usage score of channel c for phase p of the buying process
V
icpSegment i’s value of channel c for phase p of the buying process in Table 5A
Ū
iAveraged value over segment i in Table 5A
Ū
cpAveraged value over channel c for phase p of the buying process in Table 5A
i = 1, 2, or 3 signifies a segment 1, segment 2,
or segment 3 respectively;
c = 1, 2, 3, or 4 signifies website, mobile, social
media or physical store, respectively;
p = 1 or 2 signifies information search or
purchase, respectively.
Table 5B. Customer segments: Channel Usage Scores
Information search Purchase
Cluster Website Mobile Social Media
Physical
Store Website Mobile
Physical Store Number of Cases %
Segment 1 92 113 95 84 84 132 79 63 44% Mobile-oriented shoppers
Segment 2 143 136 122 127 136 130 118 29 20% Omnichannel shoppers Segment 3
71 33 57 125 65 41 157 51 36% Physical store-oriented
sample (44% in total). Respondents who frequently go to physical stores for searching product information and purchasing had a representation of 36%. Omnichannel shoppers (20%) have the smallest proportion. The two largest segments, which represent 80% of the sample, show that most consumers use one channel primarily in information search and purchase stages, although clothing brands deploy multiple channels.
4.3 Demographic Profile of the Segments
To describe the segments’ demographic structures in terms of gender and education, cross-tabulation was employed to determine the significant differences that exist among the segments, and χ2 -test was used to test the difference the segments. In addition, the average age was calculated for each segment. Table 6 contains the analysis of demographic factors. In terms of average age, there is no significant differences among the segments. Although the proportion of male respondents in mobile-oriented segment was higher than the average, there was no significant difference between the segments in terms of gender (χ2= 4.023, p=0.134).
The educational level was also not significant (χ2= 10.717, p=0.218)
Table 6. Demographic Profile of the Segments
Mobile-oriented shoppers Omnichannel shoppers Physical store-oriented shoppers Gender % Male 66.7% 44.8% 56.9% Female 33.3% 55.2% 43.1% Education %
High school degree 11.1% 3.4% 5.9%
College/HBO degree 22.2% 10.3% 13.7%
University Bachelor’s degree 31.7% 34.5% 51.0% University Master's degree 30.2% 48.3% 23.5%
Post Graduate degree 4.8% 3.4% 5.9%
4.4 Hypotheses Testing
4.4.1 Correlation Analysis
First, the mean calculation was implemented for variable perceived risks, perceived benefits, privacy concerns and willingness to share personal information. The value of overall willingness to share personal information was calculated as the mean of the willingness to share four types of personal information. Then a correlation analysis was carried out to investigate these variables and their interaction. The mean, standard deviation and
correlations of the study variables are provided in Table 7. Table 7 shows that the perceived risk was strongly and positively correlated with privacy concerns (0.715, p < 0.01), while the perceived risk was negatively corrected with respondents’ overall willingness to share personal information (-0.274, p < 0.01). The perceived benefit was negatively related to privacy concerns (-0.253, p < 0.01) and also strongly and positively related to overall willingness to share personal information (0.531, p <0.01). Last, there was a negative correlation between privacy concerns and respondents’ overall willingness to share personal information (-0.249, p < 0.01).
Table 7. Means, Standard Deviations, Correlations
Variables Mean SD Perceived risks Perceived benefits concerns Privacy Willingness to share personal information Willingness to share contact information Willingness to share financial information Willingness to share demographic information Willingness to share lifestyle information Perceived risks 3.262 1.262 1 Perceived benefits 4.693 1.206 -0.312** 1 Privacy concerns 3.221 1.330 0.715** -0.253** 1
Overall willingness to share personal
information 3.986 1.172 -0.274** 0.531** -0.249** 1
Willingness to share contact
information 4.376 1.174 -0.278** 0.469** -0.246** 0.760** 1
Willingness to share general financial
information 3.276 1.731 -0.179* 0.405** -0.172* 0.757** 0.419** 1
Willingness to share demographic
information 3.977 1.558 -0.158 0.395** -0.134 0.854** 0.537** 0.507** 1
Willingness to share lifestyle
information 4.314 1.436 -0.281** 0.435** -0.260** 0.806** 0.577** 0.376** 0.653** 1
**. Significant at the 0.01 level (2-tailed) *. Significant at the 0.05 level (2-tailed)
4.4.2 Regression Analysis
First, PROCESS v 2.16.3 by Andrew F. Hayes was used to do the Sobel test in order to test if that the level of privacy concerns would mediate the effects of perceived risks and perceived benefits on respondents’ willingness to share personal information. The results are shown in Table 8A-B. As can be seen in Table 8A-B, both the direct effect and the indirect effect of perceived risks and perceived benefits on the willingness to share personal information were not statistically significant.
Table 8A. Direct and Indirect Effects of X (Perceived Risks) on Y
Y Effect SE/Boot SE p LLCI/Boot LLCI ULCI/Boot ULCI Total effect -0.260 0.075 0.001 -0.408 -0.113 Direct effect -0.191 0.107 0.077 -0.404 0.021
Overall
willingness to share personal
information Indirect effect -0.069 0.078 0.372 -0.230 0.099
Total effect -0.263 0.076 0.001 -0.413 -0.113 Direct effect -0.202 0.109 0.067 -0.417 0.014 Willingness to
share contact
information Indirect effect -0.062 0.079 0.434 -0.238 0.114
Total effect -0.250 0.112 0.028 -0.472 -0.027 Direct effect -0.166 0.162 0.313 -0.483 0.156 Willingness to
share general
financial
information Indirect effect -0.086 0.117 0.460 -0.291 0.200
Total effect -0.200 0.120 0.053 -0.402 0.002 Direct effect -0.165 0.147 0.265 -0.456 0.126 Willingness to
share
demographic
information Indirect effect -0.035 0.106 0.741 -0.246 0.270
Total effect -0.329 0.093 0.001 -0.512 -0.146 Direct effect -0.235 0.133 0.080 -0.498 0.028 Willingness to
share lifestyle
Table 8B. Direct and Indirect Effects of X (Perceived Benefits) on Y
Y Effect SE/Boot SE p LLCI/Boot LLCI ULCI/Boot ULCI Total effect 0.501 0.071 0.000 0.361 0.641 Direct effect 0.468 0.073 0.000 0.324 0.612
Overall
willingness to share personal
information Indirect effect 0.033 0.027 0.144 -0.003 0.110
Total effect 0.455 0.074 0.000 0.308 0.602 Direct effect 0.420 0.076 0.000 0.269 0.571 Willingness to
share contact
information Indirect effect 0.035 0.032 0.137 0.000 0.129
Total effect 0.547 0.112 0.000 0.325 0.768 Direct effect 0.516 0.116 0.000 0.287 0.745 Willingness to
share general
financial
information Indirect effect 0.031 0.036 0.356 -0.024 0.136
Total effect 0.484 0.102 0.000 0.283 0.685 Direct effect 0.470 0.106 0.000 0.261 0.679 Willingness to
share
demographic
information Indirect effect 0.014 0.033 0.626 -0.037 0.094
Total effect 0.517 0.093 0.000 0.334 0.700 Direct effect 0.466 0.095 0.000 0.279 0.653 Willingness to
share lifestyle
Privacy Concerns
Hierarchical multiple regression analysis was conducted to examine the ability of perceived risks and perceived benefits to predict privacy concerns, after controlling for age, gender and education. Table 9 provides an overview of the results. In the first step of hierarchical multiple regression, three predictors were entered: age, gender and education. This model was not statistically significant F (3, 139) = 0.286, p =0.836 and explained 0.6% of variance in privacy concerns. After entry of perceived risks and perceived benefits at Step 2 the total variance explained by the model as a whole was 51.9% F (5, 137) = 29.577, p=0.000. In the final model, only perceived risks were statistically significant (β=0.708, p=0.000). In other words, if respondent’s perceived risks are a strong predictor of
respondents’ willingness to share personal information. When perceived risks increase for one, privacy concerns increase for 0.708. However, perceived benefits were not statistically significant to predict respondents’ privacy concerns (β= -0.031, p=0.629). Therefore, perceived risks have a positive impact on privacy concerns when controlling for perceived benefits. The proposed effect of perceived benefits on privacy concerns is not valid because no significant effect of perceived benefits on privacy concerns could be found when perceived risks were under control.
Table 9. Hierarchical Regression Model of Privacy Concerns
Model R R Square Adjusted R Square B Std. Error Beta T Sig
0.078a 0.006 -0.015 Age -0.020 0.034 -0.056 -0.589 0.557 Gender -0.024 0.239 -0.009 -0.101 0.920 1 Education 0.028 0.125 0.021 0.222 0.825 0.720b 0.519*** 0.502*** Age -0.022 0.024 -0.060 -0.885 0.378 Gender -0.069 0.173 -0.026 -0.397 0.692 Education 0.106 0.089 0.080 1.185 0.238 Perceived risks 0.739 0.068 0.703 10.903 0.000 2 Perceived benefits -0.027 0.072 -0.024 -0.370 0.712
a. Predictors：constant, Age, Gender, Education
b. Predictors：constant, Age, Gender, Education, Perceived risks, Perceived benefits *** Statistical significance: p<0.001
