During the 1990s, researchers and policy makers began discussing the presence of a so-called “digital divide,” a separation between people who do and do not have access to information and communication technologies (ICTs). The concept of the digital divide stems from a comparative perspective of relative inequality and depends on the idea that there are benefits associated with ICT usage and negative consequences attending non-usage; inequalities are thus posited to exist between non-users and users. Originally, the term “digital divide” mostly referred to gaps in access to computers. When the Internet became widely accessible in society and began to provide a primary means of computing, the term shifted to encompass gaps in Internet access. Defining the digital divide in terms of access to the Internet is now the usual convention. However, other digital equipment such as mobile telephony and digital television are not ruled out by some users of the term digital divide.

The so-called “Knowledge Gap” is often considered a forerunner to the concept of the digital divide. Developed forty years ago, the Knowledge Gap Hypothesis mainly provides a theory of traditional media. Tichenor, Donohue and Olien (1970) suggested that when the infusion of mass media information into a social system increases, segments of the population with higher socio-economic statuses tend to acquire this information at a faster rate than lower-status segments; this theory relies on the value judgment that more information is better. It is not possible, however, to apply the knowledge gap directly to the Internet, as the use of traditional mass media—on which the knowledge gap focuses—is comparatively straightforward and uniform (Bonfadelli, 2002). The difference in functionality among print media, radio, television and telephone is small compared with the functional differences between computers and the Internet (Van Dijk and Hacker, 2003). While the knowledge gap concerns the differential derivation of knowledge achieved through mass media and focuses, in particular, on mass media’s influence on perception and cognition, the digital divide is potentially more relevant for society with regard to the differential uses of non-mass media in all spheres of daily life.

The development of the term “digital divide” drew attention to the important issue of information inequality in scholarly and political communities at the turn of the century. Countries increasingly realized that the digital divide reduces the potential of the labor force and of innovation. Information and communication technologies were considered to be a growth sector in the economy that should be supported in global competition (Van Dijk, 2008). Between 2000 and 2004, scientific and policy conferences concerning the digital divide were exceedingly popular, but attention to this matter began to decline in 2004 and 2005 (Van Dijk, 2006). On political and policy-making fronts, many observers, particularly those in rich, developed countries, reached the conclusion that the problem was almost solved, as a rapidly increasing majority of their inhabitants obtained access to computers, the Internet and other digital technologies (Van Dijk, 2006). The common opinion among policy makers and the public at large was that the divide was closing between those who did and did not have access to computers, the Internet and other digital media. In some countries, Internet connection rates in households had reached the figure of 90 percent. Computers, mobile telephony, digital televisions and other digital media were becoming cheaper by the day, while their capacity to perform complex tasks increased. These media were introduced on a massive scale and into all aspects of everyday life. Several applications appeared so easy to use that basic literacy was supposedly the sole prerequisite for using them. However, defining the digital divide in terms of physical access to a technology is now considered superficial; physical access alone is no longer considered to correlate to information superiority. Moreover, having physical access is no longer automatically assumed to result in the possession of all the advantages the Internet has to offer. This is explained in the next section.

 

CONCEPTUALIZATIONS OF THE DIGITAL DIVIDE
An important reason for the decreasing attention given to the digital divide in the first decade of the 21st century may lie in the fact that divides in physical access to the Internet are closing in most western countries. Concerns about acquiring physical access to digital media have completely dominated public opinion and policy perspectives in the last two decades. Indeed, these concerns are still paramount, as many people think the digital divide is closed only when more than 90 percent of the population has access to a computer and the Internet. While such a percentage would put the Internet on par with television as a media source, one should note that the diffusion of the Internet in the last two decades has occurred even faster than that of television. Nevertheless, on a global scale, Internet access in 2010 was estimated between 20 and 25 percent of the population, while in many developing countries, Internet access is still restricted to less than ten percent of the population (UN/ITU). Moreover, physical access is not equal to material access. Material access includes all costs related to the use of computers, connections, peripheral equipment, software and services. These costs are diverging in many ways, and people with physical access have very different computer, Internet and other digital media expenses. In this regard, several scholars have pointed toward mobile phones as technologies that address the digital divide. Mobile phones offer a more affordable means of access to the Internet than computers do (Akiyoshi & Ono, 2008). However, one should also keep in mind that mobile phones are by no means a substitute for computers.

Whatever the case regarding mobile phones, it is certainly a misconception that physical or material access to the Internet automatically brings one all the benefits associated with Internet use. Indeed, it is rather superficial to conceive of the digital divide in terms of a binary classification between those with and without physical access to computers or the Internet. Such a belief directly links rates of access to differences in economic capital: one either does or does not have the resources to establish a connection to the Internet. Moreover, such a belief assumes that having a connection correlates with having access to all the advantages the Internet offers. Compaine (2001), for instance, relied on the Diffusion of Innovations (DI) theory of Everett Rogers (1995), who theorized that innovations would spread through society in an S-curve. The S-curve measures the relative speed with which members of a social system adopt an innovation against the length of time required for that innovation to be adopted by a certain percentage of the system (Rogers, 1995). The adoption rate features a point in time, called critical mass; at this point, an innovation has been so widely adopted that its continued adoption is self-sustaining (Markus, 1987). If Compaine were correct in applying DI theory to the digital divide, increases in physical access to the Internet would no doubt correspond to the S-curve measuring the adoption of innovations. From this point of view, the digital divide should steadily disappear as the diffusion rate reaches saturation. The likelihood of the correspondence between the declining digital divide and the increasing rate of innovation diffusion would further be augmented as a result of the migration of the Internet to platforms such as digital television and mobile phones; indeed, the mistaken notion that the digital divide is a temporary problem of physical access has been reinforced by just such migrations (Golding & Murdock, 2001). However, there are serious problems with applying the DI theory to the study of Internet diffusion (Norris, 2001; Van Dijk & Hacker, 2003). One of these problems is explained by Norris (2001), who makes a distinction between normalization and stratification models of diffusion. 

Because it is superficial to assume that physical access to the Internet automatically entails all benefits associated with Internet use, the digital divide should not be considered a divide in physical access only. In literature about the digital divide published after 2000, this conclusion has been suggested using different terminology but similar concepts. Kling (2000), for instance, has suggested a distinction between technical access (i.e., material availability) and social access (i.e., the professional knowledge and technical skills necessary to benefit from information technologies), while Attewell (2001) has distinguished between a first digital divide and a second digital divide. Hargittai (2002), however, has suggested what has become perhaps the most familiar distinction: that between first- and second-level digital divides. A year before Hargittai’s work on first- and second-level digital divides, she and DiMaggio (2001) had also suggested five dimensions in which divides may exist: Technical means (e.g., software, hardware, and connectivity quality), the autonomy of use (e.g., location of access and freedom to use the medium for one’s preferred activities), use patterns (e.g., ways of using the Internet), social support networks (e.g., availability of others for assistance with use and network sizes that encourage use), and skill (e.g., one’s ability to use the medium effectively). Aside from these dimensions suggested by DiMaggio and Hargittai, Warschauer (2003) has also argued that factors such as content, language, literacy, educational level attained, and institutional structure must be considered, while Van Dijk (2005) proposed a causal model with four types of access to ICTs: Motivational access (e.g., the lack of the elementary digital experience by people who have no interest or feel hostile toward ICTs), physical access (e.g., the availability of ICTs), digital skills (e.g., the ability to use ICTs), and usage access (the opportunity and practice of using ICTs).

Each of these scholars shares the following view: while gaps in physical access might be closing in certain respects, other digital divides have begun to grow. In his discussion of motivation access, for instance, Van Dijk (2005) argues that the wish to have a computer and a connection to the Internet precedes physical access. Thus, many of those who remain on the excluded side of the digital divide do so for motivational reasons. According to Van Dijk, in other words, there are not only ‘have-nots’ but also ‘want-nots.’ With the advent of new technologies, motivational concerns are at their strongest and create problems in accepting those technologies. In several European and American surveys conducted between 1999 and 2003, half of those individuals unconnected to the Internet explicitly stated that they would refuse to seek a connection for the following reasons: no need or significant usage opportunities, no time or liking, rejection of the medium (e.g., Internet and computer games viewed as ‘dangerous’ media), lack of money, or lack of skills (e.g. ARD-ZDF, 1999b and a Pew Internet and American Life survey: Lenhart et al., 2003). These observations lead us to one of the most confusing myths produced by popular ideas about the digital divide: that people are either in or out, included or excluded. To explain people’s motivations for using digital technologies, mental and psychological conditions are often mentioned in literature about the digital divide. Here, the phenomena of computer anxiety and techno-phobia have relevance and continue to create major barriers to computer and Internet access in many countries, especially among seniors, people with low educational background and segments of the female population. While these phenomena are decreasing in frequency, they have not completely disappeared with the further diffusion of computers and Internet access.

Lately, Internet skills are increasingly acknowledged in digital divide literature as key factors of the digital divide. Several terms are used to frame Internet skills, e.g. information literacy, computer skills, ICT literacy or web fluency. To date, very little scientific research has focused on the actual level of digital skills possessed by various populations. Many large-scale surveys have revealed dramatic differences in skills among populations, including those populations in countries experiencing the extensive diffusion of new media (Van Dijk, 2005; Warschauer, 2003). Nevertheless, these surveys measure actual levels of digital skills only by asking respondents to estimate their own proficiency. A better way to obtain valid and complete measurements of digital skills is to implement performance tests; such tests would require participants to perform those computer and Internet tasks that are regularly performed in daily life. Hargittai (2002) has begun to implement performance tests in this field (2002). Asking 54 demographically diverse Americans to perform different Internet search tasks, she discovered enormous differences in levels of accomplishment and in the time needed to complete the tasks. In the Netherlands, Van Deursen and Van Dijk (2010, 2011a, 2011b) conducted labor-intensive performance tests in a university media lab on a cross-section of the Dutch population; more than 300 people were tested. Subjects who took the test showed a fairly high level of basic operational and formal skills, but they experienced much more difficulty in processing content-related information and in exercising strategic skills. The results showed significant differences in performance between people of different ages and levels of education. Age primarily appears to be a significant contributor to the basic skills to use the Internet medium-related skills, as younger people perform better on these skills than older people do. In contrast, older individuals performed better where content-related skills, including information and strategic skills, were needed; this occurred in all instances where the older individuals possessed an adequate level of medium-related skills. However, because many seniors tend to lack medium-related Internet skills, they are seriously limited in their content-related skills. Nevertheless, this observation provides another perspective on popular notions about the abilities of the so-called ‘digital generation.’ It also shows that the skills inequality problem will not automatically disappear in the future and that life experience and substantial education of all kinds remain vital for acquiring digital skills.

Aside from divergences in skill levels, the digital divide debate has increasingly drawn attention to the actual usage of the Internet. As a dependent factor, Internet use can be measured in several ways (e.g., usage time and frequency; number and diversity of usage applications; broadband or narrowband use; more or less active or creative use). Since the introduction of broadband access, the measurement of Internet use has become an important subset of the digital divide debate. Broadband is provided by a series of technologies (e.g., telephone wire, cable, fiber optics, satellite, wireless) that enable users to send and receive data over the Internet at larger volumes and greater speeds than were possible using “dial-up” Internet access over telephone lines. Statistics regarding usage time and frequency are notoriously unreliable, as they rely on shifting and divergent operational definitions that are often determined by market research bureaus. These statistics give only some indication of the difference between actual use and physical access. It is certain, for example, that actual use diverges greatly from potential use. Furthermore, those who have a computer and/or Internet connection not always actually using them. Many assumed users actually use the computer or the Internet only once a week or a few times a month; some people never use them. It is important to understand that when a physical access gap for a particular social category closes, the comparable usage gap does not automatically disappear, as well. This discrepancy in gap closures becomes most obvious when looking at types of usage. It is generally assumed that some Internet activities are more beneficial or advantageous for Internet users than others. Some activities offer users more chances and resources to move forward in their career, work, education and societal position than others that are mainly consumptive or entertaining (e.g., DiMaggio et al., 2004; Hargittai and Hinnant, 2008; Kim and Kim, 2001; Mossberger et al., 2003; Van Dijk, 2005; Wasserman and Richmond-Abbott, 2005). In terms of the theories of capital inspired by Bourdieu (1984), one could also say that certain Internet activities allow users to accrue more economic, social and cultural capital and resources than other activities. While some sections of the population will more frequently use those applications that have the greatest advantages for accruing capital and resources (work, career, study, societal participation, etc.), other sections will choose to use those entertainment applications that have little or no advantage for accruing capital and resources (e.g., van Dijk, 2005, Bonfadelli, 2002, Park, 2002, Cho et al., 2003, Zillien and Hargittai, 2009). These differences in types of Internet use call into question the belief that growing up in a digital world results in an intuitive and unproblematic use of digital technologies (see Prensky, 2001). A difference exists between the personal and purposeful uses of technologies such as the Internet. Recently, several scholars have addressed the digital divide by attempting to classify Internet usage types. Some of these classifications take the uses-and-gratifications approach (Katz, Blumler and Gureitch, 1974) as a starting point, while others make use of the Technology Acceptance Model (Davis, 1989) or Social Cognitive Theory. Finally, there are scholars who account for differences in usage by grouping Internet users into use typologies (e.g., Ortega Egea et al., 2007).

 

THE DIGITAL DIVIDE AND INEQUALITY
What is the stake or concern of the digital divide? Do people with no access or limited access to the Internet actually experience disadvantages? A contra-argument might be that people still have pre-existing media at their disposal, which deliver information and provide needed communication channels. For those who have no Internet, many radio and television stations, as well as newspapers, are available, and for those who have no access to e-commerce, physical shops abound. People who desire new social contacts or romantic encounters do not necessarily require a social-networking site or an online-dating service, as the choice of physical meeting places is immense; meanwhile, those individuals who want to make a reservation can still pick up the phone. The issue of pre-existing media aside, another contra-argument is that the non-hierarchical nature of the Internet, in conjunction with the declining cost of computing technologies and their increasing user friendliness, encourages social leveling and undermines existing patterns of class, race, and gender inequalities (see Tambini, 2000). However, most scholars now agree that differences in Internet access among segments of the population cause social inequalities. Witte and Mannon (2009), for example, argue that the Internet is both intertwined with and consequential for inequality. Considering the Internet paradoxical, they contend that it is not only an emblem of a free and open society but also an active reproducer and possible accelerator of social inequality. It is important to understand that the digital divide is not about absolute inequalities between those with and without internet access (Van Dijk, 2006). Most observed inequalities of access to digital technology have a relative nature.
A classification of resources suggested by Bourdieu (1984) is often mentioned in discussions of the Internet’s contribution to inequality. Bourdieu reimagined both Marx and Weber’s ideas of social inequality in industrial society by defining economic, cultural, and social capital. Economic capital reflects Marx’s ideas of both monetary and property assets (i.e., those assets that can be converted directly into money), as well as those other economic possessions that might increase one’s capacities (Hoffman, 2008). Meanwhile, social capital consists of resources based on group membership, relationships, networks, and support; according to Bourdieu, social capital “constitutes the totality of current and potential resources connected to a durable network of more or less institutionalized relations characterized by mutual knowing or acknowledgements; in other words, social capital is a resource based on the affiliation to a group” (Bourdieu, 1984, p. 190). Finally, cultural capital is formed by knowledge, skills, education, and social advantages, all of which give a person a higher status in society. All forms of capital defined by Bourdieu are applicable to the Internet. In fact, Internet use and Bourdieuian forms of capital reinforce each other. On the one hand, all forms of capital affect Internet access: economic capital is a requisite for the supporting means (e.g., a personal computer and a subscription to an Internet provider); social capital is needed to learn how to connect to the Internet, how to use it and how to connect to others using it; and cultural capital is needed to cope with the diverse amount of content available to people of different cultural backgrounds. On the other hand, when these three requisites are met, the Internet begins in turn to affect the three forms of capital: economic capital can, for example, be increased by buying profitable resources online or by finding better jobs; social capital can be augmented by extending physical networks into virtual ones, which can increase civic engagement and a sense of community (Katz & Rice, 2002); and cultural capital can be increased by using the Internet for learning purposes. The result of the mutually determining relationship between capital and the Internet is that some individuals use the Internet in capital-enhancing ways, while others either do not use it or use it in less effective and less profitable ways; the Internet may, of course, also be used simply for leisure (Hargittai & Shafer, 2006; Zillien & Hargittai, 2009). Furthermore, the intensive and extensive nature of Internet use among the well-off and well-educated correlates with an elite life-style, one from which those with less capital are excluded (Van Dijk, 2005; Witte & Mannon, 2009). The result of this last observation is that groups with fewer forms of capital are likely to be affected in negative ways by the Internet: flights will be booked, concerts will be sold out, jobs will be given away, and dates will primarily be granted to those having access. Rather than encouraging equality, the Internet tends to reinforce social inequality and can lead to the formation of disadvantaged and excluded groups (Golding, 1996; Van Dijk, 2005). As Witte and Mannon (2009) contend, Internet access should be understood as a tool for the maintenance of class privilege and power; capitalist relations of production are maintained, as the inequalities upon which they rest are reproduced from one generation to the next.

Mason and Hacker (2003) support the notion that the Internet serves to increase social inequalities. To make this point, they apply Adaptive Structuration Theory, which holds that society, as well as its rules and resources, are reproduced when individuals act in ways that reinforce the systems existing prior to the use of a communication technology: “Those with power and resources outside of the IT context are the primary early adopters of the technology. They use the technology to meet their needs, and the resources and rules they brought into the IT context initially served to shape the roles and rules of those interacting via IT. This serves to reproduce the existing power relations in the social system and even strengthens them, as it opens up a new channel from which those without power and resources are further excluded. … The outcome is a technology that primarily meets the needs of those who adopted it first, and the unintended consequence of this is that those already excluded fall further behind” (p. 48).

Much theoretical work on network societies also suggests that the Internet increases social inequalities. Castells (1999) considers the divide between those who are and are not networked as one of the major axes of social inequality (Castells, 1999). Castells argues that networks are characterized by a space of flows that overwhelms and pervades the traditional space of places. He contends that the particular framing of ICTs in the context of global, informational, and increasingly de-regulated capitalism has been a major factor in the increase of inequality, in general, and of social exclusion, in particular. Because technological access is essential for both the improvement of living conditions and personal development, ICTs deepen discrimination and inequality in the absence of deliberate, corrective policies (Castells, 1999). Van Dijk (2006) defines “network society” as a form of society that organizes its relationships into media networks; these networks gradually replace or complement the social networks of face-to-face communication. For Van Dijk, media networks have become the nervous system of society and are shaping the primary means of social organization and the most important structures of modern society. In a network society, information should be considered a positional good, with some positions creating better opportunities than others for gathering, processing, and using valuable information (Van Dijk, 2006). The positions people have in networks determine their individual potential power; this concept reflects the Weberian idea that human relationships should be considered alongside economic assets when discussing social inequality. When someone in a network society has only a marginal position or no position at all, he or she experiences social exclusion; meanwhile, those who have a central position in society experience social inclusion and thereby increase their power, capital, and resources (Van Dijk, 2006). With unequal positions comes an unequal distribution of positional goods. According to the mechanism of opportunity hoarding, those individuals who have central positions in networks tend to appropriate more goods, to exert more control over particular facilities and to wield greater power within networks; they then capture the returns of these resources and use them to reproduce the boundary between themselves and those who are socially excluded (Tilly, 1999).
In conclusion, most recent scientific literature on the digital divide suggests that the Internet has the potential to strengthen traditional sorts of inequality rather than to ameliorate them. Unequal access to the Internet has varying consequences in several areas of society: the economic (e.g., acquisition and maintenance of jobs), the social (e.g., development and maintenance of social contacts), the political (e.g., voting and other kinds of political participation), the cultural (e.g., participation in cyber-culture), the spatial (e.g., the ability to lead a mobile life) and the institutional (e.g., recognition and attainment of citizens’ rights). 

Inequality among whom?

Because digital exclusion is strongly associated with traditional forms of social exclusion (e.g., exclusion based on socio-economic status, region, and deprivation (Norris, 2001)), differences between several demographic groups can be posited. The use of the Internet was initially restricted to those at the top of existing social hierarchies, and it functioned as a commodity whose distribution tended to follow existing divisions of class, race, and gender (Selwyn, 2006; Van Dijk, 2005; Willis & Tranter, 2006). Even though there is some variation in the scale of difference, the segments of the population that are most likely to be excluded can most usefully be delimited by personal and positional categories (Van Dijk, ). The following personal and positional categorical inequalities can be frequently observed in digital divide research (Van Dijk, ):

 

Personal categorical inequalities
• Age (young/old)
• Gender (male/female)
• Race/ethnicity (majority/minority)
• Intelligence (high/low)
• Personality (extravert/introvert; self-confident/not self-confident)
• Health (abled/disabled).

Positional categorical inequalities:
• Labor position (entrepreneurs/workers; management/employees; employed/unemployed)
• Education (high/low)
• Household (family/single person)
• Nation (developed/developing)

In most empirical observations, the first term listed in parentheses correlates with greater Internet access. However, such a correlation is highly dependent on the type of access in question, as categorical differences in physical access are not equivalent to categorical differences in usage access.