AN INTERNATIONAL JOURNAL
so-called “digital divide” and if there is, how important it
might be. In turn, the question is whether it will close or
widen in future years. Much of this discussion is politically charged. Old views reappear about markets and people who are supposed to solve all problems by themselves,
or not, and about the need or rejection of government intervention. In this article we attempt to postpone this ideological discussion in order to develop scientific conceptual
distinctions and to present reliable and valid empirical data
of longitudinal research on this subject matter. The concepts of digital divide, access, adoption of innovations following S-curves, and so on are carefully explicated. The
most reliable and valid data from the United States and
European countries are summarized. Finally, some results
of a large-scale official social survey in the Netherlands are
presented, as it is one of the few that tried to go beyond the
usual demographic background variables elaborating multiple regression models for the explanation of differences
found in these background variables.
The firstobstacle in research and discussion on information inequality is the multifaceted concept of access.
It is used freely in everyday discussions without an acknowledgement of the fact that there are many divergent
meanings in play. The meaning of simply having a computer and a network connection is the most common one
in use today. However, according to Van Dijk (1999), this
meaning only refers to the second of four successive kinds
of access. He distinguishes four kinds of barriers to access
and the type of access they restrict:
1. Lack of elementary digital experience caused by lack
of interest, computer anxiety, and unattractiveness of
the new technology (“mental access”).
2. No possession of computers and network connections (“material access”).
315
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
316 J. VAN DIJK AND K. HACKER
3. Lack of digital skills caused by insufficient userfriendliness and inadequate education or social support (“skills access”).
4. Lack of significant usage opportunities (“usage
access”).
Clearly, public opinion and public policy are strongly
preoccupied with the second kind of access. Many people think that the problem of information inequality in the
use of digital technology or computer-mediated communication (CMC) is solved at the moment that everyone has
the ability to obtain a personal computer and a connection to the Internet. The first kind of access problem, the
mental barrier, is neglected or viewed as a temporary phenomenon touching only elderly people, some categories
of housewives, illiterates, and the unemployed. The problem of inadequate digital skills is reduced to the skills
of operation, managing hardware and software. Sometimes this is also viewed as a temporary phenomenon to
be solved shortly after the purchase of a computer and a
network connection. Differential usage of computers and
network connections is also neglected as an important phenomenon. Because differential usage is presumed to be
the free choice of citizens and consumers in a differentiating postmodern society, it has no been viewed as important to social and educational policies. Consequently, there
is a strong material or “hardware orientation” approaching access to digital technology. We can see this in the
most prevalent manner of framing the digital divide at this
time.
According to Van Dijk (1999), access problems of digital technology gradually shift from the first two kinds of
access to the last two kinds. When the problems of mental
and material access have been solved, wholly or partly,
the problems of structurally different skills and uses become more operative. Van Dijk (1999) does not limit the
definition of digital skills to the abilities of operating computers and network connections only. Instead, he includes
the abilities to search, select, process, and apply information from a superabundance of sources. In this way, he
anticipates the appearance of a usage gap between parts
of the population systematically using and benefiting from
advanced digital technology and the more difficult applications for work and education, and other parts only using
basic digital technologies for simple applications with a
relatively large part being entertainment. Van Dijk stresses
that computers and CMC are more multifunctional than
previous communication technologies.
This position is elaborated in this article as further conceptual distinctions and the latest data presumed to be reliable and valid are described. We show that the digital
divide is a very complex and dynamic phenomenon. As a
result, it is not easy to explicate the most important tendencies and policy directions for this fairly new social
problem. Yet we attempt to do this and also to suggest options for further research when knowledge and solutions
are lacking.
SOME FACTS: A DIGITAL DIVIDE IN
THE UNITED STATES AND EUROPE?
Unfortunately, most survey data about computer and Internet penetration or use are too unreliable and invalid to
draw definite conclusions about the existence and development of digital divides. In particular, Internet statistics
are notoriously unreliable for reasons of defective sampling or the nonresponse to and poor quality of much
(marketing) telephone interviewing. Research would be
improved with large surveys that have sufficient representativeness, or with census data and official government
statistics. Further, to make statements and to test hypotheses about trends in computer or Internet penetration, it is
necessary to have longitudinal or time-series data. These
are rather rare, but they are beginning to appear now. From
1994 until 1999 we could use the biannual GVU surveys
among Internet users (GVU Centre, Georgia University,
1994–1999). Time series could be constructed from their
data. However, a major problem with these surveys is that
they involve (self-)selective sampling.
Census material and other official statistics are starting
to appear in the United States and Europe. The trends of
the 1980s and 1990s, with 1998 and 2000 as the last years
of measurement, can be derived from them. We base our
conclusions on these data: the U.S. Census Bureau data
of 1984, 1989, 1993, 1997, and 2000, the partly overlapping NTIA data about telephone and computer penetration
for 1994, 1997, and 2000, and the annual Eurobarometer
(European Union) and Dutch official statistics (by the
SCP) of 1985, 1990, 1995, and 1998. The most important
results of these official statistics in relationship to the four
types of access distinguished are presented as sources of
empirical research conducted by others. We do not present
original data of our own.
Elementary Digital Experience
Few data are available—particularly in official statistics—
concerning the first experiences of potential users of digital technology. Mental barriers of access are neglected in
the discussion about the digital divide. It is known that
large segments of (even) the developed countries marked
by high technology still have very little digital experience.
Measuring the resulting digital skills (the ability to operate digital media and to search for information in them)
one finds that even in one of the nations in Europe most
equipped for digital communication, the Netherlands, 36%
of the 1998 population had no or very few digital skills.
Among people 65 years and older this figure reached 67%
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
THE DIGITAL DIVIDE AS A COMPLEX AND DYNAMIC PHENOMENON 317
TABLE 1
Percentages of the population of computer users in the
Netherlands in 1998, 18 years and older, with a particular
score of digital skills in using 10 applications on a
1–5 scale: Windows, word processing in DOS and
Windows, spreadsheets, drawing/graphics, working
with keyboards, Internet, programming, e-mail, and
statistical programs
No or very Reasonable
few skills, skills, Good skills,
score 1.0–2.0 score 2.1–3.5 score 3.6–5.0
Total DUTCH 36 52 12
population of
computer users
Age
18–34 years 27 56 17
35–49 years 37 54 9
50–64 years 48 46 6
65+ years 67 31 2
Gender
Male 28 55 17
Female 45 49 6
Education
Low education 69 25 6
Low middle 49 46 5
High middle 30 58 12
High education 27 55 18
Note. From representative GNC Survey reported in SCP (2000).
and among people with low education 69%. The average
of women with no skills was 45% and among people with
low middle education it was 49% (see Table 1).
In another Dutch survey for digital skills, subjective
and emotional factors appeared to be responsible for this
FIG. 1. Off-liners reasons for not buying a PC in Germany, 1999. From ARD/ZDF (1999a).
lack of skills to a large degree (Doets & Huisman, 1997).
These factors entail experiences of personal shortcoming
(leading to insecurity), of being excluded, and of negative attitudes toward this technology, all factors leading to so-called “computer anxiety.” Such mental access
problems become more important when it is claimed that
there are not only information have-nots, but also information want-nots. Thus, there also are important motivational problems. In general, it appears to be possible
to live and work without digital technology at the turn
of the century. In 1999 a couple of European surveys
were published revealing that about half of the population that was not connected to the Internet also did not
want such a connection. One of these surveys was the
German Online Nonusers Survey (ARD/ZDF, 1999a).
Among the 501 nonusers in this representative sample for
Germany, 234 (54%) declared that they would not connect to the Internet for a variety of reasons of volition (see
Figure 1).
The statements listed in Figure 1 reveal everyday reasons for not using the Internet, such as “I don’t need it,” “I
don’t like it, “I can’t buy it,” and “I can’t handle it.” Similar reasons were provided by households in the United
States having a computer or WebTV in the year 2000,
but never using it to access the Internet (NTIA, 2000):
“don’t want” (31%), “too expensive” (17%), “can use it
elsewhere” (10%), and “no time” (9%). Presumably, there
are differences in motivations for using computers and the
Internet among the populations of (even) high-tech countries. Older people, those with low education, a large proportion of women, and (functional) illiterates are strongly
overrepresented among people with a lack of motivation
(ARD/ZDF, 1999a; NTIA, 2000). Further research for
the ingredients of the mixture of reasons observed here
(anxiety, negative attitude, lack of motivation) is urgently
needed.
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
318 J. VAN DIJK AND K. HACKER
Possession of Computers and Network Connections
Current discussions about the digital divide are dominated
by the (lack of) universal availability of the hardware. Increasingly, longitudinal data in official statistics are being
supplied. These data constitute strong evidence of gaps
in the possession of computer and network connections
among a number of social categories during the 1980s
and 1990s: income, education, occupation, age, gender,
FIG. 2. (Left) Gaps of income, education, age, and ethnicity, in the United States, 1984–2000. Computed from U.S. Census
Bureau data of 1984, 1989, 1993, 1997, 1998, and 2000, and data contained in NTIA (1999, 2000). (Right) Gaps of income,
education, employment, and age, the Netherlands 1985–1998. From SCP (2000).
ethnicity, and geographic location. By constructing time
series from these data, it can be shown that most of these
gaps of possession have increased during the 1980s and
1990s. Given next is a collection of figures showing this
for the variables of income, education, and age in both
the United States (see U.S. Census Bureau, 1984, 1989,
1993, 1997, 2000; Kominski & Newburger, 1999) and
the Netherlands (see SCP, 2000). Ethnicity is added as
a category in the United States (Figure 2). Gender is not
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
THE DIGITAL DIVIDE AS A COMPLEX AND DYNAMIC PHENOMENON 319
included, because both sets of official statistics employed
households as the unit of data collection. Gender differences related to possession of equipment are not sufficiently articulated in this type of research. They did appear
significant in the biannual GVU surveys among individual web users (GVU Center, 1994–1999). However, it appears that the original gender gap in actually using PCs
and the Internet decreased during the 1990s. In the year
2000 the gender difference in the possession of computers
and the Internet and the time spent using them has been
equalized in the United States (see NTIA, 2000). In the
European Union (EU), women are still catching up (see
the annual Eurobarometer surveys of the EU). However,
it is important to emphasize that a gender gap concerning
different skills and kinds of usage remains (see Tables 1,
2, and 3; Pew Internet, 2000).
The big question concerning the widening gaps is
whether these trends will continue. From a statistical point
of view, it is evident that they will not. Saturation of computer and network possession among the “higher” categories will set in, and presumably has already begun in
countries like the United States and the Netherlands. For
the “lower” categories there is much more room for catching up. Therefore, the question becomes how much current
gaps will close in the first two decades of the 21st century.
More importantly, we need to ask what kind of computers
and network connections people will possess. We return
to this crucial issue later.
Other important questions deal with the most important
factors or variables among the familiar set of background
variables. Evidently, income, education, and employment
are strongly associated. Holding the other factors constant,
it can be shown in the American and Dutch statistics that
they keep an independent effect. The Dutch SCP study,
however, employed elaborate multiple-regression analyses
for the weight of the most important variables in the possession, skills, and use of information and communication
technology (ICT). The results, summarized in Figure 3,
are informative.
The most important conclusion from the 1998 Dutch
SCP study is that household income is the most important
factor explaining differences in the possession of ICTs,
first of all PCs, and that income diminishes or even
disappears in the explanation of differences in digital skills
and ICT usage. Surprisingly high, perhaps, is the relative weight of age and gender. For age, the distribution is
curved: First possession increases with age (with a top in
the class of 30–40), and then it decreases. Women have significantly lower possession—in the EU, not in the United
States—skills, and use of ICTs. We try later to explain
why the possession of ICT is not only a matter of material
resources but also of the attractiveness of this technology
and the necessary skills to use it among people of different
age and gender.
FIG. 3. A multiple regression comparing the relative significance of the variables of income, education, age, and gender
in the possession, skills, and use of ICTs in the Netherlands in
1998. Data from representative GNC Survey 1998, SCP (2000).
Digital Skills
PCs and computer networks were renowned for their userunfriendliness until well into the 1990s. Major improvements were made with the introduction of graphical and
audiovisual interfaces. However, the situation is still far
from satisfactory if we look once again at Table 1 presenting differences of digital skills among social categories in
the Netherlands. Gaps of digital skills can be shown to
exist. The most common definition of digital skills is instrumental skills: the ability to operate hardware and software. In the study concerned, digital skills were operationalized using an index called “informacy” measuring
both skills of operating digital equipment and skills of
searching information using digital hardware and software. This means that so-called informational skills are
added to the definition. We later recommend adding a third
type of digital skills to the operational and informational
ones: the strategic skills of using information for one’s
own purpose and position. Figure 3 reveals the (perhaps)
surprising result that digital skills (instrumental and informational, together called “informacy” here) are not primarily related to educational levels but to age and gender.
It is likely that this means that real practice and motivation
are more important in acquiring digital skills than formal
education. Indeed, many studies reveal that having computer experience at work, having particular hobbies, and
having a family with schoolchildren are decisive factors
in the acquisition of digital skills by adults.
Different Uses
In this article and in other publications (Van Dijk, 1997,
1999, 2000), it is predicted that different uses of ICT will
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
320 J. VAN DIJK AND K. HACKER
TABLE 2
Usage of PC at home, United States, 1997
Word Work
processing Games E-mail Internet Bookkeeping at home Spreadsheets Databases
Total using PC 70.5 53.6 44.5 44.2 43.6 34.3 28.7 26.1
at home
Age
18–24 years 69.7 61.4 42.8 44.3 19.5 14.0 18.7 17.1
25–49 years 70.3 55.0 45.8 45.9 46.9 38.5 30.9 27.3
50+ years 71.4 44.7 41.4 39.3 48.6 34.4 28.3 28.1
Gender
Male 66.3 57.6 48.1 49.4 46.3 38.0 32.5 29.4
Female 74.7 49.6 40.7 38.9 40.8 30.6 24.7 22.7
Family income
<$25,000 69.1 57.2 40.6 38.9 35.5 22.7 22.3 20.9
$25–49,900 66.2 58.4 39.7 39.8 43.7 29.2 25.3 23.9
$50–74,900 71.2 55.4 44.9 45.1 43.9 35.5 29.7 26.7
>$75,000 75.8 47.6 52.1 52.3 47.3 44.8 34.7 31.1
Note. From U.S. Census Bureau data.
bring the most important digital and information inequalities in society. Presently, the differences observed in this
kind of access are not as large as those in differential possession and skill.
Unfortunately, data about differential usage are still
scarce and only a few years old. They are available for comTABLE 3
General usage of PCs in the Netherlands, 1998
CD-phone and Spreadsheet/ Graphics/
Word processing Games travel guide database drawing Internet E-mail
Total using PC at home 86 59 46 45 37 37 34
Age
18–34 years 89 74 45 46 41 38 36
35–49 years 84 54 47 44 35 39 35
50–64 years 83 36 45 42 35 35 31
65+ years 79 31 41 46 16 35 20
Gender
Male 89 58 55 60 45 44 42
Female 81 60 32 36 26 27 25
Family income
1st Quartile 90 74 42 43 42 33 30
2nd Quartile 82 62 47 40 36 31 28
3rd Quartile 85 57 45 39 36 37 35
4th Quartile 87 46 52 59 34 49 46
Education
Low education 73 78 47 16 32 16 12
Low middle 72 66 42 32 38 32 29
High middle 88 64 43 45 36 35 32
Note. From SCP (2000).
puter use and Internet use, both in the U.S. Census material
and the Dutch SCP investigation. However, it is our view
that only computer use has had enough time to crystallize; Internet use is only appearing as a mass phenomenon
at the turn of the century. We must wait for longitudinal
data to construct the time series we need for testing our
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
THE DIGITAL DIVIDE AS A COMPLEX AND DYNAMIC PHENOMENON 321
prediction: the rise of a usage gap. Instead, we have to turn
to data of single years like the state of affairs concerning
computer use in the Netherlands and the United States, in
1998 and 1997, respectively, presented in Tables 2 and 3.
Here we can see substantial differences in the use of PC
applications, especially among people with different levels
of age, gender, and education. With age, fairly large differences appear in using games, spreadsheets/databases/
bookkeeping (United States), and drawing and e-mail use
in the Netherlands. With gender we see that females use
all applications significantly less than males. Levels of
education appear to correlate with a different use of
games, spreadsheets/databases, Internet, and e-mail (the
Netherlands). In the United States only the data on income levels are available, revealing differences in e-mail
and Internet use and other more advanced applications:
bookkeeping, spreadsheets, databases, and work at home.
The latest Falling Through the Net study (NTIA, 2000)
reveals important differences in Internet usage by income,
education, ethnicity, and other variables, but unfortunately
only informational, educational, and work-related types
of Internet use were reported. It appears that with rising
educational levels the Internet applications of information searching, doing job-related tasks, searching for jobs,
and using e-mail increase significantly (see NTIA, 2000,
Figure A49). On the other hand, people with lower education use the Internet relatively more to take courses. When
we examine income these differences are less pronounced,
supporting the SCP conclusion that income is less important than education considering usage (Figure 3). Taking
courses and searching for jobs on the Internet is practiced
more by Americans with low incomes than with high incomes in the year 2000. The same goes for unemployed
Americans as compared to the employed (see NTIA, 2000,
Figures A47 and A50). This reveals the importance of usage access and skills access compared to their necessary
condition, material access. Having a computer and Internet connection and also having the skills to use them are
becoming increasingly important resources on the labor
market.
INTERPRETATIONS
As observed earlier, the data concerning the digital divide
have come to constitute a political battlefield. Statistics are
freely selected and judged according to political interests
or subjective perspectives. Social and political opinion has
developed four kinds of positions with an interpretation of
current state of affairs:
1. Denial of the existence of a digital divide.
2. Acceptance of some present divides, claiming that
they will soon disappear.
3. Emphasis of digital divides that are supposed to
grow and come on top of old inequalities of income,
education, age, gender, ethnicity, and geographical
location.
4. Differentiation: some gaps are decreasing while others grow.
A number of market research institutions, other corporate
interests, and conservative think tanks deny or trivialize
the existence of digital divides. Basically, their arguments
are threefold (see also U.S. Internet Council, 1999, based
on Forrester research, and Thierer, 1999):
1. The adoption rate of computers and the Internet and
the growth rates of their use are higher than any
medium before, perhaps with the exception of (color)
TV.
2. The distribution among the population approaches
normality: The averages of income, education, ethnicity, and gender rapidly parallel society as a
whole.
3. Computers and Internet connections are becoming
cheaper by the day, cheaper than a color TV system
adopted by almost every Western household. The
market is doing its work and solves all problems.
Indeed, growth rates are enormous. However, there are
some basic problems with the S-curve of adoption of innovations that usually is the basis of this argument.1 One
of them is the demarcation problem of the media supposed to be entering an S-curve: A computer and an Internet connection now are very different from a computer
and Internet connection 10 years ago. The new computer
and Internet technologies are easier to use, but they have
varying levels of complexity and options in hardware and
software.
It goes without saying that a medium that is increasingly adopted into society is approaching average parts of
the population. However, in our view digital divides are
about relative differences between categories of people.
In the 1980s and 1990s most of these divides concerning
possession of computers and Internet connections have
increased, as has been convincingly demonstrated by the
American and Dutch official statistics supplied earlier. One
is free to predict that these divides will close rapidly, an
argument to be dealt with later, but their existence in the
present and recent past cannot be denied. The argument
about cheaper hardware is correct, but only partly so. It
neglects many facts like:
1. The new media add to the older mass media that do
not disappear: One still needs a TV, radio, VCR,
telephone, and perhaps a newspaper; low-income
households continually have to weigh every new purchase (with the newspaper beginning to lose).
2. Computers are outdated much faster than any other
medium and new peripheral equipment and software
continually have to be purchased.
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
322 J. VAN DIJK AND K. HACKER
3. “Free” Internet access or computer hardware is not
really free, of course. There are nominal monthly
fees, long-term service agreements, privacy selling,
and low-quality service, for instance.
However, the most important problem of this interpretation, and of the next one, is its hardware orientation.
Perhaps the most common social and political opinion is
that the problem of the digital divide is solved as soon as
every citizen or inhabitant has the ability to obtain a personal computer and an Internet connection. In contrast, our
analysis suggests that the biggest problems of information
and communication inequality just start with the general
diffusion of computers and network connections.
The second interpretation might accept that there are or
have been gaps but asserts that they will soon disappear,
perhaps to be succeeded by other inequalities. It is simply
a question of some having the technology now and others
having it later. The early adopters pay for the innovation
and make the adoption cheaper for the late adopters. There
is a strong faith in the trend of the S-curve of adoption and
in the extension of access by market forces alone. This
interpretation comes from the authors of the Dutch SCP
survey, among others. From a statistical point of view their
position will be backed by future data automatically. The
saturation of possession by the “higher” social categories
sets in, and has already started in some Western countries,
as one can see in the slope of the top curves in Figure 2.
This argument looks like a dynamic one because it accounts for trends into the future, but in fact it is static on
several grounds. It reasons from present technologies and
their uses. The issues are what relative differences will
remain in 10–20 years and what kind of “computers” and
“Internet” will be possessed. How will they be used? What
skills will be needed? One has to remember that the comparable innovation of telephony took 70 years to acquire
an (almost) general distribution and acknowledge that penetration rates are still not complete and that usage is very
unequal, even in the rich developed countries. In the meantime, the next round of innovation in digital technology
has already appeared: broadband or high-speed Internet
access. It is striking to observe that the same divides in the
possession of this technology among people with different income and education reappear, albeit in a mitigated
form (see NTIA, 2000, Figures A18 and A19). This means
that the differences between people with broadband and
narrowband access are smaller than those between people
with and without any access to computers and the Internet.
Still, it is reasonable to say that all these differences work
in accumulation and that the relative differences in hardware possession between people with different income and
education, at least, are increasing.
Another argument of the disappearing gaps position is
that there is no digital divide in the sense of a structural
gap or a two-tiered society: The differences are of a gradual nature (see SCP, 2000). This qualification has some
basis: The two-tiered position is too simple. The digital divide is a simplifying metaphor. In fact, we see the
stretching of a whole spectrum of differentiating positions
in (post)modern society, not two classes of people (Van
Dijk, 1997, 2000). From a substantial point of view this
qualification might also be right concerning the basic possession of computers and Internet connections, though the
SCP survey’s own conclusion is that household income
is the most important factor explaining it. Yet we are not
so sure that structural divides will disappear concerning
digital skills and usage, the core of our argument.
The disappearing divide position often is politically motivated by the wish to prevent government interference.
One supposes markets will solve most problems by themselves lowering prices and offering more choice to everybody and that people in their communities and organizations will solve the rest of problems in self-regulation.
Rather than engaging in this ideological debate, we wish
to draw attention to the straightforward fact that almost
every government in the 20th century has adopted policies to promote important mass media for communication
in society, including tax policies and hardware and support subsidies for all kinds of public services. We want to
call attention to the fact that every government, including
those committed to laissez-faire capitalism as in the United
States, has implemented educational and cultural policies.
A third set of interpretations does emphasize the persistence and growth of a digital divide. It is supported
by left-wing political forces, social democrats, socialists,
progressive nongovernmental organizations (NGOs), etc.
They stress the rise of social and economic inequality in
Western society and on a world scale in general during the
last two decades. They claim information inequality only
adds another layer to increasing old inequalities of income,
education, occupation or social class, ethnicity, and gender
(see Schiller, 1996). They hold that the claim of cheaper
information and communication technology (ICT) products is a corporate trick. After the relatively cheap supply
of hardware access, the selling of expensive service and
content starts. There may be large elements of truth in
this interpretation: General inequality has increased both
nationally and internationally (see United Nations Development Program, 1998), and old inequalities do not disappear with the advent of an information or knowledge
society. However, this position underestimates the import
and complexity of changes that are occurring. Increasing
differences in the skill and usage of the new information
technologies might lead to new inequalities of a nature not
known before and to be combatted, if one chooses to do
so, with other means than the traditional ones. Moreover,
less expensive hardware with more capacity and free Internet access as a public service are very real and important
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
THE DIGITAL DIVIDE AS A COMPLEX AND DYNAMIC PHENOMENON 323
phenomena. The new technologies offer new opportunities
for citizen participation and the consumer interest.
The last set of interpretations stresses current differentiation in society in general and the use of ICTs in particular. Theoretical work on the information society or the
network society sheds another light on social inequality
(see Castells, 1996, 1998; Van Dijk, 1991, 1999, 2000).
Others even claim that the information or knowledge society will discard old inequalities and bring completely new
ones based on differential knowledge and education. Our
analysis indicates that there also is continuity, as elaborated in the next section. Concerning the different divides
discussed here, these interpretations stress that some current divides or gaps may (partly) disappear, while others
stay or increase. This interpretation matches our conclusion that the current digital divide is a very complex and
dynamic phenomenon. It is complex in the sense that, for
example, access is a multifaceted concept with many types
of problems, and it is dynamic in following the trends of
evolving technology and its uses.
SCIENTIFIC EXPLANATIONS
The differences of possession, skill, and use of ICTs usually lack scientific explanation. Even multivariate analyses
trying to weigh determining factors are rare. An exception
is made by the Dutch SCP survey constructing regression
models to explain these differences by unequal possession
of resources by individuals or households. This is a classical sociological approach in empirical research. Three
kinds of resources were distinguished: material, social,
and cognitive resources. In this survey a very narrow definition of material resources appeared to explain more (in
the regression models) than the usual variable in this respect: income. The variable constructed is narrow because
it is only composed of questions about the possession of
all kinds of equipment by households, albeit not only digital equipment. Social resources are made operational in
a number of questions about (1) having a social network
also possessing and using digital technology and (2) having social support in managing it. Cognitive resources are
threefold: literacy, numeracy, and informacy. Literacy is
the skill of reading and of searching information in texts.
Numeracy is the ability to handle numbers, figures, and tables and to compute. Informacy is equal to digital skills in
this survey. It is made operational in two ways: operating
digital equipment, and searching for information in digital
sources. The cognitive resources taken together appear to
explain more than the variable of education.
The results of the regression model based on these resources are very interesting. Striking differences of the
importance of these resources are found at the possession,
skills, and use of ICTs, respectively (Figure 4). Possession
of ICTs is explained more by informacy (instrumental and
informational digital skills) than by material resources.
FIG. 4. A multiple regression comparing the relative significance of material, cognitive, and social resources in the possession, skills, and use of ICTs in the Netherlands in 1998. Data
from representative GNC Survey 1998, SCP (2000).
Skills are explained by literacy and social resources (having a social network and support). A remarkable result is
that literacy is far more important for the explanation of
digital skills than numeracy. Apparently, people with the
ability to process textual information are more likely to
develop digital skills defined in this way than people that
are good in numbers and computing. Clearly, computers
are no longer only number crunching machines these days.
Usage is overwhelmingly determined by informacy or digital skills.
The general conclusion of the SCP research team is that
differences of skill and use are smaller than differences of
possession. After the threshold of having a computer and
network connection has been passed, material and social
resources play a relatively minor role. Social–cultural differences of age, gender, literacy, and informacy come forward. Present differences and even divides are observed—
see figures and tables already given—but according to the
SCP team they are old inequalities that are reproduced,
rather than new inequalities like the ones often related to
a knowledge or information society. The team claims that
there is no unbridgeable digital divide and that government
intervention is not needed.
It is likely that these far-reaching political conclusions
are drawn prematurely. They are based on a rather static
and superficial sociological analysis of the present situation. Constructing rather arbitrary background variables of
individualresources at a single point in time does not make
a theory that is able to relate to social and technological
development, that is to say, the level ofsociety and technology. Technology is changing rapidly; very advanced and
very simple applications are appearing side by side. And
according to many scientists and other observers, society is
evolving into an information society and a network society
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
324 J. VAN DIJK AND K. HACKER
where social (in)equality will partly be different from old
modern societies.
In an information society, information is known as a socalled primary good. Everybody needs it to function in society. However, people also need cultural capital (Bourdieu,
1986) and social capital (Mansell & Wehn, 1998) to use information in appropriate ways—that is, the skills to select
and process information and be able to use it in one’s social position and network. These kinds of “capital” are distributed very unequally in society. Moreover, information
is a positional good. This means that it becomes increasingly important to get the information first in economic,
social, and cultural competition. This is why it is so important to look at the relative differences in all inequalities
observed.
The importance of cultural and social capital for the
ability to extract relevant information from innumerable
sources and to use it for the benefit of one’s position is
even stronger in the network society, a typification in the
line of the information society. A network society consists
of social and media networks shaping the prime mode of
organization and most important structures of modern society (Van Dijk, 1999). Here the position inside and outside
networks becomes vital. This position defines one’s opportunities and power in society. Remaining outside networks
means total exclusion. Being inside might mean partial exclusion when the position occupied is a marginal one. The
position acquired at work, at school, at home, and in the
local community also determines the chances to acquire elementary digital experience, to develop further digital experience, and to use particular applications. Therefore we
want to add a third type of digital skills to the instrumental
and information skills already described: strategic skills,
the ability to use digital means too improve one’s position
in society, in work, education, and cultural practices. This
type of digital skill is closely related to the possession of
social capital or social resources as defined earlier and it is
the direct basis of usage access. In measuring social capital and resources, it is insufficient to observe only whether
one is employed or not and how big ones household is.
The precise positions at work (occupation, function, task),
at school, in the family, and in the community have to be
recorded and related to the possession, skills, and use of
ICTs. Unfortunately, these data are scarce. We have seen
that measuring skills and usage is a fairly recent research
activity, particularly in relationship to the Internet. Thus,
we are not able to look for such clear “gap pictures” as are
available on the field of the possession of computers and
network connections. The evidence is only fragmentary,
like the tables of usage we supplied earlier, or it is too recent to construct time series, like the 1993 and 1997 U.S.
Census Bureau statistics of the precise occupations and
industries using particular applications of ICT (see U.S.
Census Bureau, 1993, 1997).
Further research for the different kinds of digital skills
and for usage is urgently required. This will allow us to
investigate whether more or less structural inequalities in
skills and usage appear between social classes and people of different age, gender, ethnicity, and geographical
location. This is the hypothesis of simple versus advanced,
businesslike versus entertainment applications adopted
relatively more by particular classes of people, a suggestion made earlier.
Individual differences of ICT possession, skills, and use
observed should not only be related to the general environment of the information or network society, but also
to the particular social trends of a particular epoch. Van
Dijk (1999; pp. 153–154) has argued that in the present
epoch several trends come together to promote information inequality: social and cultural differentiation or individualization, rising income differentials, privatization
and cutbacks in social and public services, and, finally,
increasingly multifunctional and differently used digital
technology. The last trend supports Van Dijk’s vision of
the new media as trend amplifiers: Equalities and inequalities already present, growing or declining in society will
be reinforced by this technology.
CONCLUSIONS AND POLICY PERSPECTIVES
Following the line of the argument in this article, the
complexity of the picture of the so-called digital divide
comes to our mind. A number of significant gaps have
been observed and supported by relatively reliable official statistics and surveys. However, there is no question
of an absolute, yawning and unbridgeable gap between
two classes of people. Talk about “technological segregation” (NAACP President Kweisi Mfume) and “classical
apartheid” (Reverend Jesse Jackson) is exaggerated and
misses the point. The point is that the gaps observed show
first of all relative and gradual differences. This makes
them no less important. In the information and network society, relative differences in getting information and lines
of communication become decisive for one’s position in
society, more than in every society in history before. Giving everybody a computer and a network connection, banning the cutting lines of “segregation” in this way, will
not remove them. Much deeper and more clear-cut differences in skill and usage will appear as both technology
and society increasingly differentiate. The fundamental
task of future society will be to prevent structural inequalities in the skill and usage of ICTs from becoming more
intense. Inequalities become structural when they “solidify,” that is, when positions people occupy in society, in
social networks, and in media networks, or other media,
become lasting and determine to a large degree whether
they have any influence on decisions made in several fields
of society.
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
THE DIGITAL DIVIDE AS A COMPLEX AND DYNAMIC PHENOMENON 325
Another reason for the complexity of the digital divide
is that there are in fact several divides. Some are widening
while others are closing. Time series of official statistics
have demonstrated that during the 1980s and 1990s gaps
of income, education, age, and ethnicity in the possession
of computers and hardware have grown, at least in the
United States and the Netherlands. Clearly, the people at
the “better side” of these gaps have increased their lead
during these decades. Though these gaps of possession
will (at least partially) close in the next decades, if only
for the statistical reason of saturation effects, it is very
unlikely that those having acquired a big advantage will
stop and lean backward. Technology is advancing, splitting
in simple and highly evolved applications, spreading into
society and sticking to old and new social differences.
In the course of the 1990s the gender gap in the possession of ICTs has started to close. However, gender gaps in
skill and usage remain or mature, though they are much
smaller for girls and boys than for adults (see GVU, 1994–
1999; ARD/ZDF, 1999b; SCP, 2000; Pew Internet, 2000).
Large differences in digital skill and usage have been
observed recently. Here gaps might grow in the future,
although this cannot be proven at this moment due to a
lack of time-series data.
The conclusions just described have also highlighted
the dynamic nature of every digital divide. One should
not stop at a particular point in time and say that a particular technology or application will be available to everybody within a couple of years or at any projected date
in the future. Information and communication technology
will differentiate considerably in the first decades of the
21st century. Computers will be available in the simplest
(palm-top and other) forms and very advanced types of
desktops, laptops, and servers. “The Internet” will be accessible via televisions, mobile phones, and other small information appliances next to fast broadband connections.
An important policy question will be whether palm-top
computers and mobile phones or all kinds of narrowband
access will be sufficient to be called the basic connection
every citizen needs. Moreover, what does basic access to
the Internet mean: both at home and at work/school, or is
one of them sufficient, or perhaps even a connection in a
public utility?
An important characteristic of ICT in this respect is
its extended multifunctionality. Printed media, radio, television, and telephone have all been used differently by
people with high and low education in particular. However, their (difference in) functionality is small compared
to computers and the Internet. In the meantime, society
is also differentiating at an unprecedented scale. Together
those may create a usage gap that is somewhat similar to
the knowledge gap described by Tichenor et al. a long time
ago. “As the diffusion of mass media information into a
social system increases, segments of the population with
a higher socio-economic status tend to acquire this information at a faster rate than the lower status segments”
(Tichenor et al., 1970, p. 159).
Although the evidence in favor of the thesis of knowledge gap has not been conclusive (Gaziano, 1987), it might
get another chance in the information or network society
where information is a positional good. We propose to relate this gap to a usage gap, not primarily based on differential derived knowledge or information but on differential
practical use and positions in society.
The policy perspectives to be linked to this analysis
clearly depend on one’s central objectives concerning information inequality and one’s political position. Central
objectives might be twofold. The most basic one is social
inclusion. A step further is made in the objective of an
equal distribution of resources or life chances. The first
objective is backed by a coalition of forces in advanced
high-tech societies. Corporations look for a large electronic marketplace. Politicians want extended reach for
political persuasion and a grip on new channels of political communication bypassing traditional mass media.
Military people and security agencies want everybody to
be connected for purposes of control and surveillance, as
the off-liners of the future will create unknown risks. Educators are concerned about universal and public access
to all learning resources. Community builders want every
citizen to be involved in online communications linked to
offline local activities.
The second objective is more traditional and it is supported more in Europe than in the United States, for instance. The minimum is an equal distribution of chances
to every individual, an objective also having a broad support. Filling in what this means for actual material, social,
and cognitive resources reveals the differences of political
position.
Policy perspectives should be linked to the four kinds
of access that have been distinguished. Governments, civil
societies, and markets all have important roles to play in
the support of these kinds of access.
Elementary digital experience is first of all a question
of the market developing and offering ICTs that really are
user-friendly and that offer such a clear surplus value as
compared to old applications that the “information wantnots” will be convinced. Even at that point, many elderly
and less educated people and some categories of housewives will stay behind. This will be the most important
mission of adult education to be offered by governments,
community centers, and corporate training.
Concerning the general possession of computers and
networks, markets have done a good job lowering prices
for technologies with higher capacities. However, this has
not prevented the growth of digital divides in the possession of hardware, at least until very recently. Household
income is still the most important factor here. Thus, tax
Downloaded by [Vancouver Island University] at 10:27 14 January 2014
326 J. VAN DIJK AND K. HACKER
and income policies of governments certainly do make
sense. However, general tax credits or subsidies are not
effective. They have to be focused on the groups clearly
staying behind, all of them in the lowest quarter of the
income distribution. A second qualification is the need
of public or private service and guidance. Simply offering cheap boxes with computers and Internet connections
makes little sense in this situation.
Learning digital skills will be a strategic objective for
educational institutions at all levels. The official American
and Dutch surveys cited in this article indicate that present
digital skills are learned more at work than at schools or at
home. In general, formal education runs behind because
means are lacking and teachers are not sufficiently trained
or motivated. Filling in this strategic objective, it will become evident that digital skills not only mean abilities to
operate the hardware and software (instrumental skills).
Increasingly, they will mean the ability to search, select,
process, and apply information (informational skills) from
digital sources and to strategically use them to improve
one’s position in society (strategic skills). At least instrumental and informational skills have to be learned at
schools.
Improving usage opportunities for all means making
them more attractive to some people in the first place. We
have observed the surprisingly high independent effects
of age, gender, and ethnicity (in the United States) for the
actual use of ICTs. Applications should be made more attractive to many old people, women, and ethnic minorities.
This is a matter of design, culture, language, and identity included and addressed in the applications concerned.
Producers, designers, and representatives of citizens and
consumers have a job here.
NOTE
1. The S-curve of the adoption of innovations presupposes that the
medium in question is easy to identify and to mark from others. This
might be true for older mass media like a radio, a TV, or a VCR, but not
for computers and network connections. They fall apart in extremely
different types, strongly complicating the construction of any valid time
series of adoption. The second questionable proposition is the maximum
population of potential adoption. The classical S-curve presupposes
whole populations. However, some new computer and network media
are too advanced, complicated, and expensive to be ever adopted by
100% of the population.
REFERENCES
ARD/ZDF-Arbeitsgruppe Multimedia. 1999a. ARD/ZDF Online
Studie 1999. Media Perspektiven 8/99:388–409.
ARD/ZDF-Arbeitsgruppe Multimedia. 1999b. Internet—(K)eine
Mannerdom ¨ ane. ¨ Media Perspektiven 8/99:423–429.
Bourdieu, P. 1986. The forms of capital. In Handbook of theory and
research for the sociology of education, ed. John R. Richardson,
pp. 241–258. New York: Greenwood Press.
Castells, Manuel. 1996. The information age: Economy, society and
culture, Vol. I: The rise of the network society. Cambridge, MA:
Blackwell.
Castells, Manuel. 1998. The information age: Economy, society and
culture, Vol. III: End of millennium. Malden, MA: Blackwell.
Doets, Cees, and Huisman, T. 1997. Digital skills. The state of the art in
the Netherlands. ‘s-Hertogenbosch: CINOP. http://www.cinop.nl
(English summary).
Gaziano, C. 1987. The knowledge gap: An analytical review of media
effects. Communication Research 10:447–486.
GVU Centre, Georgia University. 1994–1999. GVU’s 1st–10th WWW
User Surveys. http://www.gvu.gatech.edu/user surveys/
Kominski, R., and Newburger, E. 1999. Access denied: Changes in
computer ownership and use: 1984–1997. Paper presented at the
Annual Meeting of the American Sociological Association, August,
Chicago.
Mansell, Robin, and Wehn, U. eds. 1998. Knowledge societies, Information technology for sustainable development. Oxford: Oxford
University Press.
NTIA (U.S. Department of Commerce). 1999. Falling through the
Net III: Defining the digital divide. http://www.ntia.doc.gov/
ntiahome/fttn99/part1.html
NTIA (U.S. Department of Commerce). 2000. Falling through
the Net IV: Towards digital inclusion. http://www.ntia.doc.gov/
ntiahome/fttn00/contents00.html
Pew Internet & American Life Project. 2000. Tracking online Life:
How women use the Internet to cultivate relationships with family
and friends. Washington, DC: PEW. www.pewinternet.org
SCP (Sociaal-Cultureel Planbureau/Social-Cultural Planning Agency),
Van Dijk, Liset, J. de Haan, and Rijken, S. 2000. Digitalisering van
de Leefwereld, een onderzoek naar informatie en communicatietechnologie en sociale ongelijkheid. Contains a summary in English.
Rijswijk, the Netherlands: SCP.
Schiller, Herbert. 1996. Information inequality. The deepening social
crisis in America. New York: Routledge.
Thierer, A. 2000. How free computers are filling the digital divide.
Heritage Foundation Backgrounder 1361, 20 April. http://www.
heritage.org
Tichenor, P. J., and Donohue, G., and Olien, C. 1970. Mass media
flow and differential growth in knowledge. Public Opinion Quarterly
34:159–170.
United Nations Development Program. (1998). Human development
report 1998. New York: Oxford University Press.
U.S. Census Bureau. 1984, 1989, 1993, 1997, 1998, 2000. Current
population surveys. Washington, DC: U.S. Census Bureau.
U.S. Internet Council. 1999. State of the Internet: USIC’s report on use
& threats in 1999. htpp://www.usic.org/usic state of net99.htm
Van Dijk, J. A. G. M. 1991. De Netwerkmaatschappij, Sociale aspecten van nieuwe media. Houten, the Netherlands: Bohn Stafleu
van Loghum.
Van Dijk, Jan A. G. M. 1997. Universal service from the perspective
of consumers and citizens, Report to the Information Society Forum.
Brussels: European Commission/ISPO.
Van Dijk, Jan. 1999. The network society, Social aspects of new media.
Thousand Oaks, CA: Sage.
Van Dijk, Jan. 2000. Widening information gaps and policies of prevention. In Digital democracy, Issues of theory and practice, eds.
K. Hacker and J. Van Dijk, pp. 166–183. Thousand Oaks, CA:
Sage.
Downloaded by [Vancouver Island University] at 10:27 14 January 2014