Introduction

In the midst of an arm’s race between cryptographers and hackers, where security and vulnerability clash, personal privacy is caught in the middle of a debate between the convenience of the Internet and the threatened sense of identity management. Pick a newspaper up on almost any day, and you would not be surprised to find a front-page headline on global Internet security that may also have implications for individual beliefs and behaviors. In 2013, the big issue was Edward Snowden’s whistle-blowing on the National Security Agency’s (NSA) mass surveillance program. In 2014, the buzz was all about the vulnerabilities of certain Internet databases, such as Apple’s iCloud, which was hacked causing celebrities’ nude photos to go viral. 2015 heralded the passing of the USA Freedom Act, which legislatively added checks to the NSA, while also reinforcing the prerogatives of the federal government to collect and analyze data on U.S citizens. How do the global and macro issues of Internet surveillance and privacy impact the average Joe?

Since the terrorist attacks in San Bernardino, California, on November of 2015, most of the buzz about Internet privacy and security has been surrounding the FBI-Apple encryption dispute. The FBI had requested Apple’s assistance in unlocking a work-issued iPhone 5c owned by Syad Rizwan, one of the terrorist gunmen. In February, Apple’s CEO, Tim Cook, issued an online statement to Apple users that, from their perspective, what the U.S. government was asking would jeopardize consumers’ right to the assurance of privacy and security.

The issue of Internet privacy is not simple because what it means for one to feel as if one’s privacy has been infringed upon depends on personal reflection and values. While the proliferation and complexity of Internet (and mobile) products have increased social and professional pressures to rely on Internet products, it is yet as easy as checking a box to override personal privacy, anonymity and autonomy. And because the distribution of these social and professional pressures is not necessarily equal across demographics, it makes simple sense that different groups of people are reacting differently, attitudinally and behaviorally, to these challenges.

Literature Review

The literature presented here is broken up into four major subsections. First, the concept of self will be presented through the lenses of multiple disciplines. Then the Internet will be described as a tool affording many novel opportunities for the presentation of self. With opportunity comes risk. Next, privacy models will be described, and the Internet will be considered as one area where personal privacy faces particular obstacles. Finally, literature will be considered about the work already done describing how different demographics engage differently with the Internet.

Offline Presentation of Self, Active and Passive

The construction of a self is a natural and fluid human maturation process. Social, cultural, motivational, and contextual pressures contribute to how one presents one’s self. Attributes of self are communicated both actively and passively.

Connel (1990) proposed his self-determinism theory, which states that the formation of the self is implicated in how well an individual is able to achieve three things: competence, autonomy, and relatedness. Human culture is likewise a driving force behind the forming of self-identity. Kitayama (2010) and others in the field of cultural psychology would describe how geography acted as a barrier promoting in the West and East independent and interdependent self-construals respectively. Gergen (1991) and other social scientists dictate how communication innovations over the past several centuries have stressed different themes over time. The dominant themes of class, religion, and one’s “calling” defined the romanticist view of self, along with other concepts such as honesty, predictability and stability.

In his iconic 1959 book, The Presentation of Self Everyday, Erving Goffman described humans as social creatures balancing the presentation of self differently depending on the situation and context. We behave differently with siblings, grandparents, priests, roommates, and others. Relationships we form are nurtured on unique sets of symbiotic rules. The relationship between a mother and a daughter is different from the relationship that daughter forms with friends. The daughter must traverse different avenues to achieve competence, autonomy and relatedness (Connell, 1991, Connell & Wellborn, 1991, Malley, 2012).

Dean Cocking (2008) explored and expanded upon this concept of having plural-selves, explaining that unique aspects of self are queued by environmental contexts, and at any given time, one is presenting two distinct types of self—the active and passive (Cocking, 2008, Burkitt, 2008, Jamison & Wegener, 2010). The active self is the version of self that is consciously decided upon, maintained, and exhibited by an individual. A decision here is defined as a “commitment to a course of action that is intended to yield results that...satisfy” how an individual wants to be perceived by others (Yates & Tschirhart, 2006, Cocking, 2008). The passive self represents the more intimate self. The passive self includes a person’s thoughts and feelings that s/he might seek to mask from others (Cocking, 2008).

In non-digital environments, an example of the active self could be the design of a t-shirt identifying brand loyalty. The passive-self would be more likely communicated by vocal inflections and facial expressions, or when others know an individual’s identity and history (Cocking, 2008).

However, digital spaces have multiplied the opportunities for one to present one’s self. It is less clear whether individuals apply similar conscious decisions to the presentation of their active self online. And Dean Cocking (2008) fails to attribute the boundaries for what he might constitute the communication of a passive self in an online environment.

The Proliferation of Information Communication Technologies and the Online Presentation of Self

Many scholars have sought to place the importance of our current technological revolution into the context of milestones we have previously achieved. Ubiquitous computing has created information rich environments in which humans co-exist alongside their devices and applications. And social rules and norms have emerged and evolved in digital spaces, though the rules and norms of digital spaces resemble and reflect offline social rules and norms.

Daniel Baym (2010) describes technologies that over the past several hundred years have worked to break down the communication barrier caused by physical space: the printing press, the postal service, railroads, telephones, automobiles and airplanes. The Internet is perhaps the latest communication technology innovation that impacts how humans interact. Luciano Floridi (2014) argues that the societal implications of Information Communication Technologies (ICTs) include having had an impact on how humans think of the self. ICTs today process with ease an amount of data the human brain is incapable of, and humans use ICTs to process our information-laden environment. ICTs and humans, Floridi argues, are both examples of informational organisms (inforgs), “mutually connected and embedded in an informational environment (the infosphere), which [is shared between] informational agents” (Floridi, 2014). Because ICTs are capable of engaging with the infosphere with some level of autonomy and also act as a means for humans to interact with the infosphere, Floridi encourages individuals to reconsider what it is that actually makes humans special.

The infosphere, embedded with new spaces where human users can interact, affords novel opportunities for the presentation of self. Already in the early 1990’s, scholars began studying and describing how humans interact online (Walther & Parks, 2004., Schroeder, 2002., Kraut, 1998., Kraut, 2002). E-communities, where humans would refer to one another by usernames, formed early-on and showed characteristics similar to real-world group dynamics: trust, respect, group norms, slang, rules, and consequences to rule defiance (Baym, 1997). These early communities were some of the first examples of the communication of the self online.

As the Internet grew from infancy to adolescence in the early-to-mid 2000’s, users enjoyed more nuanced opportunities for digital online self-presentation using Social Media Sites (SMSs). Ellison and boyd (2013) define SMSs as digital spaces where users can create a profile, search through a network of profiles, and connect their profile with others.’ The thought-process users can go through to create a profile designed to accomplish a particular goal can be fairly complex. (Ellison & Boyd, 2013). With improved user-interfaces and web browsers increasingly coupling static text and dynamic multimedia content, users enjoyed subsequent increases in the complexity of profile creation and mass-self communication tools (Castells, 2007, Rainie & Wellman, 2012). From hyper-specific E-Communities in the 1990’s to broad, public SMSs in the 2000’s, now SMSs are so prolific that potential users can choose how to spend their time between dozens of feature-specific SMSs. Facebook is the closest to ‘real-life’ (boyd, 2007, Ellison & boyd, 2013, Ellison, Vitak, Gray, & Lampe, 2004, Ellison, Steinfield, & Lampe, 2007). Linkedin is for professional networking. Twitter is for real-time updating (Marwick & boyd, 2011, Hussain, 2012, Shin et al, 2016). And then there are the mobile only SMSs, such as Instagram, which is primarily for mobile photography, Tinder, which is for dating/hooking-up, and Snapchat, which allows for time-limited photo-messaging. Because these platforms are intended for different purposes, users employ different strategies for how they compose their profile (Ellison, Hancock, & Toma, 2011, boyd, 2008., boyd, 2010, Hall, 2015).

It is possible to infer Connel’s (1990) Self-determination theory to digital profile design. The role of choice in using social media and other internet products seems inextricably rooted in social pressure from peers (relatedness) and the strive to maintain an edge against and/or for power and/or counter-power (competence, autonomy) (Connel, 1990, Castells, 2007, Rainie & Wellman, 2012, Hussain, 2012). As with natural human-maturation process, where the taking on of years and life experience forces one to construct an offline-self, individuals are now forced to also carefully consider how to represent the self in the larger infosphere, where the old incentives and power structures of cooperation (Ellison, Steinfield & Lampe, 2011, Scheepers, Scheepers & Stockdale, 2015) and peer pressure (Thompson, 2011), must be considered alongside invisible audiences (Marwick & boyd, 2011, Marwick & boyd, 2014) and the need for digital self-promotion (Gandini, 2011).

It is more complicated to identify what it means for users to communicate a passive self in digital contexts. Digital news distribution channels have completely disrupted print distribution (Dessauer, 2004), and E-commerce has grown into a several hundred billion dollar annual industry (Lorenzo-Romero, et al, 2014). The data trail users leave behind while using the Internet, knowingly or unknowingly, is collected and composed into a behavioral persona, which companies use to serve tailored content. Cass Sunstein (2001) argues that algorithmically-curated content served to subscribers and consumers raises new ethical concerns. If Internet giants are serving up news content particularly suited to draw in users’ attention (Pariser, 2011), couldn’t that lead to the hyper-polarization of ideologies and hive mind (Sunstein, 2001., Pariser, 2011)? And if users are coerced to buy a product they had previously searched for under the appearance of privacy, does that undermine a passive self?

Over the course of the two decades that the Internet and E-communities have existed, we have learned that human communication online tends to resemble offline communication. However, the Internet affords novel opportunities for the presentation and undermining of the self. It is still unclear whether, how, and to what extent Internet users employ traditional strategies for self-identity formation, presentation, and management.

Ethics, Privacy and Consumers’ Challenge

ICTs and the Internet present new ethical dilemmas and corresponding policy, which has led to policy vacuums. Among those affected by policy vacuums, end-users may be the only stakeholders facing personal risk. ICTs are laden with values that impact how the users are forced to confront some of these issues. Privacy is be one example where ethicists and philosophers have made strides in creating a framework broad yet thorough enough to take on new privacy challenges posed by the Internet. And consumers of Internet products have relatively little agency over or understanding about which aspects of their digital profile are amalgamated and repurposed.

In his call for better Information Ethics (IE), James Moor (2006) describes the adoption process of disruptive technologies. As a technologies’ user base grows, more ethical issues should be expected, and when many users share cheap access to, understanding of, and benefit from the technology, the social impact and implications for ethical issues increases (Moor, 2006, Brey, 2000, Tavani, 2002). For example, Brey (2010) posited that “computer systems and software are not morally neutral, and it is possible to identify tendencies in them to promote or demote particular moral values and norms,” which contradicts a popular notion that technology by itself is neutral. Floridi and Sanders (2004) have used a levels of abstraction model to argue that as ICTs become ubiquitous and their level of sophistication more precise, the values embedded in the underlying design of the technology may become increasingly polarized. Take as a conceptual example, a security system that becomes increasingly complex may pose ethical challenges. From being tripped by movement, to being tripped by movement and heat, to being tripped because of human movement, to being tripped because of human movement of a specific race or ethnicity (Floridi, 2008, Floridi & Sanders, 2004, Grodzinsky, 2008).

Privacy is one area of Information Ethics where the function and implementation of value laden ICTs can be considered in conjunction with Floridi’s infosphere. Floridi (2005) argues that one can infer the amount of privacy held by an individual as a measure of ontological friction, or the degree to which “forces oppose the information flow within (a region of) the infosphere” Consider a real-world example to describe offline ontological friction: Two people are on a road walking towards each other from far away, and they are trying to read what it says on the others’ shirt, but because they are too far away, they are unable to. What it says on their T-shirts is the presentation of an active self, the infosphere is the road, and the distance between the two people represents ontological friction. On the Internet, ontological friction could be described as the degree to which an agent is hindered from learning private information about a patient. The friction could refer to the patient’s use of privacy settings, the degree to which their different SMSs are disconnected, or the extent to which other elements of a patient’s online persona remain obscure (Floridi, 2005).

Other theories of privacy have been widely discussed and have been influential, but are limited in scope, making it difficult to cover the broad range of privacy issues on the Internet. One influential privacy theorist, Herman Tavani (2008), makes use of the Restricted Access/Limited Control (RALC) theory, which combines two influential privacy theories into one. The first, called the Restricted Access theory, was introduced in 1983 by Armstrong & Bok. Under this theory, it is up to individuals to create zones (or acknowledge contexts) of privacy for themselves. Privacy is achieved when one creates a zone that prevents others from unacknowledged prying, and where one is shines light only on desired characteristics of self to avoid or obscure access to other potentially passively conveyed attributes. (Armstrong & Bok, 1983, Tavini, 2008, Moor, 1990, Allen, 1988). Gavison’s (1980) Control Theory on the other hand, rather than acknowledging that certain information will inevitably be interpreted, stipulates that individuals inherently own and actively control all information about self (Tavani, 2008, Gavison, 1980). RALC does a good job acknowledging that some information is controlled and some is restricted, but it is insufficiently broad enough to manage IE issues that have arisen on the Internet over the past several years.

Consumers of Internet services are in a predicament trying to balance their use of tools to engage both in the presentation of self and the maintenance of their Internet privacy. With the ease of clicking Agree on 2-dozen page long privacy policies, users enjoy often free access to Internet products and services that (1) make life somehow easier and more convenient, (2) are required, formally or informally by work and education institutions, and (3) are used by family and friends. Danah boyd, among others, has written extensively how young people feel socially obligated and pressured to participate on SMSs (Jenkins, Ito & boyd, 2015). Helen Nissenbaum (1998) discusses the difficulty of protecting users privacy while they willingly provide information, both by signing up to use the service and by creating their SMS profiles, which although being guarded to some degree by privacy controls, remains relatively easy to ascertain by predators, phishers, hackers. Chhabra, et al (2011) pointed out that companies make billions in annual profit from willingly given information: search queries, product purchasing history, and demographic information. Using this information, creative companies are able to repurpose it a variety of ways, including serving up targeted advertisements. Wallace (2008) and Doyle (2009) question whether the benefits gained by users of Internet products and services outweigh what might also be viewed as intrusive, or even voyeuristic, behavior.

While Floridi’s (2008) ontological model of privacy is useful as a conceptual framework for thinking about how the flow of information in the infosphere relates to one’s personal privacy, it is unclear to what extent users of Internet products care about privacy sufficiently to impact their Internet behaviors. And without first understanding the behaviors of Internet users, it may be difficult to informedly fill policy vacuums and expectations of the companies behind popular products and services.

Millennials’ Digital Literacy and other Demographic Factors

The so-called Millennial Generation is the first cohort of present-day adults to have been born into and grown up in the presence of ubiquitous Internet technologies. The Baby boomer generation is considered the generational cohort born in the years after WW2. While Baby boomers are generally characterized as valuing honestly, stability, and consumption (Baby boomers, 2016), Millennials are more often observed prioritizing convenience, experiences, and practicality.

The first wave of Millennials matured into adulthood during the first years of the 21st century. Scholars, marketers, and content producers raced to study what makes this generation novel. Educating and working with Millennials, more than other generations, require technological adeptness among educators and supervisors (Barnes, Marateo, & Ferris, 2007., Hargittai, 2010). And Millennials are also more likely to participate in E-commerce (Thompson, 2003., Leung, 2003., Obal & Kunz, 2013). Boyd and Hargittai (2013) characterize Baby boomers, on the other hand, as more cautious and even fearful in their ICT use-strategy and approach, though what’s feared varies situationally.

Millennials are particularly well-equipped to engage with ICTs and the Internet, and the digital divide between generations. Though less well research, anecdotal evidence suggests, however, that digital literacy may not be as simply qualified as an age-gap. In his article, What Happens When Kids Don’t Have Internet at Home, The Atlantic writer Rachel Monahan (2014) describes how in Kent School District in Kent, WA, where students are issued laptops in 7th grade, much of the curriculum has gone digital. Generally this would be thought to improve digital literacy. However, about 2500 kids leave their laptops at school, lacking Internet at home, which paradoxically could further widen the digital literacy gap. This anecdotal example suggests that income and access to educational materials may differ the exposure to ICTs and possibly impact digital literacy.

Danah boyd's research has substantially qualified the idea that differences have emerged in how Millennials and Baby boomers approach ICT and the Internet. And, as Nasah et al (2010) explains, digital literacy is linked more closely to ubiquitous exposure than purely to generational differences. However, there is not substantially less empirical evidence quantifying these differences. There is even less literature interested in exploring the relationship between attitudes held about the Internet and the behaviors engaged in by those same individuals. Moreover, any proposed relationship between digital literacy and demographic differences beyond age have not at all been explored.

Research Questions and Hypotheses

Internet users must negotiate their online presentation and maintenance of self with an affinity for privacy as a concept. However, even in instances where Internet users have a strong desire to possess Internet privacy, they may be left with few useful options. Social media users use will continue to post statuses and share semi-public personal information. Companies will continue to create personalized online personas of potential customers. And more often than not, the amount of ontological friction upholding privacy in the infosphere will only ever minimally protect users.

The fact that users are using these products and services in the first place suggests that users are dealing with issues of privacy on their own terms. The purpose of this research is to better understand the relationship between the desire for increased Internet privacy and actual Internet behaviors. I hypothesize that those who express attitudes reflecting concern for privacy are also those more likely to engage in behaviors that promote Internet privacy.

A secondary purpose is to explore how this relationship varies, if at all, between demographics. It is plausible based on the above literature that attitudes and behaviors towards the Internet and Internet privacy may differ based on exposure to, understanding of, and strategy for engaging in Internet products. And while Generational differences between Millennials and Baby boomers are the most likely to be different, this research will also explore the differences between other demographic variables.

Methodology:

Data Source

Survey data was collected and provided by Pew Research Center, a reputable and well-established social scientific research non-profit. Pew Research Center is known for collecting survey data from large, diverse and representative samples and releasing the raw data as a resource for scholarly research. The data presented here was collected by Pew in 2013 as a part of an omnibus survey on perceptions of online anonymity. After some exploratory analysis, the Pew survey data seemed especially capable of generating my research questions. The sample demographic consisted of over 1000 random, adult (18+) telephonic respondents, of which about half of respondents were contacted via cell phone.

Pew also often releases summary reports of their survey data. While it is not uncommon for them to release more than one of these reports per survey, the data set used here produced only one report, called Anonymity, Privacy, and Security Online. This report discloses general statistics and reporting patterns, but it neither does a thorough job of breaking down responses by demographic, nor does it go very far beyond the purely observational. For example, the report published by Pew, while well done methodologically, is not intended to stipulate meaning from the statistical observations.

Data Preparation

This report investigates and utilizes a carefully selected subset of questions coded for yes/no from the omnibus survey. Questions were initially selected based on exploratory analysis as well as having a basis in one of two question types. For a list of all of chosen questions and their exact wording, please refer to Appendix A.

One type of selected tangentially related questions sought to gauge individuals’ attitude towards Internet privacy and anonymity. For example, does the individual worry about the amount of freely available personal information online? Does the individual perceive current privacy laws as sufficient or insufficient? And does the individual believe that Internet users should have the right to use the Internet anonymously? In one instance, a question having to do with what individuals thought about current privacy laws was recoded in order for the coded answer for feeling as if laws are inadequate to reflect other, similar questions’ coding. (However, no statistically significant findings were uncovered through analysis of this particular question.)

The second set of questions sought to gauge the extent to which an individual has taken action on the Internet to maintain and protect some aspect of their Internet privacy. Some of these behavioral questions related to one’s personal privacy and the active presentation of self. For example, has the individual ever used a fake name or otherwise concealed her/his identity on the Internet? Other behavioral questions dealt with the notion that individuals can leverage Internet tools to maintain or protect what I’m calling one’s passive presentation of self online, the trail of data and metadata following many of the actions taken by users and often repurposed by Internet companies for some particular fiscal gain. For example, a question might gauge users’ use and disuse of cookies, which could be inferred as a means to avoid tailored Internet advertisements. For a full list of questions selected for the purposes of this research, please navigate to Appendix A below.

A third set of demographic questions was used to tease out potential differences between four demographic variables. The full dataset was stratified based on these four variables: age, education, income, and political affiliation. Age is a proxy variable for generational cohort; survey respondents born between 1945 and 1964 are ‘Baby boomers’ (n~390); those born between 1980 and 2000 are ‘Millennials’ (n~209). This clustering corresponds with commonly understood definitions of these respective cohorts. The variable for education attainment is a proxy for a more complex and difficult to measure potential for competency and exposure to ICTs; individual respondents are grouped based on whether s/he had achieved some college accreditation or degree (n~487) or not (n~511). Income is an imperfect, but potentially viable proxy for socio-economic status. Respondents are grouped into broad categories without regards for geographic location: lower annual income (<$30,000; n~287) and higher annual income (>$100,000; n~290) Lastly, self-reporting on political affiliation is a used here as a fairly rudimentary proxy for broader ideological differences. Pew made no effort to assess ideological stances on socio-technical issues, though it seems possible that strong stances could potentially impact certain attitudes and behaviors related to Internet privacy: Republicans (n~260) and Democrats (n~280).

The number of respondents for each demographic vary slightly from question-to-question based on the willingness of the sample to answer each optional question. Refer to Table 1 for detailed question-by-question sample demographic information. Neither Income nor Political Affiliation yielded any significant results and so are neither and reported in Table 1.

Table 1: Number of respondents on a question-by-question basis

Data Analysis

Data analysis was conducted using R Statistical Programming software, which produced descriptive output, cross-tabulations of variable pairs, and tests of significance related to the distribution responses for any given variable pair. Analyses were conducted several ways. First, question-pairs were analyzed using responses from the sample as a whole to uncover whole-sample distributions (See Table 2 and Table 3). Next, individual questions were analyzed intra-demographically (i.e. Baby boomers vs. Millennials) to understand the extent of difference between how different demographic answered individual questions differently (See Tables 4, 5, 6, 7, 8, and 9) Then question-pair analysis was conducted intra-demographically (i.e. Baby boomers vs. Millennials) to better unpack the relationship between attitudes and behavior (See Tables 5.1, 6.1, 7.1, 8.1, 8.2, and 9.1). Chi-squared, Tests of Homogeneity and Independence was the statistical test used on resulting data matrices to measure the extent of significance. For the purposes of the research presented here, a significance level of less than 1% was required to consider the finding significant.

Results

This section lays out the data analysis in a series of [ ] results tables and provides a table-by-table analysis of the primary findings. The first two tables shows general differences in reported behavior between all respondents across the entire Pew sample who either do and do not express worry about the amount of personal information about them online, and between all respondents who either do or do not think that people should be able to use the Internet anonymously. The next table investigates generational differences between Millennials and Baby boomers on the issue of anonymity. A series of 5 tables explores generational differences in engaging in pseudonymous Internet behavior and the role that worry plays. Then a series of two tables explores generation differences on avoiding Internet advertisers. The next set of 5 tables explores generational and educational differences in how Internet users approach Internet cookies.

Worry and Desire for Anonymity, Privacy for Active-self Online

Concealing one’s identity online is a primary strategy for increasing personal ontological friction while navigating digital spaces. Table 2 compares those who reported having engaged in different privacy-enhancing behaviors with those who have not on a basis of whether they also (1) reported worrying about the amount of publicly available personal information online, and (2) reported having a belief that people should have the right to use the Internet anonymously. Privacy-enhancing behaviors in this context means the individual reported having done one of the following in the past: used the Internet in a way that hid or masked his/her identity, used a fake name online, and gave misleading information about self online.

Table 2: Using the Internet while maintaining privacy for one’s Active-self

Table 2 demonstrates that those who worry about their personal privacy and those who believe in the right to use the Internet anonymously are significantly more likely than their counterparts to have engaged in behaviors thought to help maintain their personal privacy. It is perhaps not surprising that individuals that do not worry and that do not care about anonymity are less likely to engage in privacy-enhancing behaviors. While not necessarily the main point to be communicated through Table 2, it may also be relevant to note that while those who worry make up about half of the sample respondents, significantly more respondents believe in the right to anonymity than those who believe that people should not be able to use the Internet anonymously.

Worry and Desire for Anonymity, Privacy for Passive-self Online

While quite thorough in his description of the active presentation of self online, Dean Cocking (2008) is much less explicit on what exactly constitutes the presentation of the passive-self in the context of digital environments. Increasingly, data driven Internet companies collect and amalgamate troves of personal information related to Internet searches and other behavior. The data is often repurposed to tailor advertisements to a user at the cost of their privacy and wallet. For the purposes of this research, the online passive-self, the self that is composed of “conflicting, less chosen and controlled” attributes, will be considered communicated online through the trail of data and metadata trailing most online activity. In this report, the passive-self in the digital environment refers to the trail of data and metadata that follows invisibly behind Internet users’ actions. One concrete way to limit the collection and repurposing of this type of data by Internet companies is to leverage certain browser options, such as deleting and disabling cookies. Table 3 compares those who have reported engaging in these types of behaviors with those who have not on the basis of whether the respondent also reported, as above, (1) worrying about the amount of publicly available personal information online, and (2) believing that people should be able to use the Internet anonymously.

Table 3: Using the Internet while maintaining privacy for one’s Passive-self

Table 3 demonstrates that those who worry about their personal information and those that believe in the right to be able to us the Internet anonymously are significantly more likely than their counterparts to have engaged in behaviors intended to increase Internet passive-self privacy. It is important to note that compared to Table 2 above, it is in general much more common to have used browser tools to both clear and disable cookies. Because the online passive-self is often less tied to one’s offline identity, it may be that individuals are more able to conveniently engage in privacy-enhancing behaviors without the social stigma that may accompany active-self privacy-enhancing behaviors, such as using a fake name on Social Media Sites (SMSs).

Millennials and Baby boomers and the Right to Anonymity

The above two tables demonstrate differences across the entire Pew Research sample. However, by sub-setting the data by demographic, it is possible to identity more minute and otherwise muted trends concerning how individuals as members of a particular cohort may be differently approaching issues surrounding Internet privacy.

Given danah boyd’s extensive qualitative research on generational differences between Millennials and Baby boomers, the Net Generation, and digital literacy, it is not surprising to find quantitative results suggesting differences between how Millennials and Baby boomers perceive issues such as Internet privacy and anonymity. Table 4 compares Millennials and Baby boomers on their reported attitude about whether or not people should have the ability to use the Internet anonymously.

Table 4: Millennials and Baby boomers, the right to use Internet anonymously

A significantly larger portion of Millennials (71.7%) than Baby boomers (57.7%) share the opinion that users of the Internet should be able to use the Internet anonymously. This generational difference may provide supporting evidence for what “digital literacy” means in practice: a high comfort level with ICT use and confidence in taking advantage of the affordances of anonymity. It is also plausible that the ubiquitous exposure of Millennials to varied digital spaces might encourage Millennials to identify differently with public and intentional SMSs and information search tools such as Google. The relationship between Baby boomers and ICTs, on the other hand, is characterized by foreignness. Though Baby boomers in general have widely adopted ICTs, less life long exposure to such technologies comparative disadvantages Baby boomers. While several factors likely impact Baby boomers attitudes towards ICTs, it is possible that those attitudes reflect societal values where individuals are less concerned with issues of privacy. One possible interpretation stipulates that Baby boomers, not entirely grasping the functional properties of the Internet, perhaps conflate online anonymity as a right with someone having something to hide.

Millennials using the Internet in a way that masks their identity

Concealing one’s identity on the Internet is a primary strategy for protecting privacy on digital spaces where one must actively present their self. Table 5 compares Millennials and Baby boomers on a basis of whether or not they have reported having used the Internet in a way that conceals their identity.

Table 5: Millennials and Baby Boomers and concealing one’s identity online

Millennials (27.1%) are significantly more likely than Baby boomers (13.2%) to have tried to use the Internet in ways that conceal their personal identity. Perhaps sheer exposure to more opportunities for the digital presentation of self, most clearly exemplified through their use of Social Media Sites, but also in forums, E-communities, video gaming, and other contexts, has encouraged in Millennials a more developed understanding of digital etiquette, where on certain online environments the use of anonymity or identity concealment is normative. Table 5 establishes and supports the hypothesis that a measurable difference exists between Millennials and Baby boomers and maintenance of active-self privacy on the Internet.

Previous tables have demonstrated the role that worrying plays in encouraging an Internet user to engage in behaviors thought to enhance Internet privacy. For privacy-enhancing behaviors thought to increase privacy for one’s active self online, those behaviors are to be relatively uncommon when looking at the Pew sample population as a whole. Even among those who worried, only one in five had used the Internet in a way to conceal his/her identity. Table 5.1 is a parsed-out, more targeted comparison between Millennials and Baby boomers that reported worrying about the amount of publicly available personal information online on the basis of whether or not they reported having masked their identity online.

Table 5.1: Worrying and concealing one’s identity online

Millennials (38.8%) that worry about the amount of publicly available personal information online are significantly more likely than Baby boomers worriers (16.8%) to have attempted to use the Internet is a way that masks their identity online. Comparing Table 5 and 5.1, while worrying about personal information greatly incentivizes Millennials’ decisions to conceal their identity online, Baby boomers remain distinctly less likely to have attempted to conceal their identity. Perhaps Baby boomers’ relative lack of exposure to opportunities for online presentation of self limits either their technical capacity or general desire to conceal their identity online.

Millennials, Baby Boomers, and Pseudonymity

Dead Cocking (2008) describes the active presentation of self online as almost entirely voluntary process, where users are able to keep many aspects of the self private. Floridi’s (2008) Ontological Model Privacy views privacy as a measure of ontological friction present in the infosphere. Giving oneself a fake name online or otherwise misconstruing personal information is one possible strategy to increase the amount of ontological friction in SMSs, message boards, blogs, comment threads, and other frequented digital spaces. Table 6 compares Millennials and Baby boomers on the basis of reported pseudonymous Internet behavior.

Table 6: Millennials and Baby boomers use of pseudonymous behavior

Table 6 supports the notion that Millennials are significantly more likely than Baby boomers to have attempted to leverage pseudonymous behavior to protect aspects of their identity online. As active SMSs users, Millennials are more sternly and ubiquitously confronted with dilemmas presented by invisible audiences. One knows not whose prying eyes peer and prey over personal data or why. After all, as danah boyd (2008) and others have observed, when it is easy and convenient, young ICT users have been known to warp the affordances of the technology to solve some certain problems using novel methods. Maintaining a semblance of privacy may be one such problem.

It is perhaps surprising that no significant differences were found between the proportions of Millennials and Baby boomers that reported worrying about the amount of openly available personal information about them on the Internet. It is difficult to suppose what makes either Millennials or Baby boomers worried about their personal information. However, it may be possible to better understand the role and relationship between worrying about personal information and engaging in pseudonymous behavior. Table 6.1 compares Millennials and Baby boomers who reported worrying about the amount of personal information available online on the basis of whether or not they have engaged in certain pseudonymous behaviors.

Table 6.1: Millennials and Baby boomers, worrying and pseudonymnity

Table 6.1 demonstrates that not only are Millennials more likely to engage in protecting their privacy, but that worrying about the amount of one’s openly available personal information on the Internet increases the likelihood of an individual to engage in privacy-enhancing behaviors. Millennials, especially worried Millennials, are leveraging pseudonymous behavior as a means of increasing at least the feeling of privacy (or the amount of ontological friction). However, for Baby boomers, worrying about one’s personal information did not make them more likely to have engaged in pseudonymous behavior. The ability of Millennials to engage in this way could be driven by digital literacy, as could the inability of Baby boomers to do so. After all, Millennials are more ubiquitously exposed to Internet environments where there may be some social or personal gain in using a fake name or obscuring personal information.

Passive-Self Presentation and Avoiding Online Marketers

As stated above, for the purposes of this research, the online passive-self will be considered the self composed by the amalgamation of personal data and metadata trailing Internet activity. There are many little behaviors one could engage in to make personal data more obfuscated and it more difficult for Internet companies to collect and leverage these passive qualities of self. Table 7 compares Millennials and Baby boomers on the basis of whether or not they reported having tried to use the Internet in a way that keeps advertisers from being able to see what they have read, watched, or posted online.

Table 7: Millennials and Baby boomers avoiding advertisers

Table 7 demonstrates that Millennials (40%) are significantly more likely than Baby boomers (26.2%) to have attempted to avoid online advertisers. This finding could also stem from the comfort and ease with which Millennials use ICTs and the Internet. Because online advertisers almost always attempt to use aspects of one’s passive-self to compose potential customer personas and tailor advertisements, having the ability to manage Internet use in a way that minimizes contact with online marketing is one possible strategy for minimizing person exposure. Although one might have assumed that digital literacy alone could promote the privacy-enhancing Internet behaviors among Millennials, Table 7 suggests that digital literacy, or lack there of, may also act as a barrier preventing Baby boomers from engaging in privacy-enhancing behaviors.

Table 7.1 compares Millennials and Babyboomers who reported worrying about personal information online on the basis of whether or not they have attempted to avoid keep advertisers from being able to see what they have read, watched, or posted online

Table 7.1: Worrying while trying to avoid Internet advertisers

Millennials (53.8%) that worry about personal information are significantly more likely than Baby boomers (30.7%) that worry to have attempted to avoid Internet advertisers. Worrying also significantly increases the likelihood that Millennials will have attempted to use the Internet in this way. Given the perceived advantages and effects of digital literacy, it is perhaps not surprising to have found that worrying promotes having tried to avoid online advertisers more for Millennials more than for Baby boomers. However, the notion that worrying about the amount of available personal information online does not greatly alter the likelihood that Baby boomers are able to engage in privacy-enhancing behaviors could, ironically, be for them worrisome.

Cookies: Generational & Educational differences

One typical strategy for avoiding online advertisers is through the frequent clearing of cookies and browser history. Cookies are small text files created by websites that allow websites to recognize, track and store users’ Internet behavior. It should not be a surprise that cookies play a big role in E-commerce and tailored advertisements. Table 8 displays data from two demographics, generation and education. Table 8 compares Millennials and Baby boomers and individuals who have and who have not received a college degree on the basis of where they reported having cleared their Internet browser’s cookies.

Table 8: Generational and educational differences in clearing cookies

By this point, generational differences between Millennials (78%) and Baby boomers (61.3%) has been established and communicated through many tables in this report. Digital nativism or near-nativism equips Millennials significantly better to engage and interact with Internet tools and operations. Deleting cookies is one strategy employed by Millennials enabling them to evade Internet advertisers. And in terms of the presentation of self, deleting cookies might be viewed as just one of many examples where Millennials’ digital literacy allows them to exist on the Internet with more autonomy and anonymity.

College and cookie management go hand-in-hand. Those who have attended college (71.9%) are significantly more likely to have deleted their Internet cookies or cleared their browsing history than those who have not received a college degree (56.6%). This could stem for a variety of reasons, and this report does not intend to stipulate one reason as necessarily more likely than another. It is true that today college requires nearly ubiquitous use of personal computing both to remain competitive against peers and to keep up with the rigor demanded by these higher education institutions. However, that was not necessarily the case for those college graduates that are also Baby boomers. College as a stimulating environment encourages critical thinking and problem solving, so it is possible that those who have attended college are simply more aware of their desire for autonomy and anonymity online.

Table 8.1 compares Millennials and Baby boomers that reported worrying about the amount of personal information online on the basis of whether or not they also reported having cleared their cookies and browsing history.

Table 8.1: Worrying and clearing one’s cookies

In congruence with many of the findings presented in this report, Table 8.1 demonstrates that Millennials (78.5%) who have reported worrying about the amount of publicly available personal information are significantly more likely than their worrying Baby boomer (67.4%) counterparts to have reported clearing their cookies and browsing history. However, worrying did not significantly impact the proportion of Millennials that reported clearing their cookies, though it did modestly increase the proportion of Baby boomers that reported doing so. Even without worrying, 78% of Millennials reported clearing their cookies, so it is possible to surmise that the act of clearing one’s cookies may be one behavior already ubiquitous among Millennials.

While there was no significant difference between those worriers who did and did not receive a college degree, the belief that people should have the right to anonymity drove further both generational and educational significant findings. Table 8.2 compares Millennials and Baby boomers and those who have and have not received a college degree on the basis of whether or not they Internet users should have this right to use the Internet anonymously.

Table 8.2: Right to anonymity and clearing one’s Internet cookies

Consistent with the findings presented in Table 8, Table 8.2 suggests that those who are clearing their Internet cookies and browser histories generally tends to be those that believe Internet users should have the right to browse the Internet anonymously. This is a clear example of individuals behaving in accordance with their attitudes. And although difference in income alone did not reveal itself as a good predictor of attitudes or behavior, it is important to note the relative correlation between age, college and income, particularly for questions whose subject matter revolve around cookies.

Disabling cookies, perhaps requiring commitment

From a user experience perspective, it is important to note that cookies generally enhance the experience of browsing the Internet. Cookies enable auto-fill passwords and other forms, and for those that participate in E-commerce, tailored advertisement may be helpful. Already, Millennials as a generation have been characterized as prioritizing convenience, so it may not be too surprising then that while the vast majority knows how and have cleared their cookies, there is no significant difference between the proportion of Millennials and Baby boomers that reported disabling cookies entirely. However, educational differences seem to be a more robust predictor of cookie-disabling behavior. Table 9 compares those who have earned a college degree with those who have not on the basis of whether or not they have set their browser to disable cookies.

Table 9: College grads disable cookies

Those who have earned sort of college degree (50.6%) are significantly more likely than those who have not (40.5%) to have set their browser to disable Internet cookies. While the act of clearing cookies is relatively simple and lacks permanence, the decision to disable cookies can perhaps be considered both more permanent and more decidedly against what cookies represent and how cookies function. Disabling cookies can have user experience ramifications because many Internet companies’ products rely on cookies to save password, searches, and tailored content.

Table 9.1 compares those who earned a college degree with those who have not and that either worry about the amount of available personal information online or believe users should be able to use the Internet anonymously, on the basis of whether or not they reported having set their browser to disable cookies.

Table 9.1: Worry, the right to anonymity, and disabling browser cookies

College graduates 57.7%) who worry about privacy are significantly more likely than non-college graduates (42.2%), and college graduates (55%) that believe users should be able to use the Internet anonymously are significantly more likely than non-college graduates (42.5%) to engage in privacy-enhancing. However, neither worrying about personal information nor the belief in the right to online anonymity greatly changes the likelihood of having disabled an Internet browser’s cookies. This caveat is important to understanding the limitations of this report. At no point can one make serious conjectures at answering the why behind a particular privacy-enhancing behavior. Digital literacy clearly has something to do with the ability and/or desire for an individual to engage with his/her Internet privacy, but worrying about personal information and believing in the right Internet anonymity could be more descriptive characteristics than motivating factors.

Discussion

The issue of Internet privacy is one with more questions than answers. Scholars in a wide-range of academic disciplines, from psychologists to Information ethicists, from philosophers to social scientists, are tackling the issue from many vantage points. A broad, overarching conceptual goal of this research was to better understand how individuals are navigating and negotiating their presentation of self in digital environments. At this point in 2016, the Internet is colossal and the opportunities for the online presentation of self are vast. By using data from a reputable Pew Research survey, the research presented here provides a quantitative window into individuals’ attitudes towards Internet privacy and behaviors they have taken to maintain a semblance of their Internet privacy.

Strong Attitudes Encourage (or relate to) Privacy-Enhancing Behaviors

Holding attitudes that reflect worry or entitlement encourages users to be more mindful of and active in bolstering of their Internet privacy. Individuals who worry about their privacy online, for example, are significantly more likely to have engaged in pseudonymous Internet behavior such as using a fake name or providing misleading personal information. Individuals who worry are also more likely not only to regularly clear their Internet cookies, but also to disable them altogether.

Generational Differences and Digital Literacy

The results presented in this report also support the notion that Millennials are significantly more equipped to deal with the problems surrounding maintaining their Internet privacy. A lot of their preparedness is likely rooted in digital literacy. Millennials, after all, are ubiquitously exposed to using ICTs and the Internet not only on a daily basis, but also as a part of more social and professional encounters. And along with this ubiquitous exposure, technical savvy Millennials are simultaneously exposed to Internet rules and norms. Driven by norms, sticky privacy-enhancing behaviors engaged in by Millennials can be viewed in this light as business-as-usual. Armed with digital literacy, Millennials who worry about their personal information or desire to use the Internet anonymously are even more able and likely to have engaged in privacy-enhancing behaviors.

Conversely, it is plausible that while digital literacy is making possible these sorts of options for Millennials, the lack of digital literacy actually may act as a barrier preventing Baby boomers from engaging in privacy-enhancing behaviors that would suit their interests. Imagine Millennials who worries about the amount of personal information available about them on the Internet, but who are able to do something about it, be it using a pseudonym or attempting to avoid the prying eyes of Internet advertisers. Perhaps these behaviors lend the Millennials some peace of mind. Now imagine Baby boomers with that same worry who, lacking the know-how to engage in those same privacy-enhancing behaviors, end up doing nothing productive. It is possible that, unable to do much about it, they become more and more worried.

College Education and Digital Literacy

This report has also highlighted significant differences between behaviors engaged in by those who are and are not college-educated. It is more than plausible that digital literacy is not simply a set of skills that comes to fruition simply because of an individual’s age. More likely, digital literacy comes about through continued, prolonged exposure with ICTs and Internet tools. It is possible that those with a college education also engage more ubiquitously with ICTs in professional settings. It is also possible that for those with a college education, ICTs are more affordable and accessible in general. Another explanation could be that critical thinking skills learned in college encourage more proactive privacy-enhancing behavioral tendencies.

Active and Passive Presentations of the Digital Self

Another goal of this research was to interrogate the notion that certain privacy-enhancing behaviors would more fully benefit either the active-self or passive-self. For the purposes of this research, what is meant by the active presentation of self in the context of digital environments relates to those digital spaces where information about self is communicated highly voluntarily, such as on SMSs. As a whole, engaging in behaviors thought to enhance privacy for the active self, such as using pseudonyms or providing false of misleading personal information, was not very engaged in. From the whole population sample, only about 20% reported having engaged in those types of behaviors. When parsed by demographic, Millennials had engaged in such behaviors the most by far, and even among Millennials who worry about personal information, fewer than half had engaged in pseudonymous behavior. One interpretation could be that SMSs are often coupled with real-life offline networks, so individuals are wary of audiences who might consider pseudonymous behavior inappropriate.

Still, the proportions of those engaging in active-self privacy-enhancing behaviors is low, which begs consideration to how exactly respondents interpreted the survey questions. Consider, for example how respondents might have interpreted what was meant by “While you have used the Internet, have you ever used a fake name?” If one were only thinking in terms of public SMSs, perhaps these low figures make sense. Considering the plethora of online E-communities where ambiguous usernames are normative, one might have expected more Internet users to report using a fake name online. This could highlight a disconnect between how users are thinking about self online, where it might not be common for individuals to consider their online avatars as representations of self.

In the context of digital environments, this research considers the passive-self as a reference to the involuntary trail of data trailing much of a users’ Internet activity. Taking steps to attempt to protect this more private version of digital self was evidently far more common. For example, a majority of respondents reported having cleared one’s cookies, a common defense against unsolicited Internet advertising. It is possible that enhancing one’s privacy in relation to their passive-self is a more tangible and convenient task. It does not require deep contemplation on social expectations and audience reception. Rather, it requires a simple baseline understanding of how to operate ICTs and Internet browsing tools.

Political Ideology and Income

The analysis of demographic characteristics other than generational cohort and educational level yielded few significant results. As described in the methodology section above, the same analysis methods were run on sub-sets of the Pew Research surveyed sample. Reported U.S. political affiliation was chosen as an imperfect proxy reflecting the large and polarized ideological divide between the two main U.S. political parties, Republicans and Democrats. Income, parsed out between those that make under $30k and over $100k was intended to serve as a less than perfect proxy to reflect Socio-economic status. In almost all instances, no significant differences were found intra-demographically between either of these groups.

Conclusion and Limitations

Given the exponential explosion of ICTs and Internet products and services over the course of the past couple of decades, we find ourselves in a polarized world ripe with policy vacuums, ethically convoluted corporate practices, and not a lot of answers about how to deal with issues of Internet privacy. And yet users of ICTs and Internet products and services confront these types of issues on a personal basis practically every day, both consciously and unconsciously. The research presented here sought to contextualize individuals’ confrontation with certain facets of these issues related to Internet privacy by repurposing reputable and nationally-representative data, which gauged a selection of attitudes and behaviors related to Internet privacy and anonymity.

In a perfect world, questions and responses analyzed would be specifically designed and tailored to answer my specific research questions. However, it would have been impossible in the time allotted for an understand honors thesis to produce a national dataset as reputable and robust as Pew’s. Moreover, only a subset of questions were selected from the Pew survey, and only a subset of data tables were presented in the Results, making it difficult to separate myself from whatever unintended personal biases that I personally hold.

Moreover, performing logistic regression could have further parsed out and bolstered the findings reported here, especially by highlighting the extent of inter-demographic confoundedness. However, purposes of this report were to explore the correlative relationship between attitudes and behaviors of Internet users. While logistic regression could have helped informed a causal understanding between attitudes and behaviors, it did not seem necessary to determine causality because of the high correlation that was uncovered.

Still, the research presented here strongly encourages further research in two specific areas. (1) Future research could further explore more minute aspects of how individuals are dealing with the task of their digital presentation of active- and passive-selves while simultaneously engaging in privacy-enhancing behaviors. (2) Future research could further quantitatively explore demographic differences especially between how Millennials and Baby boomers are dealing with issues related to their overcoming or leveraging of digital literacy, related to their strategies for their digital presentation of self, and related to Internet privacy.