Content uploaded by Alexander J.A.M. Van Deursen
Author content
All content in this area was uploaded by Alexander J.A.M. Van Deursen on Mar 16, 2020
Content may be subject to copyright.
Article
Corresponding author:
Alexander J.A.M. van Deursen, University of Twente Faculty of Behavioral Sciences, Department of Media,
Communication and Organization, Cubicus Building, PO Box 217, 7500 AE Enschede, The Netherlands
Email: a.j.a.m.vandeursen@utwente.nl
Internet skills and the
digital divide
Alexander van Deursen and Jan van Dijk
University of Twente, The Netherlands
Abstract
Because of the growing amount of information on the internet and people’s increasing
dependence on information, internet skills should be considered as a vital resource in
contemporary society. This article focuses on the differential possession of internet
skills among the Dutch population. In two studies, an in-depth range of internet skills are
measured by charging subjects assignments to be accomplished on the internet. Subjects
were recruited by applying a random stratified sampling method over gender, age, and
education. While the level of operational and formal internet skills appeared quite high,
the level of information and strategic internet skills is questionable. Whereas education
appeared an important contributor to all skill levels, age only appeared a significant
contributor to operational and formal skills. The results strengthen the findings that the
original digital divide of physical internet access has evolved into a divide that includes
differences in skills to use the internet.
Keywords
digital divide, inequality, information, internet, internet skills, literacy, online
Introduction
The term ‘digital divide’ initially referred to gaps in access to a computer. When the
internet diffused rapidly into society and became a primary type of computing, the term
shifted to encompass gaps in not only computer but also internet access. Early research
on the digital divide focused mainly on a binary classification of physical access.
Theories of internet adoption have recognized this limitation and an increasing number
of researchers have argued that more attention should be paid to social, psychological,
and cultural backgrounds (Van Dijk, 2006). This has resulted in several conceptualizations
new media & society
13(6) 893–911
© The Author(s) 2010
Reprints and permission:
sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/1461444810386774
nms.sagepub.com
894 new media & society 13(6)
of how to approach digital divide research (e.g., DiMaggio and Hargittai, 2001;
Mossberger et al., 2003; Van Dijk, 2005; Warschauer, 2003). These conceptualizations
reveal that while gaps in physical access are being addressed, other gaps seem to widen.
One of the factors that appears to be important is the differential possession of digital
skills. Changes in society demand new skills, especially those related to the internet as
one of the most important means of communication in contemporary society. Because
of the growing amount of information on the internet and people’s increasing depend-
ence on information, internet skills should now be considered as vital assets. When these
skills are unequally divided among the population, the consequences of this skills ine-
quality may even exacerbate existing societal inequalities (Van Dijk, 2005; Witte and
Mannon, 2007).
Very few measurements and scientific investigations of the actual level of internet
skills possessed by populations at large have been conducted. There are several empirical
studies that address a specific aspect, e.g., navigation (Ford and Chen, 2000), orientation
(Ahuja and Webster, 2001), selecting search results (Aula and Nordhausen, 2006; Pan
et al., 2007), defining search queries (Spink et al., 2001), or evaluating information
(Moharan-Martin, 2004). Unfortunately, the measurements do not tend to be admin-
istered on representative samples. From a sociological point of view, there are only
few studies that address internet skills. These studies typically use survey questions ask-
ing respondents for an estimation of their own level of internet skills. This method has
significant validity problems (Hargittai, 2005; Merritt et al., 2005; Talja, 2005). A gen-
eral impression that can be drawn is that the divides of skills tend to become bigger than
the divides of physical access and that, while physical access gaps are more or less clos-
ing in the developed countries, the skills gap tends to grow (Van Dijk, 2005).
Besides using less valid measurement methods, the few conducted internet skills stud-
ies among populations at large also fail to explain what the measured skills exactly com-
prehend. In most cases, only the command of hardware and software is considered. An
explanation might be the overabundance of internet skills-related concepts, while opera-
tional definitions are almost non-existent. A deeper understanding is required in order to
escape the simplification of early digital divide research in which only binary classifica-
tions of physical access were considered. Now a new simplification might appear: the
simple duality of the skilled and the unskilled.
The main contribution of this article is the measurement of an in-depth range of internet
skills actually commanded by the Dutch population. Since the best way to obtain a valid
measure of internet skills is directly testing that skill, two large-scale performance tests are
conducted. The following section contains the research questions and hypotheses.
Literature review
Internet skills
For measuring internet skills among populations at large, studies that use a range of inter-
net skills with a sequential and conditional nature are very interesting. Steyaert (2002)
and Van Dijk (2005) introduced three general types of digital skills that are also applica-
ble to the internet. Steyaert (2002) distinguished between instrumental skills (the opera-
tional manipulation of technology), structural skills (related to the structure in which
van Deursen and van Dijk 895
information is contained), and strategic skills (proactively looking for information, infor-
mation-based decision-making, and scanning for relevant information). Van Dijk (2005)
changed Steyaert’s definition to operational skills (the skills to operate computer and
network hardware and software), formal information skills (the ability to understand and
to handle the formal characteristics of a computer and a network such as file structures
and hyperlinks), substantial information skills (the ability to find, select, process and
evaluate information in specific sources of computers and networks), and strategic skills
(the capacities to use information as the means for specific goals and for the general goal
of improving one’s position in society). These definitions enable in-depth measurements
of internet skills and provide an opportunity to investigate how the different skills levels
are distributed among social segments in the population. Furthermore, they go beyond the
more traditional definitions of media literacy by suggesting a more (inter)active use.
While traditional media enable active mental processing, digital media require users to
interact with interfaces. A minimum level of active engagement with the medium is
required, and the possibility of interactions, transactions, and interpersonal communica-
tion is offered. Using the internet constitutes action, interaction, and transaction.
To encourage research to focus on in-depth skill measurement and to support the
achievements of digital divide research, Van Deursen and Van Dijk (2009a, 2010) elab-
orated the range of internet skills by proposing:
• Operational internet skills. These are derived from concepts that indicate a set of
basic skills in using internet technology.
• Formal internet skills. These relate to the hypermedia structure of the internet
which requires the skills of navigation and orientation.
• Information internet skills. These are derived from studies that adopt a staged
approach in explaining the actions via which users try to fulfill their information
needs.
• Strategic internet skills. These are the capacity to use the internet as a means of
reaching particular goals and for the general goal of improving one’s position in
society. The emphasis lies on the procedure through which decision-makers can
reach an optimal solution as efficiently as possible.
This division of four internet skills provides opportunities to investigate how these
different internet skills levels are distributed among segments in the population. The
definition avoids a technologically deterministic viewpoint by both accounting for
aspects related to the use of the internet as a medium (operational and formal) and sub-
stantive aspects related to the content provided by the internet (information and strategic).
The four types of internet skills have a sequential and conditional nature. Content-
related skills somehow depend on the medium-related skills because the absence of
medium-related skills means that one will not even come to perform the content-related
skills. Altogether, these skills are considered necessary for the general population to
function well in an online environment. The conceptual definition of internet skills is
listed in Table 1. The first research questions is:
RQ 1: What are the levels of operational, formal, information, and strategic internet skills of
Dutch citizens?
896 new media & society 13(6)
Internet skills determinants
Since digital exclusion is strongly associated with traditional forms of social exclusion,
differences between several demographic groups with regard to internet skills can be
suggested. The second research question is:
RQ 2: Which factors determine internet skills levels?
Table 1. Medium- and content-related internet skills
Medium-related internet skills
Operational internet skills Operating an internet browser:
Opening websites by entering the URL in the browser’s
location bar;
Navigating forward and backward between pages using the
browser buttons;
Saving files on the hard disk;
Opening various common file formats (e.g., PDFs);
Bookmarking websites.
Operating internet-based search engines:
Entering keywords in the proper field;
Executing the search operation;
Opening search results in the search result lists.
Operating internet-based forms:
Using the different types of fields and buttons;
Submitting a form.
Formal internet skills Navigating on the internet, by:
Using hyperlinks embedded in different formats such as texts,
images, or menus.
Maintaining a sense of location while navigating on the internet,
meaning:
Not becoming disoriented when navigating within a website;
Not becoming disoriented when navigating between websites;
Not becoming disoriented when opening and browsing
through search results.
Content-related internet skills
Information internet skills Locating required information by:
Choosing a website or a search system to seek information;
Defining search options or queries;
Selecting information (on websites or in search results);
Evaluating information sources.
Strategic internet skills Taking advantage of the internet by:
Developing an orientation toward a particular goal;
Taking the right action to reach this goal;
Making the right decision to reach this goal;
Gaining the benefits resulting from this goal.
Source: (Van Deursen and Van Dijk, 2009a, 2010).
van Deursen and van Dijk 897
Regarding gender, the outcomes are not consistent. Goulding and Spacey (2002) claim
that men possess more knowledge about the internet and the way to use it than women
since the latter have been slower to start using the internet than men have. Wasserman and
Richmond-Abbott (2005) found that the level of internet use was related to web knowl-
edge, and that this was higher among men than among women. Schumacher and Morahan-
Martin (2001) concluded that men possess greater internet skills than women do. Hargittai
and Shafer (2006) found that men and women do not differ greatly in their online abili-
ties, but that women’s self-assessed skill is significantly lower than that of men.
H1. There are no differences of internet skill levels between men and women.
Older people are often regarded as ‘laggards’ in the diffusion process for innovations
(Rogers, 1995). Young people get to know the internet at an early age and are considered
more skillful than seniors. It is often believed that the so-called digital generation possesses
the highest level of internet skills. Seniors have never had the opportunity to acquaint
themselves with the internet at school and lag behind in their use of the internet as well as
their digital skills (De Haan and Huysmans, 2002). In line with these statements, Hargittai
(2002) concludes that age is negatively associated with one’s level of internet skills.
H2. With age an increasing number of adults show a lower level of internet skills.
Education is the most consistent global predictor of the use of ICTs. The higher edu-
cated more often own computers, have internet access at home, and connect through
broadband and spend more time online (Buente and Robbin, 2008). Strongly related to
educational attainment are cognitive resources that are largely responsible for differences
in internet use and in the digital skills of different educational groups (De Haan et al.,
2002). Goldin and Katz (2008) argue that the more highly educated are able to keep up
with technological advancements and therefore increase their lead over people who are not
able to keep up. However, they also concluded that education in digital skills cannot keep
up with technological developments, which has resulted in wage inequality in the US.
H3. With educational level an increasing number of people show a higher level of internet
skills.
People who spend more time online will acquire more knowledge about the internet
and thus develop better online skills (Hargittai, 2002, 2005). Moreover, people who
have been internet users for a longer period of time are expected to be better at finding
information online because they have more experience to draw upon (Hargittai, 2002,
2005). Indeed, Hargittai (2002) found years of experience and intensity of use to serve
as strong predictors of internet skills. In general, for both computers and the internet,
the length of previous experience and the amount of current usage have been associated
with greater technological expertise (Schumacher and Morahan-Martin, 2001).
H4. With internet experience an increasing number of people show a higher level of internet
skills.
898 new media & society 13(6)
H5. With internet usage time an increasing number of people show a higher level of internet
skills.
Social resources consist of access people have to other sources of help and training
(Robinson et al., 2003). According to Warschauer (2002), users should participate
within a social setting: ‘digital literacy’ is a social practice, involving access to physical
artifacts, content, skills, and social support. Van Dijk (2006) considers the social con-
text of internet users to be a decisive factor in the opportunities they have for learning
internet skills. Thus, uptake and also further usage of the internet may be significantly
affected by the amount of social support to which a user has access.
H6. People who have access to social support when using the internet show a higher level of
internet skills than people who do not.
Other factors that might be significant in the possession of internet skills are socio-
economic position, the location of internet use, and participation in an internet course.
Students and people in jobs are more likely to use the internet than people that are retired
or unemployed (although also undergraduate students with a relatively high level of
education demonstrate limitations in internet skills (e.g., Davis, 2003; Volman et al.,
2005).
H7. With socio-economic status increasing, people show a higher level of internet skills.
People forced to use the internet at school or in libraries have less time to practice
which might result in a restriction of their internet skills.
H8. People who use the internet most often at home will show a higher level of internet skills
than people who most often use it elsewhere.
Finally, a myriad of internet courses are offered both in education and in labor organ-
izations. Those that take part in these courses and learn to use the internet are more
likely to possess sufficient internet skills than those who do not (Anandarajan et al.,
2000). These might help people to improve their level of internet skills.
H9. People who participated in an internet course show a higher level of internet skills than
people who did not.
Method
Sample
Two studies were conducted to measure the level of internet skills among the Dutch
population. In line with the procedure used by Hargittai (2002), we used the condition of
invitation that the subject used the internet at least once every month for more than just
e-mail. This condition excluded approximately 20 percent of the Dutch population, but
ensured that low-frequency users who were nonetheless familiar with the internet were
van Deursen and van Dijk 899
included. The subjects were not informed about the exact intention of the studies. The
invitation policy put people who feared a test at ease.
To be able to generalize from the findings, the subjects were recruited by applying a
stratified random sampling method. First, a sample was randomly selected from a tele-
phone book of an eastern region in the Netherlands. Subsequently, a selective quota
sample was drawn from the strata of gender, age (equal number of subjects in the catego-
ries of age 18–29, 30–39, 40–54, and 55–80), and educational level of attainment (equal
number of subjects in the categories low, middle, and high) to reach equal subsamples.
Digital divide research has repeatedly shown that access to and use of the internet is
heavily stratified by these variables. When respondents indicated they were willing to
participate, contact details were recorded, and a time for the research session was sched-
uled. Respondents received a confirmation letter in the mail. The day before the study,
respondents were reminded of the session by phone. After the session (which took
approximately one and a half hours), subjects were rewarded with 25 Euros.
Procedure
The first study took place between September and December 2007 and the second study
between September and December 2008. Both performance tests were conducted in a uni-
versity office. Although this forced subjects to use a computer that might be configured
differently from the machine they normally use, this approach does control for quality of
the hardware/software and internet connection and ensured that the setting is similar for all.
After arriving at the laboratory, subjects were given verbal instructions about the pro-
cedure. Prior to the test, a ten-minute questionnaire was administered to gather personal
data. After the subjects completed the questionnaire, they were given a sequence of nine
assignments, one at a time. Subjects themselves decided when they were finished or
wanted to give up on an assignment. No encouragements were given because the pressure
to succeed is already higher in a laboratory setting than at home. After a specific maxi-
mum amount of time had passed (determined from pilot-tests), the test-leader gently asked
the subjects to move on to the next assignment. If the correct answer was not found, the
task was rated as not completed. The test-leader directly measured successful completion
of the tasks and the time spent but refrained from influencing the subjects’ strategies.
During the assignment completion, subjects used a keyboard, a mouse, and a 17-inch
monitor. These were connected to a laptop from which the test-leader could watch the
subjects’ actions. The laptop had access to a high-speed university network for internet
use and was programmed with the most popular internet browsers. This allowed subjects
to replicate their regular internet use. No default page was set on the browsers, and all the
assignments started with a blank page. To ensure that subjects were not influenced by
previous users’ actions, the browser was reset after each session by removing temporary
files, cookies, and favorites. In addition, downloaded files, history, form contents, and
passwords were removed, and the laptop was rebooted.
Assignments
Two assignments (consisting of eight tasks) were used to measure operational internet
skills, two (consisting of four tasks) for measuring formal internet skills, three for
900 new media & society 13(6)
measuring information internet skills, and two for measuring strategic internet skills. The
total outcome for every type of internet skills is measured as the number of tasks solved
successfully and the time spent on these tasks. All assignments were fact-based and have
a specific correct action or answer. Open-ended tasks are avoided because of the ambigu-
ity of interpretation of the many potential answers.
In the first study, the assignments related to government information and services.
The second study used a more popular context, in which general, leisure-related assign-
ments geared toward the consciousness of all internet users had to be completed (see
http://www.alexandervandeursen.nl/nms/ for the complete list of the assignments). In the
operational internet skills assignments, subjects were for example asked to save a file,
bookmark a website, or fill out an online form. The formal internet skills assignments
included for example navigating different website designs and surfing between different
websites and search results. The information skills assignments charged subjects with
finding information. The assignments ranged from looking up a Michelin-awarded res-
taurant in Amsterdam to finding minimum wage levels in a specific year. Finally, in order
to complete the strategic skill assignments, subjects, for example, had to work in a struc-
tured manner and make decisions based on retrieved information. They were for example
asked to book a specific trip as profitably as possible, or find the political party that best
matches some particular positions.
Explanatory variables
All discussed explanatory variables are accounted for in the internet skills measure-
ments. Gender is included as a dichotomous variable and age as a continuous variable.
Data on education were collected by degree. These were subsequently divided into three
overall groups of low, medium, and higher educational attainment. Internet experience
was measured as the number of years people have been using the internet. The amount
of internet use was measured by the number of hours respondents spend browsing the
internet weekly (the week before the survey was used as reference). Furthermore, partici-
pation in an internet class, whether at school or elsewhere, was considered dichotomous
(no/yes). In the regression analyses reported later, the data on social support, location of
internet use, and working situation, are also transformed to dichotomous variables.
Working situation was divided into active (employers, employees, and students) or inac-
tive (retired, disabled, homemakers, unemployed) groups. This transformation was per-
formed because the number of subjects in some of these groups is very low (see Table 2)
and because adding these as dichotomous variables provides information about the vari-
ables as a whole, instead of each separate group.
Missing data
The number of missing values was low. During the assignment completion, the test-
leaders were instructed to check all data carefully for missing answers. This resulted in
only two missing values in the first study and three in the second study. The available
data of the subjects that forgot to answer a question were still used, given the limited
number of subjects in both studies.
van Deursen and van Dijk 901
Results
Sample
In both studies, 109 subjects performed the tests (response rates to the invitation were
28% and 32% respectively). Table 2 contains the characteristics of the subjects. The
quota sample applied was designed to ensure a sufficient number of subjects in different
gender, age, and educational groups to obtain valid data. The average years of internet
Table 2. Subjects over gender, education, age, location of internet use, requiring assistance,
socio-economic status, and participation in an internet course
Study 1 Study 2
n%n%
Gender
Male 51 47 57 52
Female 58 53 52 48
Age
18–29 25 23 27 25
30–39 27 25 23 21
40–54 27 25 28 26
55–80 30 28 30 28
Education
Low (e.g., Primary school) 32 29 34 31
Middle (e.g., High school) 37 34 34 31
High (e.g., College and University) 40 37 41 38
Primary location of internet use
At home 95 87 104 95
At work 13 12 5 5
At school 1 1 0 0
At friends or family 0 0 0 0
At a library 0 0 0 0
Assistance when using the internet
No 51 47 45 41
Yes, from family 32 29 29 26
Yes, from friends 18 17 27 25
Yes, from colleagues 5 5 5 5
Yes, from a helpdesk 2 2 3 3
Socio-economic status
Employee 54 50 52 47
Retired 18 17 19 18
Student 14 13 16 15
Houseman/housewife 10 9 11 10
Employer 7 6 5 5
Disabled 5 5 3 3
Unemployed 1 1 2 2
Participation in an internet course
No 84 77 81 74
Yes 25 23 28 26
902 new media & society 13(6)
experience was 8.1 (SD = 3.0) and 8.3 (SD = 3.2) and the average amount of internet use
was 9.7 (SD = 9.7) and 9.5 hours a week in studies 1 and 2 respectively. Overall, the people
who participated in the studies represented a diverse group of internet users.
Levels of internet skills
According to Table 3, the subjects in the first study altogether completed 7.2 of the nine
operational tasks (80%). They completed an average of 2.9 of the four formal skills tasks
(72%) and an average of 1.9 of the three information skill tasks (62%). Most problematic
are the two strategic tasks of which the subjects only completed 0.5 overall (25%). Only
11 percent of the subjects were able to complete both strategic skill tasks. In the second
study, the results are similar. Here, the subjects altogether completed an average of 6.6 of
the nine operational tasks (73%), an average of 3.3 of the four formal skills tasks (83%),
an average of 1.6 of the three information skill tasks (53%), and an average of 0.6 (30%)
of the two strategic skill tasks. In both studies, the time spent on the information and
strategic tasks varies substantially.
Table 4 reveals that 39 percent of the subjects were able to complete all operational
internet skills tasks, 33 percent were able to complete all formal internet skills tasks, 21
percent were able to complete all information internet skills tasks, and only 11 percent of
the subjects were able to complete both the strategic skills tasks in the first study. In the
second study these figures were 25, 55, 16, and 13 percent respectively. As much as 62
percent of the subjects in the first study and 56 percent of the subjects in the second study
could not complete any of the two strategic internet skills tasks successfully.
Internet skills determinants
To identify factors that influence the level of internet skills, two linear regressions for all
four internet skills were conducted: one with the number of assignments completed suc-
cessfully as dependent variable and one with the time spent on these assignments as
dependent variable.
According to Table 5, age is the main contributor for number of operational skill
assignments completed and the time spent. The older the subject, the fewer assignments
are completed and the more time is spent on the assignments. Educational attainment is
Table 3. Overview of successful task completion and time spent
Task completion Time (s) spent
Study 1
M(SD)
Study 2
M(SD)
Study 1
M(SD)
Study 2
M(SD)
Operational tasks (8) 6.3(1.9) 5.9(1.9) 553(254) 409(185)
Formal tasks (4) 2.9(1.0) 3.2(1.0) 616(255) 443(214)
Information tasks (3) 1.9(0.8) 1.6(1.0) 939(449) 919(327)
Strategic tasks (2) 0.5(0.7) 0.6(0.7 1,466(575) 1,628(534)
van Deursen and van Dijk 903
Table 4. Number of tasks the subjects failed to complete successfully
# Failed tasks % of subjects
Study 1 Study 2
Operational internet skills 0 39 25
1 17 13
2 10 16
3 11 18
4 12 8
5 5 9
6 5 6
7 1 4
8 0 0
Formal internet skills 0 33 55
1 32 22
2 22 16
3 11 6
4 1 0
Information internet skills 0 21 16
1 52 45
2 18 22
3 8 17
Strategic internet skills 0 11 13
1 28 31
2 62 56
Table 5. Linear regression results of the number of operational internet skills tasks completed
successfully and time spent
# Tasks completed
successfully
Time spent
β – Study 1 β – Study 2 β – Study 1 β – Study 2
Gender (m/f) –.06 –.07 –.08 .05
Age –.30*** –.50*** .43*** .57***
Education (low-medium-high) .32*** .20** –.27*** –.06
Years online .15* .03 –.18** –.20**
Hours online weekly .04 .15 –.10 –.15*
Internet class (no / yes) .03 –.04 –.01 .06
Assistance required (no / yes) –.12 –.07 .13 .11
Primary location of internet use
(at home / elsewhere)
.08 –.16* –.07 .04
Working situation (inactive / active) –.15 –.11 –.16* .08
R2.52 .55 .64 .65
F14.02*** 15.46*** 22.34*** 23.09***
*p <.05, **p <.01, ***p <.001.
904 new media & society 13(6)
also significant for the number of tasks completed and in the first study also for the time
spent. Other factors appear less important. Internet experience is significant in the first
study for the number of tasks completed, and in both studies for the time spent.
Regarding formal internet skills, age again appears the most important predictor in
both studies. In addition, educational attainment is significant for the number of formal
tasks completed. See Table 6.
The regression results reported in Table 7 indicate that educational attainment is the
main contributor to the number of information tasks completed in both studies. Age is
Table 6. Linear regression results of the number of formal internet skills tasks completed
successfully and time spent
# Tasks completed successfully Time spent
β – Study 1 β – Study 2 β – Study 1 β – Study 2
Gender (m/f) .08 .09 -.15 -.05
Age -.25** -.35*** .46*** .48***
Education (low-medium-high) .26* .27*** -.16* -.11
Years online .13 .21 -.13 -.17*
Hours online weekly -.02 -.02 -.13 -.13
Internet class (no / yes) .07 .10 -.02 -.11
Assistance required (no / yes) -.26** -.03 .13 .10
Primary location of internet use
(at home / elsewhere)
-.18* -.24** -.05 .19**
Working situation (inactive / active) .12 -.17 -.09 .09
R2.49 .51 .57 .55
F12.39*** 11.36*** 16.46*** 15.64***
*p <.05, **p <.01, ***p <.001.
Table 7. Linear regression results of the number of information internet skills tasks completed
successfully and time spent
# Tasks completed successfully Time spent
β – Study 1 β – Study 2 β – Study 1 β – Study 2
Gender (m/f) -.13 -.10 -.01 .15
Age -.12 -.05 .23 .31*
Education (low-medium-high) .36*** .28*** -.22* .16
Years online .07 .13 -.04 -.25
Hours online weekly -.11 .03 .02 .02
Internet class (no / yes) .02 -.02 .00 .17
Assistance required (no / yes) .00 -.21* .19 .11
Primary location of internet use
(at home / elsewhere)
.11 -.08 -.07 -.05
Working situation (inactive / active) -.04 -.06 -.16 .03
R2.13 .27 .23 .18
F2.82*** 3.99*** 4.67*** 3.55***
*p <.05, ***p <.001.
van Deursen and van Dijk 905
significant for the time spent on the tasks. Finally, using help from peers is significant for
the number of tasks completed in the second study.
According to Table 8, educational attainment is the only predictor for the number of
strategic tasks completed in both studies. There are no significant predictors for the time
spent on the strategic tasks.
Hypotheses
Hypothesis H1 – that there are no differences of internet skill levels between men and
women – is supported. Gender did not appear as a significant contributor for any of the
four internet skills.
Hypothesis H2 – that with age an increasing number of adults show a lower level of
internet skills – is partly supported. The elderly perform more poorly than the younger
generations with regard to operational and formal internet skills. However, age did not
appear as a significant contributor to the level of information and strategic internet skills.
Hypothesis H3 – that with educational level an increasing number of people show a
higher level of internet skills – is supported. The level of educational attainment affects
all four types of internet skills.
Hypothesis H4 – that with internet experience an increasing number of people show
a higher level of internet skills – is only supported for the operational internet skills. It
appears that formal, information, and strategic internet skills do not grow with years of
internet experience.
Hypothesis H5 – that with internet usage time an increasing number of people show a
higher level of internet skills – is rejected with the partial exception of operational skills.
Internet usage time only negatively contributed to the time spent on the operational tasks
in the second study.
Table 8. Linear regression results of the number of strategic internet skills tasks completed
successfully and time spent
# Tasks completed successfully Time spent
β – Study 1 β – Study 2 β – Study 1 β – Study 2
Gender (m/f) -.06 -.10 -.11 .14
Age -.17 .02 -.03 .02
Education (low-medium-high) .42*** .42*** .13 .14
Years online .02 .06 .06 .10
Hours online weekly -.15 -.06 -.14 .03
Internet class (no / yes) .03 -.01 .05 .21
Assistance required (no / yes) -.16 -.12 .14 .07
Primary location of internet use
(at home / elsewhere)
-.05 -.09 -.03 -.06
Working situation
(inactive / active)
.14 -.07 -.08 -.02
R2.30 .25 .01 .09
F6.09*** 3.75*** .84 1.07
***p <.001.
906 new media & society 13(6)
Surprisingly, Hypotheses H6 – that people who have access to social support when
using the internet show a higher level of internet skills than people that do not, and H7 –
that with socio-economic status increasing, people show a higher level of internet skills,
and H9 – that people who participated in an internet course show a higher level of inter-
net skills than people who did not – are all rejected. Only the primary location of use had
some minor influence on mainly operational and formal internet skills. People who use
the internet most often at home, rather than at work, libraries, with friends or in internet
cafés seem to perform slightly better on these skills. Thus Hypothesis H8 – that people
who use the internet most often at home will show a higher level of internet skills than
people who most often use it elsewhere – is partly supported.
Research questions
Regarding research question 1, it appears that the Dutch population on average has a
fairly high level of operational and formal internet skills, but that the levels of informa-
tion and especially strategic internet skills attained are much lower. The latter leave
much room for improvement. It is important to understand that operational and formal
internet skills are not sufficient for effective use of the internet and that information and
strategic internet skills are more troublesome. The general assumption that assistance
can always be provided to those who have insufficient skills might be partly true for rela-
tively basic operations, but certainly not for more complicated ones that require informa-
tion and strategic internet skills. The results of the performance tests force policy makers
and new media developers to adjust their beliefs that, with the exception of some seniors,
everybody has access to and can use the internet.
Regarding research question 2, it appears that age and educational attainment are the
most important contributing factors. Age is significant for operational and formal skills.
The younger generation performed better on operational and formal skills, but not on
information and strategic skills. Educational attainment appears significant for all inter-
net skills. Other research results revealed that people learn digital skills more in practice,
by trial and error, than in formal educational settings (De Haan and Huysmans, 2002;
Van Dijk, 2005). We argue that this is probably the case for operational and formal skills,
but certainly not for information and strategic internet skills. Internet experience only
contributes to the level of operational internet skills. It appears that formal, information,
and strategic internet skills do not grow with years of internet experience. Furthermore,
amount of time spent online weekly only negatively contributed to the time spent on the
operational tasks in the second study. Survey research that uses internet self-efficacy as
a dependent variable often finds prior internet experience to be a very strong predictor.
An explanation for the observed weak relation of internet experience and time spent
online with operational and formal skills might be the fact that people often keep repeat-
ing similar mistakes when using the internet. Computer users tend to rely on acquired
skills, even when they are aware that they could learn more efficient procedures for
achieving the same results (Cahoon, 1998). This probably also accounts for internet use.
People learn by trial and error, but when they more or less achieve the goals they had in
mind, people will persist in making the same mistakes online.
van Deursen and van Dijk 907
Discussion
The two studies discussed intended to measure an in-depth range of internet skills actu-
ally commanded by the Dutch population. While other definitions of internet skills often
focus on specific aspects of internet use, we have applied a definition of internet skills
that is derived from multiple research directions and subsequently arranged in a particu-
lar order. An important characteristic of the applied definition is the sequential and con-
ditional nature of four types of internet skills. The results of the study emphasize the
value of this particular distinction by revealing that especially the information and strategic
internet skills appear to be the most difficult in performance tests and leave much room
for improvement. A second characteristic of the applied definition is that it concentrates
on task- and goal-oriented internet use. This goes beyond the more traditional definitions
of media literacy by suggesting a more (inter)active use (consisting of interaction with
programs or people, transactions in goods and services, and making decisions). A third
characteristic of the applied definition is that the considered internet skills should be
commanded by every internet user if they want to effectively use this medium at all.
Although the results of both studies are similar, the temptation to draw conclusions
about absolute levels of performance should be avoided. There are a few other perform-
ance tests of users’ internet skills, but these did not distinguish between the four skills
applied here. Unfortunately, this means that there are no direct standards of comparison
within the Netherlands, or any other country. Four general claims can be made:
1. The Netherlands is a country with very high household internet penetration (93%
in 2010) and a high level of educational attainment. It is to be expected that per-
formance in many other countries in the world will be lower.
2. Evidently, the results depend on the difficulty of the assignments. In the first
study, relatively simple tasks of using online government information were
selected. In the second study, an even more feasible context was used, as virtually
everyone is familiar with the leisure activities used for the internet assignments.
The results of both studies appeared similar, suggesting that previously existing
knowledge about government issues seems to have had a minor effect on the
level of internet skills.
3. The two studies were administered to Dutch-speaking subjects only. Since the
amount of content available in other languages, especially English, is larger than
available content in the Dutch language, employing information and strategic
internet skills in these languages might be even more difficult.
4. In actual internet use outside the artificial test situation, performances can be
expected to be lower. In a test situation, subjects are often more motivated to
complete an assignment (though, in this case, they were not explicitly spurred).
In their own environments, many of them would have grabbed the phone or run
to a service desk or someone else in their social environment to get the answer.
Other research among the Dutch population indicates that users of public web-
sites often give up and turn to the telephone or a front desk (Pietersen and Ebbers,
2008).
908 new media & society 13(6)
In our studies we did not account for communication skills since this would have
made the performance tests that already required 1.5 hours from the subjects an unreal-
istic effort. The skills necessary for content creation and content sharing have also been
neglected. However, in our view, both information and strategic internet skills are also
crucial for these activities. Active participation and user-generated content require a high
level of internet skills, particularly for ‘serious’ as compared to entertainment applica-
tions. Both limitations are tasks for future research.
We do not know whether the lack of information and strategic skills we have observed
appears to the same extent in the information retrieval of traditional media. Further
research should address this question by including comparable information and strategic
skills performance tests in the use of media other than the internet. A comparison of the
results of these tests would reveal whether internet use produces better test results or
introduces another barrier because many people do not master the special skills required
for appropriate use of the internet. Finding information in a traditional library might be
more difficult for inexperienced information seekers than finding the same information
on the internet using a ‘simple’ search engine. However, the internet makes information
seeking and improving literacy more difficult, as they assume a number of new opera-
tional and formal skills. This raises yet another barrier in addition to the skills of reading
and writing. Confirmative research is necessary to reveal whether operational, formal,
information, and strategic skills together increase the gap between people of differing
ages and educational and occupational backgrounds in terms of the new (as opposed to
traditional) media.
Conclusion
The results of the conducted studies strengthen the findings that the original digital
divide (defined as the gap between people who have and do not have physical access to
computers and the internet) has developed a second divide that includes differences in
the skills to use the internet. For a better understanding of these skills divides, it is impor-
tant to consider both medium- and content-related internet skills in future measurements.
In digital divide research, the conclusion that operational and formal internet skills are
not sufficient for an effective use of the internet so far only received little attention.
Information and strategic internet skills are also required. In contemporary (and future)
information society these skills increasingly determine people’s positions in the labor
market and in social life. Unfortunately, these skills appear to be the most problematic
and a large part of the Dutch population seems to be struggling to equip themselves with
the skills they need to participate in contemporary society.
If people with low levels of internet skills fail to find information online while an
increasing amount relevant to daily life become easiest to access on the internet, they
become increasingly disadvantaged. The results of the studies strongly indicate that large
parts of the population will be excluded from actual and effective internet use. This espe-
cially goes for less educated populations. While these groups have always been socially
disadvantaged, their life chances are now further in danger. They are increasingly
excluded from all the benefits the internet now has to offer, ranging from economic
opportunities such as privileged access to jobs, health opportunities such as better diets
van Deursen and van Dijk 909
or improved exercise habits, or political opportunities such as online services and partici-
pation. This exclusion appears over all age groups within society. In the beginning,
groups with fewer internet skills will be persuaded in negative ways; flights will be
booked, concerts will be sold out, jobs will be given away, and dates will primarily be
granted to those having access. Continuing the transformations towards the internet in
the most important domains of life will eventually lead to serious problems if the lack of
internet skills among large parts of the population is not accounted for. Unfortunately, only
the lack of operational and formal internet skills might be considered as a temporary prob-
lem (until a better accessible technology appears). The lack of information and strategic
skills seems to be more structural. These skills strongly relate to education and intellectual
capacities and should therefore gain a more central position in future research. While
originally the digital divide could be ‘easily’ addressed by providing physical access, this
now seems to be much harder when content-related internet skills are considered. Although
several recommendations can be provided (see, for example, Van Deursen and Van Dijk,
2009b), one might seriously question whether the digital divide can be closed at all.
Funding
This research received no specific grant from any funding agency in the public, commercial, or
not-for-profit sectors.
References
Ahuja JS and Webster J (2001) Perceived disorientation: An examination of a new measure to
assess web design effectiveness. Interacting with Computers 14(1): 15–29.
Anandarajan M, Simmers C and Igbaria M (2000) An exploratory investigation of the antecedents
and impact of internet usage: An individual perspective. Behaviour & Information Technology
19(1): 69–85.
Aula A and Nordhausen K (2006) Modeling successful performance in Web searching Journal of
the American Society for Information Science and Technology 57(12): 1678–1693.
Buente W and Robbin A (2008) Trends in internet information behavior 2000–2004. Journal of the
American Society for Information Science and Technology 59(1): 1743–1760.
Cahoon B (1998) Teaching and learning internet skills. New Directions for Adult and Continuing
Education 78: 5–13.
Davis PM (2003) Effects of the web on undergraduate citation behavior: Guiding student scholar-
ship in a networked age. Libraries and the Academy 3(1): 41–51.
De Haan J and Huysmans F (2002) Differences in time between internet users and nonusers in the
Netherlands. IT&Society 1(2): 67–85.
De Haan J, Huysmans F and Steyaert J (2002) At Home in a Digital World: Acquiring Digital Skills
between Home and School. The Hague: Social and Cultural Planning Agency.
DiMaggio P and E Hargittai (2001) From the ‘Digital Divide’ to ‘Digital Inequality’: Studying
Internet Use as Penetration Increases. Working Paper Series 15, Princeton University Center
for Arts and Cultural Policy Studies.
Ford N and Chen SY (2000) Individual differences, hypermedia navigation, and learning: An
empirical study. Journal of Educational Multimedia and Hypermedia 9: 281–312.
910 new media & society 13(6)
Goldin C and Katz LF (2008) The Race between Education and Technology. Cambridge, MA:
Harvard University Press.
Goulding A and Spacey R (2002) Women and the Information Society: Barriers and Participation.
Paper presented at 68th IFLA Council and general conference, Glasgow, 18–24 August.
Hargittai E (2002) Second-level digital divide: Differences in people’s online skills. First Monday
7(4).
Hargittai E (2005) Survey measures of web-oriented digital literacy. Social Science Computer
Review 23(3): 371–379.
Hargittai E and Shafer S (2006) Differences in actual and perceived online skills: The role of gen-
der. Social Science Quarterly 87(2): 432–448.
Merritt K, Smith D and Renzo JCD (2005) An investigation of self-reported computer literacy: Is
it reliable? Issues in Information Systems 6(1): 289–295.
Morahan-Martin JM (2004) How internet users find, evaluate, and use online health information:
A cross-cultural review. CyberPsychology and Behavior 7(5): 497–510.
Mossberger K, Tolbert CJ and Stansbury M (2003) Virtual Inequality: Beyond the Digital Divide.
Washington, DC: Georgetown University Press.
Pan B, Hembrooke H, Joachims T, Lorigo L, Gay G and Granka L (2007) In Google we trust: Users’
decisions on rank, position, and relevance. Journal of Computer-mediated Communication
12: 801–823.
Pieterson W and Ebbers W (2008) The use of service channels by citizens in the Netherlands:
Implications for multi-channel management. International Review of Administrative Sciences
74(1): 95–110.
Robinson JP, DiMaggio P and Hargittai E (2003) New social survey perspectives on the digital
divide. IT&Society 1(5): 1–22.
Rogers EM (1995) Diffusion of Innovations. New York: Free Press.
Schumacher P and Morahan-Martin J (2001) Gender, internet and computer attitudes and experi-
ences. Computers in Human Behavior 17: 95–110.
Spink A, Wolfram D, Jansen BJ and Saracevic T (2001) Searching the web: The public and their
queries. Journal of the American Society for Information Science 53(2): 226–234.
Steyaert J (2002) Inequality and the digital divide: Myths and realities. In: Hick S and McNutt J
(eds) Advocacy, Activism and the internet. Chicago, IL: Lyceum Press, 199–211.
Talja S (2005) The social and discursive construction of computing skills. Journal of the American
Society for Information Science and Technology. 56(1): 13–22.
Van Deursen AJAM and van Dijk JAGM (2009a) Using the internet: Skill related problems in
users’ online behavior. Interacting with Computers. 21: 393–402.
Van Deursen AJAM and van Dijk JAGM (2009b) Improving digital skills for the use of online
public information and services. Government Information Quarterly 26: 333–340.
Van Deursen AJAM and van Dijk JAGM (2010) Measuring internet skills. International Journal
of Human-Computer Interaction. 26(10): 891–916.
Van Dijk J (2005) The Deepening Divide Inequality in the Information Society. London: Sage
Publications.
Van Dijk J (2006) Digital divide: Research, achievements and shortcomings. Poetics 34: 221–235.
Volman M, van Eck E, van Heemskerk L and Kuiper E (2005) New technologies, new differ-
ences: Gender and ethnic differences in pupils’ use of ICT in primary and secondary education.
Computers & Education 45(1): 35–55.
van Deursen and van Dijk 911
Warschauer M (2002) Reconceptualising the digital divide. First Monday 7(7).
Warschauer M (2003) Technology and Social Inclusion: Rethinking the Digital Divide.Cambridge,
MA: The MIT Press.
Wasserman IM and Richmond-Abbott M (2005) Gender and the internet: Causes of variation in
access, level, and scope of use. Social Science Quarterly 86: 252–270.
Witte J and Mannon S (2007) The Internet and Social Inequalities. New York: Routledge.
Alexander J.A.M. van Deursen is a researcher at the University of Twente in the
Netherlands. His dissertation, entitled ‘Internet skills, vital assets in an information soci-
ety’ has just been released. His research focuses on digital inequality with specific atten-
tion to internet skills.
Jan A.G.M. van Dijk is Professor of Communication Science and the Sociology of the
Information Society at the University of Twente, the Netherlands. He is chair of the
Department of Media, Communication and Organization and Director of the Center for
eGovernment Studies at the same university.