Privacy Protection in Products

Privacy Protection in Products

Is there market potential for a privacy protection addition as a product feature?

3

Privacy Protection in Products

Is there market potential for a privacy protection addition as a product feature?

5

Management Summary

In this era of digital data, consumer information is collected to create personalized products or services. With this rise of personalization, a large range of concerns perceived by consumers for the security of their personal information rises as well and can be seen as an invasion of their personal lives. Eliminating these concerns has become a new business issue for many companies. An industry that widely uses these types of private information are mobile devices, and two of these product types, smartphone and wearables, are gaining popularity. These smartphones and wearables use wide amounts of personal data which can lead to the creation of large privacy concerns for their users. Companies recognize these privacy concerns and are examining how to integrate privacy into their businesses. By protecting and adding privacy features in their products or services, every firm is able to create additional value for their brand or company and can be seen as a source of competitive advantage. The aim of this thesis is to show the market potential of these privacy protection features in privacy concern evoking products and will attempt to analyze the effect of adding a privacy protection element to a smartphone and wearable, which results in the following main research topic:

How do consumers value a privacy protection addition to smartphones and wearables, and what is the willingness to pay for this protection?

This research contributes to the work field by discussing how consumers really value a privacy protection addition to their products and if, (e.g. why) companies should consider to add these additions to their products. This research is divided in three parts, a literature review of general privacy concerns and protection, and for mobile devices specific, will be discussed. Followed by a description on the method of research and results of a conjoint analysis to show the utility of these additions. The last part of this research will show the findings and limitations of the research. A review of prior literature on the subject of privacy concerns and the valuation of a privacy protection addition to wearables and smartphones gave directions to the suspected outcome of the researchs and leads to the following hypotheses:

6

(features) and in which way this affects their preferences for certain products. This type of empirical study is suited for the goal of this article because there is shown how consumers value the importance of an addition of such a privacy protection to their products and the willingness-to-pay for an addition in privacy protection can be calculated.

Looking at the conclusions of the conjoint analysis, a high privacy protection holds the highest utility for consumers for both smartphones and wearables in comparison with levels of lower privacy protection. For both smartphones and wearables, consumers have the highest probability of choosing high privacy protection over the other levels as well. When looking at the willingness-to-pay for this high privacy protection it is shown that the amount consumers are willing to spend for an upgrade, from no to a high level, is considerably low. This willingness-to-pay for smartphones is slightly higher (€5,72) than the amount the respondents want to spend for a wearable (€5,10). The relative importance of privacy protection of smartphones show that after brand, respondents allocated the most relative importance to privacy protection and the other attributes have a lower importance. For wearables privacy protection is third, after brand and the no purchase option. Price is following closely as well. This means brand and privacy protection are by far more important for smartphones, while for a wearable price and no purchase are almost equally important as well. This means the hypotheses specified in the literature review has to be rejected based on the empirical evidence. This hypotheses does not hold as the empirical findings show the opposite is true. Although for both product categories the utility for high privacy protection is highest in comparison with the other privacy protection levels. The relative importance and willingness-to-pay indicate privacy protection has a higher importance for smartphones than a wearable product holds. Although the hypotheses cannot be stated as true, the importance allocated to the addition of privacy shows there is potential for such additions, although consumers do not want to pay for it. In addition to this thesis research

7

Preface

This thesis was written during the last semester of my master Marketing Intelligence at the University of Groningen. During the master I gained the knowledge that was needed to

conduct and write this research. Although this study was constructed in a short period of time, I can state that I personally reached a new learning level. I felt fascinated and eager to dive into the theory of privacy and enjoyed creating my own empirical research on the subject. It is probably unnecessary to say it took some pain to reach the end state of this thesis and I am thankfull for everyone who took the time to assist me in my way to completion. I would like to thank Dr. Lara Lobschat for giving me the chance to take a dive in this research topic and for the help when confronted with problems. Last but definitely not least I would thank all friends, students, family and other helpers for aiding me by filling in my survey and the moral support.

9

Introduction

In this era of digital data, consumer information is collected to create personalized products or services which are becoming more important every day. Personalized message created by the use of consumer data can create experiences and value for consumers and lead to an increase in customer loyalty (Awas 2006). Yet the rise of personalized services can have a downside. With this rise, a large range of concerns perceived by consumers for the security of their personal information rises as well (Awad 2006). A survey by Accenture in 2010 showed that more than 50% of their respondents were worried about the privacy of their information and a smartphone was quoted to be ‘a potential portable, personal spy’ (Sutanto, Palme, Tan & Phang 2013). This personal security of information can be described as privacy, owning the facts of your life and having the freedom to reveal those facts (McrCeary 2008). These privacy details can include information about one’s preferences, finances, health and taste (McrCeary 2008). The collection and use of these personal details has raised serious concerns about a consumers’ privacy and can be perceived by them as an invasion of their personal lives. Eliminating these concerns has become a new but important business issue for companies, which use and profit from this private information (Lee, Ahn & Bang 2011). An industry of products that widely uses private information are mobile devices (Hoffman 2014). The expectation is, by 2025, mobile devices generate more information than all of the data generated in today’s economy, combining all sources (Maddox 2015). A more recent product line of mobile devices, wearables (clothes, smartwatches etc.), which gather and display personal data, are gaining popularity (The Economist 2015). Wearables can be used in a wide amount of activities and can therefore collect wide amounts of personal data about their users and environment, store data and transmit this information on a continuous base (Motti & Caine 2015). Wearables create highly personal data which can lead to the creation of privacy concerns for their users. Yet these wearable models currently offer a much smaller amount of personal information for interested companies who would like to use this

10

compromised if data is misused or a device is stolen or lost. These risks evoke privacy concerns of consumers (The Economist 2015).

Companies recognize these privacy concerns and are examining how to integrate privacy into their businesses and how this can add value for their customers (Hoffman 2014). By thinking strategically firms can transform concerns evoked by privacy into market advantages

(Cavoukian 2008). By protecting and adding security and privacy features in their products or services and creating new methods to protect personal information of customers, every firm is able to create additional value for their brand and company. These privacy concerns and the need for security control can be seen as a source of competitive advantage. A wide range of new products and services created from these privacy concerns are entering the market. Examples are the Gatekeeper, which automatically makes sure you are logged of on synchronizing devices when consumers are not using them, or SmartCipher, which ads protection by encrypting the emails of consumers (Gabrielson 2015). Recent studies reveal, new revenues can be generated by giving consumers the control of who learns, and what they learn from them (Hoffman 2014) but in today’s economy firms still earn more money by using consumer data then by creating protection for these types of personal information (Grant, Jentzsc, Harrison & Maertens 2014). This suggests a shift has to be made towards these new business opportunities because consumers seem to become more aware of other options, instead of just giving away their personal data to firms (Cardozo, Cohn, Higgins, Opsahl & Reitman 2014). This current research will show insights into these new business opportunities. One way to reduce privacy risks for consumers is for firms to embed protection into their existing products and services (Maddox 2015). Experts on privacy stimulate

11

research topic: How do consumers value a privacy protection addition to smartphones and wearables, and what is the willingness to pay for this protection?

This research will contribute to the work field by discussing how consumers really value a privacy protection addition to their products and if, (e.g. why) companies should consider to add these additions to their products. This thesis will contribute to literature because many studies are focused on concerns that are evoked by privacy and state that consumers become more prone to avoid these risks themselves by not sharing their information or overall

avoiding privacy concerning products. Recent work is suggesting firms protecting consumers’ privacy themselves can create a competitive advantage to differentiate on but little is known about the actual value consumers place towards these features when available to them. This research is divided in three parts. First a literature review of general privacy concerns and protection, and for mobile devices specific, will be discussed. Followed by a description on the method of research and results of a conjoint analysis to show the utility of these additions. The last part of this article will show the findings and limitations of the research.

Literature Review

Privacy Concerns

Privacy can be described as individuals controlling the terms under which their personal information is acquired and used (Culnan & Bies 2003). This is an important factor to consumers and companies in this current time. A wide amount of firms are creating

personalized services based on the collection of personal consumer information (Awad 2006). These personalized services can create advantages to the customer in ways of experiences and higher perceived value, yet it can create concerns about the security of their personal details as well (Awad 2006). The possibilities of data collection for companies are almost endless. They can collect a consumer’s browsing behaviour, history of purchases and location

information, over time. It can be used in multiple databases and shared with other parties. The benefits seem amazing for companies but concerns might be a larger downside than

12

phone records and information in email, chat, visual imagery and transfers are viewed by a variety of online companies as Facebook and Google (Tzanou 2013). As a recent study showed, consumers are aware of the risk their personal data can be compromised and a lot of companies think they are living up to what consumers expect them to, with respect to assuring privacy, which is overestimated in most cases (Conroy, Milano, Narula & Singhal 2014). These days, protecting their privacy becomes more top of mind for consumers and can

influence attitudes towards stores and brands (Hoffman 2014). Releasing personal information is perceived as a risk in a consumers mind because they are not sure how these details are handled by companies and know they can not control this. When consumers perceive these risks, this will increase their concerns about privacy. This may have negative consequences and might lead to risk aversion, a consumer refraining from use of concerning material which can be in the form of lower purchase and information sharing intentions (Youn 2009). A survey used in the research of Conroy, Milano, Narula and Singhal (2014), show that for 59% of consumers it would have a negative effect on how likely they would be to buy a brand that had a databreach (Conroy, Milano, Narula & Singhal 2014). Making sure your personal information is secure is currently a wide spread concern for consumers, and trying to limit the extent to which your information becomes public yourself, is becoming very difficult as everywhere you go, a digital trace is left behind (Tapscott 2016). Especially individuals with a higher education and income are more aware of the danger regarding their privacy, as well as young adults perceive more concerns than adults (Miller 2014). When consumers believe their details to be valuable for companies, this will decrease their intention to share these details (Kyung 2013). Consumers can be compromised in their personal information in three ways, their personal data can be compromised while in their own hands, their data can be

compromised while transferring to a company and consumer details that are already in hands of a company, are still always at risk as well. An example is JetBlue Airlines which gave records of the traveling behaviour of their customers to another company in secret, which in turn gave it to another company who posted records online. This shows personal details always have the possibility to be disclosed to others (Milne, Rohm & Bahl 2004). Previous research examines that consumers do want the benefits of sharing their personal data and creating openness but prior research implies as well they want a perfect security of their privacy (The Economist 2015). In conclusion to these research findings there can be said that using personal data can create personalized messages and attributes which better serves the need and preferences of consumers and might lead to a higher adoption rate. Yet in

13

as disclosing personal information makes consumers more vulnerable (Aguirre, Mahr, Grewal, de Ruyter & Wetzels 2015).

Privacy Protection

Because of the digital breadcrumbs every consumer leaves behind, it is very difficult to protect yourself and your data (Tapscott 2016). Consumers often are informed about privacy protection, but the use of these strategies of protection is very low (Dommeyer & Gross 2003). Consumers might weigh their privacy concerns about their personal information against their advantages, therefore it can have a negative effect on purchase or use (Rapp, Hill, Gaines & Wilson 2009). Industry research shows only 37% of consumers thinks their privacy is protected enough by companies (Conroy, Milano, Narula & Singhal 2014). As a response, firms can transform concerns evoked by privacy in to market advantages

(Cavoukian 2008). A new product is pushed forward, privacy (Angwin & Steel 2011), as consumers are creating awareness of how valuable their information is, protection is what some seek for (Peterson 2014).

These days not many firms see privacy protection as anything else than following rules, which is not very likely to create a lot of trust and sharing, especially not when consumers are not aware of their protection (Conroy, Milano, Narula & Singhal 2014). Companies do not want consumers restricting the information they share, and therefore have to offer better services to protect privacy (Tapscott 2016). For example, Facebook recognized the concern to manage an individual’s privacy better as they were worried consumers would share less information or abandon the company completely. Their research showed that the growth of the company was largely depending on how consumers have the trust to share their personal details (Goel 2014). This means protecting a consumer’s privacy is not just making sure their data cannot be breached, it is also important to let consumers know they can trust your product and brand (Conroy, Milano, Narula & Singhal 2014). Personalized messages might appeal to consumers more as it lies within their line of interests but it can be considered as violating privacy as well. When resistance becomes the reaction towards personalized messages, it holds no value for companies. Protecting privacy might minimize this resistance because of the perception of their increased information safety and can improve the impact of the personalized message (Tucker 2014). Research shows a link between how save consumers perceive their

14

take the safety of their data into consideration in the process of buying and even 80% states to be positively influence to buy from a company that they believe protects their data. Research shows as well that 81% of the consumers believe companies have the responsibility to protect their data, although this perception is not always recognized by the companies itself. Also the link between the protection of one’s personal details and the decision to buy can be larger than a lot of companies realise (Conroy, Milano, Narula & Singhal 2014). Consumers are vulnerable to the collection of personal information because it can be disguised with certain online entertainment practices. They often feel like they are not capable of the protection of their privacy themselves due to lack of knowledge (Youn 2009). In example, transactional data or details about website visits can be intercepted if encryption protocols are not used but these encryptions are a highly technical skilled process to accomplish (Milne, Rohm & Bahl 2004). This can evolve in two strategies to handle concerns, approach and avoidance

strategies. Where an approach strategy leads to learning to protect themselves and avoidance strategy will result in ignoring the perceived risks, withholding or removing personal details and transactions or entirely refusing the use of the product (Youn 2009)(Bélanger 2011). Recent study focusses mainly on the protection of privacy as a fair information practice of companies through showing their notices on the use of privacy to consumers. This means companies mostly inform consumers about the use of private information, not actively taking measurements to protect it (Milne, Rohm & Bahl 2004). Next to current theoretical research it is necessary to examine empirically how privacy is valued by consumers (Friedewald, Wright, Gutwirth & Mordini 2010). It is important research can show companies how privacy is really appreciated by consumers, to adjust their policies towards it (Steijn & Vedder 2015) as

15

balance has to be found where these benefits are experienced with protection of privacy (Siyavooshi, Sanayei & Fathi 2013). With the stakes of privacy protection being this high, companies like Apple will have to adapt their approach, as said by their vice president of software engineering, Craigh Federighi, although security of personal information is believed to be an endless race, personal information should be protected as well as possible (Perlroth & Benner 2016).

Privacy Paradox

Although research indicates that consumers can feel an increase in privacy concerns and want their privacy protected, their action sometimes contradict these believes and the privacy paradox is created (John 2015). People say they have concerns about privacy but it does not automatically result in them staying away from privacy concern creating products (Cardozo, Cohn, Higgins, Opshal & Reitman 2014). Consumers say they perceive a wide range of privacy concerns on mobile phones and do not believe companies are protecting their data, yet this has no immediate effect on the usage of the products, and therefore their

concerns do not limit the extend of information that is handed over (Miller 2014). Consumers often see the benefits of using products, which gather personal information, like mobile devices. These devices can aid in aspects of a consumers life and create value because of their small size and pervasiveness in for example healthcare or sport related activities. Mobile devices have therefore become valuable markets for companies as well because a lot of consumers trust a lot of personal data to these devices (Perlroth & Benner 2016). The privacy paradox shows, consumers see these benefits, using privacy concerning products like mobile devices give and therefore are willing to accept the risks these might bring (Motti & Caine 2015). A survey from Pew Research Center shows that their respondents do not perceive any way of communication or channel, as save to pass along private information. Of all channels used for communication, landlines are rated highest on trust yet no actions are shown on this statement as usage of this product is in a strong decline (Miller 2014). People may be willing to let go of their privacy in exchange for certain advantages the services and products offer them (Sutano, Palme, Tan & Phang 2013) or it is perceived by consumers that they do not have a choice with regards to their contradicting actions because they have invested too much in the service or product already. When an individual’s social contacts all run through

16

consumers even saw value in the use of their personal details to make the services more efficient (Miller 2014). This might mean that consumers, in contradiction to their rising concerns, might not be willing to invest in an addition of privacy protection towards their products (Technology Quarterly 2014).

Prior research has indicated that the negative consequences privacy concerns bring might lead to risk aversion, a consumer refraining from use of concerning material (e.g. lower purchase and information sharing intention) but if they perceive previously discussed benefits of the disclosure of personal information, this might lead to an decrease of privacy concerns and weakens the intention and motivation to engage in privacy protection. This effect is only encountered when the benefits outweigh the concerns evoked by the risk perceived (Youn 2009). Protecting the privacy of a consumer has become a continuingly growing issue over the past years and made companies sensitive to the concerns a consumer might have. Still most consumers do not withhold personal information that might have negative consequences for them. They do agree privacy is important and worth protecting but their actions often say otherwise as they tend to share their details for rather small benefits like discounts or just to complete transactions. Research on this phenomenon therefore shows contradicting behavior (Kyung 2013). A declaration of concerns might not always lead to protecting measurements as behaviour as response to these concerns might diverge (Monteleone 2015).

Consumers face constant trade-offs between the benefits of products and services that better fit their needs and the risk of compromising their privacy, yet not many research has been done on how concerns and intention to protect privacy actually results in market behavior (Norberg, Horne & Horne 2007). This thesis will examine how consumers actual value privacy protection as an addition to existing product attributes.

Privacy Concern Evoking Products

As mentioned before, by 2025, it is expected that the information generated by mobile devices will exceed all data in today’s economy combined (Maddox 2015). These new medias are affecting our behaviour with respect to working, thinking, learning but also playing, yet there can be said that because of the rise of mobile devices, the digital environment can be a confusing and dangerous place (Tapscott 2016)(Morey, Forbath & Schoop 2015).

17

prices (The Economist 2015). Wearable devices can aid in different aspects of a consumers life and creates value because of their small size and pervasiveness in for example healthcare or sport related activities. Because these products can be used in multiple activities, they collect a wide amount of personal data about their users, as well as surroundings. They store data and transmit information on a continuous base (Motti & Caine 2015). By using this product, consumers create data about their personality, health and other highly personal data, which can lead to a risk to be compromised and therefore create privacy concerns (The

Economist 2015). If a company attracts more data, especially sensitive data, there is perceived that the risk of data breaches is greater (Conroy, Milano, Narula & Singhal 2014). These privacy concerns might lead to a lower adoption rate for these type of devices (The Economist 2015).

Still, despite the wide amount of data created by wearables, it is currently a smaller amount of information in comparison to a smartphone (The Economist 2015). Current smartphones are to be compared with a modern day computer and gathers a wide range of personal information as the products go, where the owner goes (Sutanto, Palme, Tan & Phang 2013). Mobile

phones have become valuable markets for companies because a lot of consumers trust a lot of personal data to these devices (Perlroth & Benner 2016). Research on smartphones shows most users don’t really know which data is created by using the product and have no

18

this medium is rapidly increasing (Siyavooshi, Sanayei & Fathi 2013). Yet it creates sensitive data as they track location data for example. This personal information can be used or even misused for a range of activities, from innocent advertising purposes to potential stalking. Protection of this data is important as consumers do not want to experience the use of mobile devices as the feeling of being under surveillance (King & Raja 2013) and increasing

concerns can make mobile device products less attractive over time (Siyavooshi, Sanayei & Fathi 2013).

It can be concluded that using a wearable and smartphone can create privacy concerns. Consumers are concerned on how their information on location and context is gathered and used (Haque 2015). Research shows that consumers rate attributes like a camera and

microphone with high privacy concerns but sensors that measure facts like heart rate and the tracking of activities evoke fewer privacy concerns within a consumers’ mind, this is most likely because consumers are less aware of the implications of this type of information that is collected. The most important concerns consumers say they have are created by GPS sensors (Motti & Caine 2015).

Because mobile technology is increasingly evolving the need to offer protection and privacy becomes more important. Prior research on behaviour targeting shows more than 50% of the consumers are afraid of their privacy on mobile devices and about 80% does not read the notices companies offer to them. This suggests mobile companies need to find another, for consumers perceived simpler way, to protect their private information. Next to most current practices which withold being informed and having a choice about the disclosure of private information, consumers desire another addition, the actual implementation of procedures on a technical, administrative and physical way to protect their privacy (Timpson & Troutman 2009). With this knowledge, adding privacy protection to smartphones and wearables could be profitable. Research by the economist show that respondents say their intention to use a wearable would increase with around 28% if they would be more discreet (The Economist 20150) and although it may need an investment at first, the long term results can hold more willingness to share information and trust (Conroy, Milano, Narula & Singhal 2014). Consumers might even be willing to pay for more trust (Reichheld 2003).

Although research on concerns about smartphones and wearable devices show they are perceived as similar on a wide base, specific attributes of wearables have a larger

19

privacy becomes more critical (Motti & Caine 2015). In collecting data, wearables have fewer stopgaps in the records of data and therefore delivers more continuous information for firms (Alton 20150). It can therefore be expected that the effect of privacy concerns on the adoption of a wearable device might be larger (Motti & Caine 2015) and leads to the following

hypotheses:

H1: A smartphone will have a lower utility of privacy protection addition and a lower willingness to pay for privacy protection attributes in comparison to a wearable.

Methodolgy

To empirically test the hypotheses constructed in the literature review, a conjoint analysis is executed. A conjoint analysis provides a clear image of how consumers perceive different attributes (features) and in which way this affects their preferences for certain products (Schlegelmilch & Ambos 2004). This type of empirical study is suited for the goal of this article, defining the market potential for privacy protection additions, because there will be shown how consumers value the importance of an addition of such a privacy protection to their products. A conjoint analysis asks the respondents to choose alternatives with a variety of features, shuffled in choice tasks. By choosing between different alternatives of attributes (features), attention is drawn towards features (attributes) that are most important to the respondents (Meisner, Musalem & Huber 2016). This will show how much utility is allocated towards such a privacy addition and if this affects their choice preference of products.

Measures

For this conjoint analysis two product of the mobile devices category, smartphones and wearables, are used to examine the utility of an addition of privacy protection and these utilities are to be compared later on to examine their differences. These two types of products are chosen because they tend to evoke high concerns for privacy because of their features and amount of intimate information, as shown in the literature review. When choosing product attributes for the conjoint analysis, three criteria can give a guideline. It is stated that the attributes used, should be important for the consumer and play a role in the decision process. The attributes also have to be actionable and should all be independent of each other

20

To find the most important attributes for wearables, comparison websites are examined and provided evidence to use the following attributes and their levels.

For wearables the three best-selling brands are used as well as the three best-selling colors and two types of most wanted features as this is essential for the goal of usage of the product. In addition three price levels are used, ascending from a lower price level towards a higher price (Colon 2016). These attributes are accompanied by the addition of privacy protection with four levels, no, low, medium and high to examine the importance of each level of protection. Brand:  Apple;  Samsung;  Pebble. Price:  €199;  €299;  €399. Color  Black;  Silver;  Gold. Features:  Accelerometer;  Accelerometer + Heartratemonitor. Privacy Protection:  No;  Low;  Medium;  High.

21 Brand:  Apple;  Samsung;  HTC. Price:  €399;  €499;  €599. Color  Black;  Silver;  Gold. Memory:  16GB;  32GB. Privacy Protection:  No;  Low;  Medium;  High.

To control for variating results, the following factors are taken into consideration in addition to the conjoint variables, explained with the source, the items and scaling are based on:

Demographic

Gender (Venkatesh, Davis & Davis 2003):  Male

 Female

Age (Venkatesh, Davis & Davis 2003):  Younger than 18 years old;  18-24 years old;

22

 65-74 years old;  75 years or older.

Highest received diploma (Snibbe, Conner & Markus 2005):  No education;  Primary education;  LBO/MAVO/VMBO  MBO  HBO/WO  Other.

Houshold income (Weeden & Grusky 2005):

(An average income is considered to be around €34.500 in the Netherlands (CBS 2015)).  Below €34.500;  Around €34.500;  Around €51.500;  Around €69.000;  More than €69.000. Additional variables

Involvement in the product category (Chandrasekaran 2004):

Answer how much you agree with the following statements about mobile devices

(smartphones and wearables). Measured on a 5-point scale (totally disagree – totally agree).  I am interested in mobile devices.

 Given my interest, mobile devices are not very relevant to me.  Overall, I am quite involved when I purchase mobile devices.

Involvement is measured to account for variating results caused by differences in levels of involvement of respondents in purchasing mobile devices. As low involvement might indicate one has a lower interest in purchasing a mobile device, thinking about important attributes and therefore allocates less importance towards protecting ones privacy.

Brand preferences (Niedrich & Swain 2003):

Divide 100 points over the following four brands of mobile devices (smartphones and wearables of Apple, Samsung, HTC and Pebble), in such a way that it reflects your preferences (the most points for the brand you most prefer).

23

respondent with a preference for a certain brand might always choose that brand, with no regard for other attributes.

Products already owned (Summers, Belleau & Xu 2006):  None;

 Smartphone;  Wearable;

 Both smartphone and wearable.

Already owned products are measured to account for variating results in purchase intention/choice caused by differences in prior possession of the product(s).

Results

Sample

The data used in this research is gathered using an electronic tool called Preference Lab, and was send out to respondents by means of a direct mailing and social media. The sample used in this conjoint analysis holds the results of 239 respondents who filled out a structured questionnaire between the 25th of April 2016 and May 16th. The respondents were presented with a choice-based conjoint analysis because this type of conjoint analysis is most close to how products are evaluated in a real life situation, which make the results more valid (Green & Srinivasan 1990). A fractional factorial design is used as when using a full factorial design the respondents would have to evaluate an amount of choices too large to complete this conjoint research. The design was made as efficient as possible by being balanced and orthogonal with minimal overlap in choice sets and use of non-dominated choice sets (Green & Srinivasan 1990. The number of alternatives is set to three (usually between two and five) and the number of choice sets is set to eight, (usually between eight and sixteen) for each product type separately, to make the effect of fatigue as low as possible. When using more than twelve choice sets, a motivation for consumers is needed and because the conjoint analysis for the two products combined holds sixteen choice sets in total, an incentive is used (Ding, Grewal & Liechty 2005) in this case, all respondents of the research had a chance to win two gift-certificates.

24

buying a smartphone or wearable. A principal component analysis was conducted after recoding all three statement results into the same direction and an orthogonal rotation (varimax) was used and showed one factor for the three items as this explained 65,9% of the variance and had an eigenvalue over 1. The rotated component matrix, shows the items belong to one factor as well and can therefore be combined into one variable: Involvement. Combining the three variables into one variable gives a Cronbach’s Alpha of .74, the value would not increase if variables would be deleted, and the new variable can therefore be considered internally consistent and reliable as the value is higher than the threshold of .70.

Descriptive Statistics

The descriptive statistics are presented in table 1. Most of the respondents were male (57,3%) while 42,7% was female. The largest group (40,2%) of the respondents are between 18 and 24 years old and have an household income of below €34.500,- which is 49,80% of the sample. 59,40% of the respondents are HBO or WO educated, followed by the second largest group with a MBO education (22,60%). The mean time respondents took to fill in the questionnaire was 246 seconds, e.g. 4 minutes. The involvement of the respondents can be described as neutral. Out of a 100 preference points, Apple gets an amount in mean points of 48,2, Samsung gets 32,1 points, HTC gets 14,3 and Pebble is allocated 5,4 points. This means the brand Apple is most preferred, followed by Samsung.

25 Income < €34.500 119 49.8% Around €34.500 39 16,3% Around €51.750 35 14,6% Around €69.000 15 6,3% > €69.000 31 12,0% Owning a product No 10 4,2% Smartphone 193 80,8% Wearable 2 0,8% Both 34 14,2%

Table 1 - Descriptive Statistics Emperical findings

The needed utilities of all (smartphone and wearable) attributes are calculated in Latent Gold and will construct a model which reflect the structural relationships between the respondents and the utilities they allocated. Classes are constructed that represent categories of respondents with similar preferences and show the utilities belonging to these classes. To show which model (number of classes) shows the best fit to the data, the BIC value, 𝑅2 and hitrate are examined. The model with the lowest BIC value will often fits best to the data, as well as a higher 𝑅2 value as this indicates how much of the variance in the data can be explained by the model. It will show the goodness of the model prediction. The hitrate shows how much of the observed choices in data are predicted correctly by the model and will represent the prediction error, which should be as low as possible for the best model.

Smartphone results

After conducting a model estimate of the utility of smartphones in Latent Gold, it is shown that this specific model has to be treated as a 2-Class choice model as this holds the lowest BIC value (1910,9) and improves the 𝑅2 from 17,9%, the value for the aggregate model with 1 class, to 34,5%. The aggregate model with all respondents in one class has a hit rate of 62,6% with a prediction error of 0,37, the 2-class model has a hit rate of 70,1% with a

26

Class description

Consumers belonging to the first class allocated a mean points of 46,2 points of 100 towards Apple as an indicator of their preferences and 36,0 towards Samsung. Consumers belonging to the second class allocated a mean points of 55,8 out of 100 towards Apple as an indicator of their preferences and in second place, Samsung with 20,3 points. This means this class has a higher preference for Apple and much lower for Samsung as the first class. The probability that someone in the first class only has a smartphone is 66,8%, while 24,3% has a wearable as well. The probability that someone in the second class has a smartphone is 88,9%, while only 65,7% has a wearable as well. This means the first class has a much higher probability of owning a wearable than the second group. Differences in the conjoint preferences are discussed in the following paragraphs.

Utility

In table 2 the utilities respondents, divided in the two classes, allocate to the different

attributes and levels are shown. Class 1 gives the highest relative utility, in comparison to the different levels, towards Samsung, a price of €399, memory capacity of 16GB, a black color and a high privacy protection. A purchase in comparison to no purchase, has the highest utility. Class 2 has a preference for Apple, allocates the most utility to a price of €399, 16GB and a high privacy protection as well as class 1 and prefers a silver color. Purchase, in comparison to no purchase, has the highest utility in this class as well.

27 No -0,7961 -0,2295 Low -0,3309 -0,4518 Medium 0,3421 0,6013 High 0,7849 0,8406 Purchase Intention No Purchase -0,0225 -0,3853 Purchase 0,0225 0,3853 Table 2 - Utility of Smartphone Attributes Choice probabilities

Table 3 shows the probability a person in the previously defined two classes would choose one of the given levels over the other levels. A consumer in class 1 has the highest probability to choose Samsung with a price of €399, a memory of 16GB and a black color. Choosing a high privacy protection has the highest probability as well as the probability of purchasing the product although this does not differ much with the probability of not purchasing the product. A consumer in class 2 has the highest probability to choose Apple with a price of €399, a memory of 16GB and a silver color. Choosing a high privacy protection has the highest utility as well as the probability of purchasing the product.

28

Purchase Intention

No Purchase 0,4888 0,3163 Purchase 0,5112 0,6837 Table 3 – Probabilities Smartphone Attributes Importance of the Attributes

Figure 1 shows the relative importance of the attributes, perceived to be by the respondents of the research. This importance measures how much influence each attribute has on peoples’ choices. For both classes brand and privacy are the most important.

Figure 1 – Relative Importance Attributes Smartphone

Wearable results

After conducting a model estimate of the utility of wearables in Latent Gold, it is shown that this specific model has to be treated as a 4-Class choice model as this holds the lowest BIC value (5725,4) and improves the 𝑅2 from 13,4%, the value for the aggregate model with 1 class, to 52,1%. The aggregate model with 1 class has a hit rate of 57,7% with a prediction error of 0,43, the 4-class model has a hit rate of 79,9% with a prediction error of 0,30 and therefore, based on the 𝑅2 and hitrate, the 2-class model fits well and can explain more of the choices of the respondents than the aggregate model.

Class description

Some differences in the four classes are that consumers belonging to the first class have the largest probability to have preference for Samsung (43,9 mean points out of a 100), the consumers in the second class the largest probability to have a preference for Apple (74,3

0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5

Brand Price Memory Color Privacy Protection No Purchase

RELATIVE IMPORTANCE OF ATTRIBUTES SMARTPHONE

29

mean points out of a 100), the third class the largest probability to have a preference for Apple as well (85,1 mean points out of a 100) and the fourth class the largest probability to have a preference for Samsung again (44,6 mean points out of a 100). The probability that someone in the first class only has a smartphone is 86,3%, while 10,9% has a wearable as well. The probability that someone in the second class only owns a smartphone is 65,6%, and 16,3% has a wearable as well. The probability that someone in the third class only owns a smartphone is 88,2% and a wearable as well is only 9,5%. The probability someone in the fourth class only owns a smartphone is 64,1% while 65,8% owns a wearable as well. This means the fourth class has a much higher probability of owning a wearable than the other groups. Differences in the conjoint preferences are discussed in the following paragraphs.

Utility

In table 4 the utilities respondents allocate to the different attributes and levels are shown. Class 1 prefers Samsung as brand, a price of €199, an accelerometer + heartrate monitor and a black color. A high privacy protection has the most utility allocated to it, as well as a

purchase. Class 2 prefers Apple as brand, a price of €299, an accelerometer + heartrate monitor as well and a silver color. This class allocates the highest utility to high privacy protection as well although no purchase is given the highest utility. Class 3 prefers Apple as well with a price of €199, a black color and an accelerometer + heartrate monitor. High privacy protection has the highest utility, as well as purchase over no purchase. Class 4 prefers Samsung with a price of €199, an accelerometer + heartrate monitor and a gold color. High privacy protection has the highest utility again but the no purchase option as well.

30 Silver 0,0135 0,1593 0,0692 -0,1807 Gold -0,1523 -0,2004 -0,2831 0,1240 Privacy Protection No -0,5389 -0,6741 -0,6262 -0,6759 Low -0,1948 -0,1583 -0,2910 -0,2972 Medium 0,0850 0,3899 0,3323 0,0788 High 0,6487 0,4425 0,5849 0,8943 Purchase Intention No Purchase -1,3378 0,3060 -2,7058 0,9839 Purchase 1,3378 -0,3060 2,7058 -0,9839 Table 4 – Utility Wearable Attributes

Choice probabilities

The following table shows the probability a person in the previous defined classes would choose one of the given levels over the other levels. A consumer in class 1 has the highest probability to choose Samsung, a price of €199, an accelerometer + heartratemonitor with a black color and high privacy protection. This class has a very high probability for purchasing the product. A consumer in class 2 hast he highest probability to choose Apple, a price of €299, an accelerometer + heartratemonitor with a silver color and high privacy protection but have the highest probability of not purchasing the product. A consumer in class 3 has the highest probability to choose Apple, a price of €199, an accelerometer + Heartratemonitor with a black color and high privacy protection. This class has the highest probability of purchasing the product over not purchasing. A consumer in class 4 has the highest probability to choose Samsung with a price of €199 , an accelerometer + Heartratemonitor and a gold color with high privacy protection but have the highest probability of not purchasing the product.

31 Heartrate monitor Color Black 0,3803 0,3435 0,4043 0,3498 Silver 0,3355 0,3866 0,3498 0,2579 Gold 0,2842 0,2698 0,2459 0,3742 Privacy Protection No 0,1323 0,1159 0,1196 0,1064 Low 0,1867 0,1941 0,1672 0,1554 Medium 0,2470 0,3359 0,3118 0,2264 High 0,4340 0,3540 0,4014 0,5117 Purchase Intention No Purchase 0,0644 0,6484 0,0044 0,8774 Purchase 0,9356 0,3516 0,9956 0,1226 Table 5 – Probabilities Wearable Attributes

Importance of the Attributes

Figure 1 shows the relative importance of the attributes of the wearable, perceived by the respondents of the research. This importance measures how much influence each attribute has on people choices. For class 1, the none-option is most important, followed by price and privacy protection. For class 2, brand is most important followed by privacy protection. For class 3 the none-option is most important, followed by brand and privacy protection. For class 4 the none-option is most important as well followed by privacy protection.

0 0,05 0,1 0,15 0,2 0,25 0,3 0,35 0,4 0,45 0,5

Brand Price Features Color Privacy

Protection

No Purchase RELATIVE IMPORTANCE OF ATTRIBUTES WEARABLE

Class 1 Class 2 Class 3 Class 4

32

Wearable and Smartphone comparison

When comparing the relative importance the respondents allocated towards the attributes of both products it is shown that for both types of mobile devices a privacy protection attribute is placed high. For the wearable privacy protection holds the third place and is not as important as brand and the no purchase option, this might indicate the purchase intention for this product type is not high in this respondent sample. For smartphones privacy protection was secondly most important after brand. This means privacy protection is considered to be more important in smartphones than in wearables when influencing their choices for an alternatie.

Willingness-to-pay

The willingness-to-pay can be examined to research the effect of adding a privacy protection addition to the price a consumer is willing to spend for a smartphone or wearable. In the wearable product category, consumers in class 1 are willing to spend €1,58 for an upgrade from no privacy protection to high privacy protection. For consumers in class 2 this amount is €3,85. Class 3 is willing to spend €3,03 for an upgrade to high privacy protection and class 4 is found to be willing to spend €5,10 and is therefore the group who has the highest

willingness-to-pay for privacy protection. For the smartphone attributes, consumers in class 1 are willing to spend €4,09 for an upgrade from no privacy protection to high privacy

protection. For consumers in class 2 this amount is €5,72. This means for smartphones,the respondents are willing to spend a small amount more to upgrade their privacy protection than for the wearable product category, but for both product categories the willingness-to-pay for privacy is small.

Conclusion

Conclusion findings

When looking at the utility of privacy protection for both smartphones and wearables, a high privacy protection holds the highest utility for consumers in comparison with the other levels (no, low, medium). For both smartphones and wearables, consumers have the highest

33

privacy protection for smartphones and wearables become more visible. After brand, respondents allocated the most relative importance to privacy protection and the other attributes have a larger difference in importance, as for wearables privacy protection is third after brand and the no purchase option, price is following closely as well. This means brand and privacy protection are far more important than the other attributes when choosing a smartphones, while for a wearable price and no purchase are almost equally important as well. This means the following hypotheses specified in the literature review has to be rejected based on the empirical evidence.

H1: A smartphone will have a lower utility of privacy protection addition and a lower willingness to pay for privacy protection attributes in comparison to a wearable. The hypotheses of the thesis does not hold as the emperical findings show the opposite is true. Altough for both product categories the utility for high privacy protection is highest in

comparison with the other privacy protection levels. The relative importance and willingness-to-pay indicate privacy protection has a higher importance for smartphones than a wearable product holds. Although the hypotheses can not be stated as true, the importance allocated to the addition of privacy shows there is potential for such additions, although consumers do not want to pay for it.

Contribution

The literature review showed there is a market for privacy protection although uncertainty was established about consumers’ responses towards these additions as concerns for privacy might not always result in action towards privacy protection. A lot of research has been focused on explaining theoretical responses about privacy while few journal articles focus on measuring empirical responses to additions of privacy protection to products. The results of this research, in line with the literature review, showed privacy does have a lot of influence on respondents choices as most of the time it is rated as one of the most important attributes of smartphones and wearables. Yet never the most important attribute, brand for both product types and the option not to purchase, and followed closely by price for wearables are often one step more important.

Discussion

34

wearables or the fact that a large part of the respondents has no intention to buy a wearable and therefore have a different view towards the privacy concerns evoked by this product type as this is not relevant to them.

The willingness-to-pay for an upgrade of no privacy protection towards a high privacy protection for both products is examined to be very low. Reasons for this phenomenon might be found in the privacy paradox. Altough the respondents hold privacy protection as relatively important, this importance is not reported in their actions as they are not willing to pay more to protect their privacy. Another reason for the surprising willingness to spend on privacy protection might be hidden in the calculation of these spendings, for example when the respondents are not as price sensitive or when they react positively to increases in price.

Implications

This thesis results in implications for companies of mobile devices that are questioning the importance of privacy for their products. These firms should consider to add a privacy protection addition to their products as the theoretical and current empirical evidence of this thesis shows respondents hold the importance of such an addition in high regards. This indicates there is a market potential for these additions although the respondents are not willing to spend additional amounts on an upgrade in privacy protection. This means investments in privacy protection should come from within the company but still can have improving results as shown in the literature review. By adding such a protection consumers might be less concerned of the use of these products and therefore will not move away from use or purchase and might trust more information to these products.

Limitations

This study has several limitations for which further research has to be conducted. The empirical evidence used to write this thesis was based on a student-based sample and

therefore problems might arise in making the research generalizable. Further research should be conducted in a more diverse sample of respondents to compare the results. In addition to a a more diverse sample, other products can be examined as well to compare differences in importance of privacy protection and willingness-to-pay for such an addition.

35

Bibliography

Aguirre, E., Mahr, D., Grewal, D., de Ruyter, K., & Wetzels, M. (2015). Unraveling the Personalization Paradox: The Effect of Information Collection and Trust-Building Strategies on Online Advertisement Effectiveness. Journal of Retailing 91, 34-49. Alton, L. (2015, Sep 12). How Wearable Tech Could Spark A New Privacy Revolution.

TechCrunch.

Angwin, J., & Steel, E. (2011, February 28). Web's Hot New Commodity: Privacy. The Wall Street Journal.

Awad, N. F. (2006). The Personalization Privacy Paradox: An emperical Evaluation of Information Transparency and the Willingness to be Profiled Online for

personalization. MIS Quarterly Vol. 30 No. 1, 13-28.

Bélanger, F. (2011). Privacy in the Digital Age: A Review of Information Privacy Research in Information Systems. MIS Quarterly vol. 35 no. 4, 1017-1041.

Cardozo, N., Cohn, C., Higgins, P., Opsahl, K., & Reitman, R. (2014). Who has your back? Protecting your data from goverment requests. Electronic Frontier Foundation. Cavoukian, A. (2008). The Privacy Payoff: How building privacy into your communications

will give you a sustainable competitive advantage. International Association of Business Communicators, (pp. 1-45). New York.

CBS. (2015, September 7). Gemiddeld inkomen; particuliere huishoudens naar diverse kenmerken. Retrieved from Centraal Bureau Statistiek:

http://statline.cbs.nl/statweb/publication/?dm=slnl&pa=70843ned&d1=a&d2=0&d3=0 &d4=a&hdr=g1,g2,g3&stb=t&vw=t

Chandrasekaran, R. (2004). The influence of Redundant Comparison Prices and Other Price Presentation Formats on Consumers' Evaluations and Purchase Intentions. Journal of Retailing vol. 60 issue 1, 53-66.

Colon, A. (2016, May 9). The Best Smartwatches of 2016. PCMAG.

36

Culnan, M., & Bies, R. (2003). Consumer Privacy: Balancing Economic and Justice Considerations. Journal of Social Issues Vol. 58 No. 2, 323-342.

Ding, M., Grewal, R., & Liechty, J. (2005). Incentive Aligned Conjoint Analysis. Journal of Marketing Research Vol.XLII, 67-82.

Dommeyer, C., & Gross, B. (2003). What Consumers Know and what they Do: An Investigation of Consumer Knowledge, Awareness, and Use of Privacy Protection Strategies. Journal of Interactive Marketing vol. 17 no. 2, 34-51.

Friedewald, M., Wright, D., Gutwirth, S., & Mordini, E. (2010). Privacy, Data Protection and Emerging Sciences and Technologies: Towards a Common Framework. The European Journal of Social Science Research vol. 23 no. 1, 61-67.

Gabrielson, C. (2015, November 23). Privacy Products You Might Want To Know About. The Privacy Advisor.

Goel, V. (2014, May 22). Some Privacy, Please? Facebook Under Pressure, Gets the Message. The New York Times.

Grant, J., Jentzsc, N., Harrison, R., & Maertens, F. (2014). Monetizing privacy and data protection: Can privacy be profitable? CDPC Conference. Brussel: IPaCSO.

Green, P., & Srinivasan, V. (1990). Conjoint analysis in Marketing: New developments with implications for research and practice. Journal of Marketing vol. 54 iss. 4, 3-19. Haque, U. (2015, February 5). Managing Privacy in the Internet of Things. Harvard Business

Review.

Hoffman, D. (2014). Privacy is a business opportunity. Harvard Business Review. John, L. (2015). We say we want privacy online, but our actions say otherwise. Harvard

Business Review.

King, N., & Raja, V. (2013). What Do They Really Know About Me in the Cloud? A Comparative Law Perspective on Protecting Privacy and Security of Sensitive Consumer Data. American Business Law Journal vol. 50 no. 2, 413-482. Kumar, R., Gopal, R., & Garfinkel, R. (2010). Freedom of Privacy: Anonymous Data

37

Kyung, E. (2013). Behind the Privacy Paradox: Decreasing Disclosure by Viewing

Information as a Constrained Resource. Advances in Consumer Research, 156-160. Lee, D.-J., Ahn, J.-H., & Bang, Y. (2011). Managing Consumer Privacy Concerns in

Personalization: A strategic Analysis of Privacy Protection. MIS Quarterly Vol. 35 No. 2, 423-444.

Maddox, T. (2015, Octobre 21). The Dark Side of Wearables: How they're secretly jeopardizing your security and privacy. TechRepublic.

Marr, B. (2015, April 21). How Big Data Is Changing Healthcare. Forbes. McrCeary, L. (2008). What was Privacy? Harvard Business Review.

Meisner, M., Musalem, A., & Huber, J. (2016). Eye Tracking Reveals Processes that Enable Conjoint Choices to become Increasingly Efficient with Practice. Journal of Marketing Research vol. 53. , 1-17.

Miller, C. (2014, November 12). Americans Say They Care About Privacy But Act As If They Don't. The New York Times.

Milne, G., Rohm, A., & Bahl, S. (2004). Consumers' Protection of Online Privacy and Identify. Journal of Consumer Affairs vol. 38 no. 2, 217-234.

Monteleone, S. (2015). Addressing the ‘Failure’ of Informed Consent in Data Protection. Journal of International Law & Commerce, 70-110.

Morey, T., Forbath, T., & Schoop, A. (2015, May). Customer Data: Designing for Transparency and Trust. Harvard Business Review.

Motti, V. G., & Caine, K. (2015). Users' Privacy Concerns about Wearables: impact of form factor, sensors and type of data collected. Clemson: McAdams Hall.

Niedrich, R., & Swain, S. (2003). The Influence of Pioneer Status and Experience Order on Consumer Brand Preference: A Mediated-Effects Model. Journal of Marketing Science vol. 31, 468-480.

38

Okazaki, S., Li, H., & Hirose, M. (2009). Consumer Privacy Concerns and Preference for Degree of Regulatory Control: A study of mobile advertising. Journal of Advertising vol. 38 no. 4, 63-77.

Perlroth, N., & Benner, K. (2016, 03 22). Apple Policy on Bugs May Explain Why Hackers Would Help F.B.I. New York Times.

Peterson, A. (2014, September 26). Privacy is tech's latest marketing strategy. The Washington Post.

Rapp, J., Hill, R. P., Gaines, J., & Wilson, M. (2009). Advertising and Consumer Privacy: Old practices and New challenges. Journal of Advertising vol. 38. no. 4, 51-61.

Reichheld, F. (2003). Loyalty Rules: How today's leaders build lasting relationships. Harvard Business Review Press.

Saeri, A., Ogilvie, C., Macchia, S., Smith, J., & Louis, W. (2014). Predicting Facebook Users' Online Privacy Protection: Risk, Trust, Norm Focus Theory, and the Theory of

Planned Behavior. Journal of Social Psychology 154, 352-369.

Schlegelmilch, B., & Ambos, B. (2004). Multi-utility: Strategic option in Deregulated markets? An Empirical Assessment using Conjoint Analysis. Journal of Strategic Marketing 12, 57-68.

Segan, S. (2016, June 7). The Best Phones of 2016. PCMAG.

Siyavooshi, M., Sanayei, A., & Fathi, S. (2013). SMS Advertising and Consumer Privacy: Analysis of Factors Affecting Consumer Willingness to send and Receive Information in Permission and Data based SMS Advertising. New Marketing Research Journal, 101-124.

Snibbe, Conner, & Markus. (2005). You can't always get what you want: Educational

Attainment, Agency and Choice. Journal of Personality and Social Psychology vol. 88, 703-720.

39

Summers, T., Belleau, B., & Xu, Y. (2006). Predicting Purchase Intention of a Controversial Luxury Apparel Product. Journal of Fashion Marketing and Management vol. 10 iss. 4, 405-419.

Sutano, J., Palme, E., Tan, C.-H., & Phang, C. W. (2013). Adressing the Personalization-Privacy Paradox: An Emperical Assessment from a Field Experiment on Smartphone users. MIS Quarterley Vol. 37 No. 4, 1141-1164.

Tang, Z., Hu, Y., & Smith, M. (2008). Gaining Trust Through Online Privacy Protection: Self-regulation, Mandatory Standars, or Caveat Emptor. Journal of Management Information Systems vol. 24 no. 4, 153-173.

Tapscott, D. (2016, 03 11). After 20 years, it's harder to ignore the digital economy's dark side. Harvard Business Review.

Technology Quarterly. (2014). Hiding from Big Data. The Economist, 1-5.

The Economist. (2015, August 15). Data Privacy, we'll see you, anon. The Economist. The Economist. (2015, March 24). The Wear, Why & How. The Economist.

Timpson, S., & Troutman, M. (2009). 'The Importance of a Layered Policy on all Mobile Internet Sites and Mobile Marketing Campaigns. International Journal of Mobile Marketing Vol. 4. No. 1, 57-62.

Tucker, C. (2014). Social Networks, Personalized Advertising and Privacy Controls. Journal of Marketing Research, 546-562.

Tzanou, M. (2013). Is Data Protection the Same as Privacy? An Analysis of

Telecommunications Metadata Retention Measures. Journal of Internet Law, 1-15. Venkatesh, V., Davis, F., & Davis, G. (2003). User Acceptance of Information Technology:

Toward a Unified View. MIS Quarterly, 425-478.

Weeden, K., & Grusky, D. (2005). The Case for a New Class Map. Journal of Sociology vol. 111, 213-248.