See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/271073805

Content creation on the Internet: a social cognitive perspective on the

participation divide

Article in Information Communication and Society · January 2015

DOI: 10.1080/1369118X.2014.991343

CITATIONS READS

94 1,124

3 authors:

Christian Pieter Hoffmann Christoph Lutz

Leipzig University BI Norwegian Business School
187 PUBLICATIONS 2,615 CITATIONS 150 PUBLICATIONS 5,737 CITATIONS

SEE PROFILE SEE PROFILE

Miriam Meckel
University of St.Gallen
239 PUBLICATIONS 1,948 CITATIONS

SEE PROFILE

All content following this page was uploaded by Christoph Lutz on 20 January 2015.

The user has requested enhancement of the downloaded file.

This article was downloaded by: [Universitaet St Gallen]
On: 19 January 2015, At: 03:15
Publisher: Routledge
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered
office: Mortimer House, 37-41 Mortimer Street, London W1T 3JH, UK

Information, Communication & Society

Publication details, including instructions for authors and
subscription information:
http://www.tandfonline.com/loi/rics20

Content creation on the Internet: a

social cognitive perspective on the
participation divide
a a a
Christian Pieter Hoffmann , Christoph Lutz & Miriam Meckel
a
Institute for Media and Communications Management, University
of St. Gallen, Blumenbergplatz 9, 9000 St. Gallen, Switzerland
Published online: 19 Jan 2015.

Click for updates

To cite this article: Christian Pieter Hoffmann, Christoph Lutz & Miriam Meckel (2015): Content
creation on the Internet: a social cognitive perspective on the participation divide, Information,
Communication & Society, DOI: 10.1080/1369118X.2014.991343

To link to this article: http://dx.doi.org/10.1080/1369118X.2014.991343

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the
“Content”) contained in the publications on our platform. However, Taylor & Francis,
our agents, and our licensors make no representations or warranties whatsoever as to
the accuracy, completeness, or suitability for any purpose of the Content. Any opinions
and views expressed in this publication are the opinions and views of the authors,
and are not the views of or endorsed by Taylor & Francis. The accuracy of the Content
should not be relied upon and should be independently verified with primary sources
of information. Taylor and Francis shall not be liable for any losses, actions, claims,
proceedings, demands, costs, expenses, damages, and other liabilities whatsoever or
howsoever caused arising directly or indirectly in connection with, in relation to or arising
out of the use of the Content.

This article may be used for research, teaching, and private study purposes. Any
substantial or systematic reproduction, redistribution, reselling, loan, sub-licensing,
systematic supply, or distribution in any form to anyone is expressly forbidden. Terms &
 Conditions of access and use can be found at http://www.tandfonline.com/page/terms-
and-conditions
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015
 Information, Communication & Society, 2015
http://dx.doi.org/10.1080/1369118X.2014.991343

Content creation on the Internet: a social cognitive perspective on the

participation divide
Christian Pieter Hoffmann*, Christoph Lutz and Miriam Meckel

Institute for Media and Communications Management, University of St. Gallen, Blumenbergplatz 9, 9000
St. Gallen, Switzerland
(Received 23 April 2014; accepted 12 November 2014)
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Sociodemographic variables are held to impact Internet users’ willingness and ability to
productively use online media. This effect can create a ‘participation divide’ between
distinct user groups. Recently, studies have enhanced our understanding of the participation
divide by differentiating types of online content creation. They found that
sociodemographics may only affect specific forms of online participation. We suggest that
social cognitive theory (SCT) helps explain why and how sociodemographic variables
influence different forms of online participation. Based on SCT, we analyze the mediating
effect of two cognitive constructs, self-efficacy and privacy concerns, on different types of
online content creation. We conduct a survey among German Internet users and apply
structural equation modeling to compare three distinct theoretical models. We find that
considering the mediating effects of cognitive constructs, based on SCT, improves our
understanding of which sociodemographic variables affect which type of online content
creation – and why.
Keywords: online participation; participation divide; social cognitive theory; content creation;
social media; digital divide

1. Introduction
To an unprecedented degree, social media have made it easy for lay users to publish their musings
and opinions and make them accessible to a wide audience (Blank & Reisdorf, 2012; Correa,
2010; Hargittai & Walejko, 2008; Schradie, 2011). Based on personal profiles, users can
quickly connect to like-minded citizens and become members of lively communities of interest
(Gil de Zúñiga, Veenstra, Vraga, & Shah, 2010; Woodly, 2007; Zhang, Johnson, Seltzer, &
Bichard, 2010). These new media affordances have triggered significant research interest in
what is called ‘online participation’: Users employing new media to create and share content
with interested audiences in order to affect their social environment.
Online content creation and participatory Internet uses are held to generate social capital, pro-
viding both group- and individual-level benefits (Hargittai & Walejko, 2008; Shah, Kwak, &
Holbert, 2001). Across a number of civic domains, online participation allows for the identifi-
cation and coordination of communities of interest, fostering self-help and mutual support and
facilitating agenda-setting efforts (Epstein, Rosenberg, Grant, & Hemenway, 2002; Livingstone
& Bober, 2004; Lutz, Hoffmann, & Meckel, 2014; Sandaunet, 2008). Online participation has

*Corresponding author. Email: christian.hoffmann@unisg.ch

© 2015 Taylor & Francis

 2 C.P. Hoffmann et al.

been shown to be closely associated with offline engagement – a relationship most frequently con-
firmed in the context of political participation (Boulianne, 2009; Towner & Dulio, 2011; Vitak
et al., 2011).
Yet, studies have consistently shown that not all users benefit equally from the participatory
opportunities provided by new media. Online participation and content creation require a more exten-
sive skill set than mere consumptive Internet uses (Hargittai, 2002; Van Dijk, 2005). Digital divide
research has shown that sociodemographics differentiate participatory web uses (Hargittai, 2002,
2010; Hargittai & Hinnant, 2008; Van Deursen & Van Dijk, 2011; Van Dijk, 2005; Zillien & Hargittai,
2009). Accordingly, research was quick to expand the notion of a ‘digital divide’ to include the socio-
economic stratification of online participation, that is, the ‘participation divide’ (Blank, 2013a; Blank
& Reisdorf, 2012; Correa, 2010; Hargittai & Walejko, 2008; Schradie, 2011).
Recently, conflicting findings have triggered a controversy on the scope and prevalence of the
‘participation divide’: In a study of British Internet users, Blank (2013a) differentiated three forms
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

of online content creation. He found that sociodemographic variables such as age, gender, or edu-
cation do not necessarily affect online participation across all domains, but rather only select
forms, political participation most notably. In a response, Schradie (2013) pointed to the persist-
ence of large social inequalities when it comes to online content creation. Both authors do not
disagree on the existence of a participation divide, but rather on its form and scope – which ulti-
mately also affect evaluations of social effects.
In this study, we will turn to social cognitive theory (SCT) to provide a more nuanced under-
standing of why sociodemographic variables might affect online content creation – and thereby
contribute to the ongoing debate on the scope of a participation divide. SCT suggests that environ-
mental influences shape cognitive predispositions which, in turn, affect behavior (Bandura, 1977,
1986). We argue that sociodemographic variables indicate environmental influences, such as train-
ing or use experience, which shape cognitive factors driving use behavior (Frenkel, 1990; Wei, Teo,
Chan, & Tan, 2011). Accordingly, cognitive factors mediate the effect of sociodemographic vari-
ables on online content creation (cf. Correa, 2010; Hargittai & Walejko, 2008, Schradie, 2013).
We will focus our analysis on the mediating role of two cognitive factors in particular: self-
efficacy and privacy concerns. Both have been shown in previous studies to significantly affect
the use of information and communications technology (ICT) (Compeau, Higgins, & Huff,
1999; Gefen & Straub, 1997; Venkatesh & Bala, 2008) and have been applied in digital divide
research (Hsieh, Rai, & Keil, 2011). Taking the mediating role of these cognitive factors into
account will allow us to better explain seemingly conflicting findings on the scope of the partici-
pation divide.
We base our analysis on a large-scale online survey conducted in Germany (N = 1488). Our
study will apply the differentiation of online content creation proposed by Blank (2013a). In order
to analyze the effect of the cognitive variables on online content creation, we compare three
alternative models conducting structural equation modeling (SEM): a direct, a fully mediated,
and a partially mediated model. We will show that considering the mediating role of self-efficacy
and privacy concerns not only increases the explanatory power of the analysis, but it also allows
for a differentiation of direct and indirect effects of sociodemographic variables on content cre-
ation, and thereby a deeper understanding of the participation divide.

2. Theoretical background
2.1. Online content creation and participation
While an increasing number of studies address antecedents, forms, and outcomes of online par-
ticipation, the concept itself remains somewhat ill-defined (Rice & Fuller, 2013). Empirical
 Information, Communication & Society 3

investigations, especially of political participation, have subsumed vastly different activities

under the umbrella of online participation, such as donating money, signing e-petitions, writing
e-mails to a government representative, or sharing photos (Best & Krueger, 2005; Calenda &
Meijer, 2009; Di Gennaro & Dutton, 2006; Hoffman, 2012; Rojas & Puig-i-Abril, 2009; Ward,
Gibson, & Lusoli, 2003).
A number of authors agree that the creation of online content constitutes a key element of
online participation (Jenkins, 2006; Schradie, 2011). In fact, some consider online participation
synonymous with content creation (Hargittai & Walejko, 2008). Others point out that the
concept of online participation goes beyond content creation, since content creation may be
strictly unidirectional in nature and address an unknown, dispersed audience. Online partici-
pation, instead, is commonly more focused on specific reference groups and more social and inter-
active in nature as its purpose is to engage in a public discourse and affect the social environment
(Gil de Zuñiga et al., 2010; Park & Perry, 2008; Schradie, 2011).
A recent literature review of online participation, accordingly, has defined the concept as ‘the
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

creation and sharing of content on the Internet addressed at a specific audience and driven by a
social purpose’ (Lutz et al., 2014). Such a broad understanding of online participation is not
limited to the political domain, but encompasses other civic spheres (Verba, Schlozman, &
Brady, 1995), as in the case of cultural, business, or educational participation (Lutz et al.,
2014). Also, online participation may lead to or be closely associated with participation in the
offline world. For example, Moy, Manosevitch, Stamm, and Dunsmore (2005) found that inter-
active and community uses of the Internet are positively associated with offline civic engagement
(cf. Dutta-Bergman, 2006).
Studies of online participation agree that participation can be distinguished from more passive
or consumptive Internet uses (Hoffman, 2012). Yet, consumptive Internet uses may facilitate
online participation: Bakker and de Vreese (2011) found that using the Internet for information
purposes, such as news consumption, is positively associated with both online and offline partici-
pation (cf. Pasek, More, & Romer, 2009; Shah et al., 2001; Wang, 2007).
In summary, there exists a significant overlap between the concepts of online participation and
content creation, in some cases rendering both indistinguishable (Schradie, 2011). Online partici-
pation is considered more resource intensive than other, more consumptive Internet uses, requir-
ing a more extensive skill set (Hargittai, 2002; Van Dijk, 2005). It can also be considered a more
capital-enhancing use of the medium (Hargittai & Walejko, 2008). This may be one reason why
research interest has been migrating from the ‘digital divide’ to the so-called participation divide.

2.2. From the digital divide to the participation divide

Digital divide research suggests that offline inequalities are reproduced online. More specifically,
socioeconomic differences tend to be replicated online, with socioeconomic disadvantages limit-
ing user access to Internet and benefits from Internet use. Antecedents frequently explored in the
digital divide literature include education, income, gender, and age. While early studies of the
digital divide tended to focus on access to the Internet or the extent and frequency of Internet
use, more recent studies began to differentiate forms of Internet use as well as the associated
skill requirements:
Socioeconomic status (SES), as measured by income, occupational status, social background,
or education, is a key construct in the digital divide literature. Those with high SES are held to
more easily take advantage of the Internet, because they command the necessary resources
(material, human, and social capital) allowing better access to modern ICTs (DiMaggio, Hargittai,
Celeste, & Shafer, 2003; Van Dijk, 2006) – for example, broadband Internet connection,
 4 C.P. Hoffmann et al.

smartphones, or tablets. They also possess the necessary skills to use new media (Gui & Argentin,
2011; Hargittai, 2002, 2010; Hargittai & Shafer, 2006; Van Deursen & Van Dijk, 2011).
Hargittai and Walejko (2008) found that higher social status is associated with more expres-
sive Internet uses. Compared with low SES users, those with high SES are expected to use the
Internet in more capital-enhancing ways (Hargittai & Hinnant, 2008; Zillien & Hargittai,
2009). Education and income have both been shown to positively impact online participation
(Gibson, Lusoli, & Ward, 2005). Even among youth, male, higher status, and better educated citi-
zens are more engaged online than their female, lower status, and less educated counterparts
(Livingstone, Bober, & Helsper, 2005).
As for gender, although differences in access to the Internet have almost leveled out in many
Western countries, inequalities remain (Helsper, 2010; Li & Kirkup, 2007; Ono & Zavodny,
2003). Online games or sexual content, for example, have been shown to be male-dominated
uses, while online health information seems to be more popular among female users (Helsper,
2010: pp. 356–357). Studies tend to find that men are more active and eager to participate
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

online than women (Calenda & Meijer, 2009; Di Gennaro & Dutton, 2006).
Age is found to be a strong predictor of Internet use and skills, with younger users being more
active and skillful (Bridges, Appel, & Grossklags, 2012; Dahlgren, 2011; Hargittai, 2002, 2010;
Jugert, Eckstein, Noack, Kuhn, & Benbow, 2013). On the other hand, the effect of age on online
engagement can be moderated by the users’ interest – for example, older users being more inter-
ested in political affairs (Gibson et al., 2005; Wang, 2007).
In summary, the digital divide literature provides substantial support for the notion that socio-
economics not only impact if, but also how individuals use the Internet. These findings have
implications for the participatory effect of new media, as different use patterns are more or less
conducive to individuals’ participation in a number of civic domains. Given the provided defi-
nition of online participation, ‘participation divide’ can be understood as differences in the
online creation and sharing of purpose-driven content with specific audiences due to socioeco-
nomic influences. An increasing number of studies have tried to examine the participation
divide, resulting in some controversy and need for further exploration.

2.3. Differentiating participation divides

Current findings on the antecedents of online participation are closely in line with those developed
in digital divide studies. It is noteworthy, though, that many studies of the participation divide
define and operationalize online participation strictly as online content creation (Correa, 2010;
Hargittai & Walejko, 2008). Recently, SES has been shown to significantly affect content cre-
ation, while gender affects the type of content created by users (Hargittai & Walejko, 2008).
Schradie (2011) finds that the effect of SES is even more pronounced on participatory than on
consumptive Internet uses. Variables such as income and education, particularly, affect users’ pro-
pensity to create and share content. Correa (2010) finds that among college students, gender, age,
and race (but not SES) affect the level of online content creation. As for age, the consistent finding
is that younger users are more prone to make use of the participatory forms of the Web, such as
blogs (Schradie, 2012).
Previous studies – while significantly extending our understanding of the participation divide –
are associated with some limitations (see Table 1): Some studies did not rely on random samples,
but focused on college students alone (Correa, 2010; Hargittai & Walejko, 2008). Schradie (2011)
based her analysis on a representative sample of the US population; Blank (2013a) on one of the
UK population. Most studies focus on data collected in the United States (cf. Blank, 2013a).
Methodologically, previous studies relied on either linear or logistic regression, alone.
 Information, Communication & Society 5

Table 1. Key studies of online content creation.

Hargittai and
Study Walejko (2008) Correa (2010) Schradie (2011) Blank (2013a)
Main Age and online Psychological factors SES affects user Differentiates types of
findings skills predict (motivation, skills) propensity to content; the effect
online sharing of and demographics create and share of SES on content
creative content. predict content content, effects creation differs by
Production of creation, SES does differ by online type of content
creative content not activity (mainly political
also depends on content)
SES
Sample First-year college US college students Pew sample OxIS sample
students from an (N = 3139) representative of representative of
urban US public US populationa UK populationb
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

university (N =
1060)
Types of Four types of Aggregates 10 Focus on activities Differentiates
content creative content different types of rather than political, skilled,
content into one content type and social and
measure entertainment
content
Method Logistic regression, Linear regression, Logistic regression, Logistic regression,
stepwise stepwise stepwise stepwise
a
The Pew Research Center’s Internet and American Life Project conducts regular and nationwide surveys on US
American’s use of Internet and mobile technology and attitudes toward Internet-related questions. http://www.pewinternet.
org/.
b
The Oxford Internet Survey (OxIS) is a broad survey on Internet topics that is conducted biennially in the UK. http://oxis.
oii.ox.ac.uk/.

Recently, Blank (2013a) pointed out a limitation of previous studies that goes to the heart of
the established understanding of the participation divide: a lack of differentiation of the type or
form of content produced by users. Given that content creation is such an integral element of
online participation, research should take account of the complexity of the concept by taking a
closer look at the content produced. In his analysis, Blank (2013a) differentiated three forms of
online content creation using exploratory factor analysis: skilled content, social and entertainment
content, and political content. Skilled content includes activities such as writing a blog, maintain-
ing a website, and posting writing (or other creative content). Social and entertainment content
captures the use of social network sites (SNS) for posting pictures, and uploading video or
music files. Finally, political content entails disseminating political content and commenting on
political/social issues.
Using logistic regressions, Blank (2013a) differentiates the antecedents of the three forms of
content creation. He finds that skilled content is produced by young, technically savvy people
who use multiple devices and are comfortable revealing personal information. Social status did
not affect this form of content creation. Social and entertainment content is produced by
young, technically skilled people of lower incomes. It is more likely to be created by nonelite
users. Finally, political content is produced by well-educated users who are either students or
use the Internet at work. They constitute a highly educated elite.
In other words, Blank finds that the shape and scope of the participation divide are contingent
upon the type of content created by Internet users. More specifically, the positive effect of SES on
content creation is limited to the political domain. The creation of social and entertainment
content, instead, appears to be more of a nonelite phenomenon, with higher SES being negatively
 6 C.P. Hoffmann et al.

associated with online participation. These findings and implications triggered a heated exchange
on the antecedents of online participation, the prevalence of social inequalities on the Internet, and
methodological challenges when addressing these issues (cf. Blank, 2013b; Schradie, 2013).
Schradie (2013) notes that stronger theoretical foundations are necessary to explain the differing
effects of sociodemographic variables on content creation.
We propose that our understanding of online content creation and participation divides can be
enhanced, and the effect of sociodemographic variables better explained if we consider learnings
from SCT. The next segment will provide a brief overview of SCT and its application to Internet
use, in order to derive our research models.

2.4. A social cognitive perspective

Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

In a study of US citizens’ access to information and the resulting effects on political participation,
Bimber (2001) found that socioeconomic variables do affect access to information, but cannot
directly explain effects on participation. He suggests that ‘cognitive pathways’ need to be con-
sidered to understand the participatory effect of new media. SCT, particularly, has repeatedly
been applied to the adoption of new communication technology and the ability of users to produc-
tively use new media. It provides a suitable approach to obtaining a more differentiated under-
standing of the sociodemographic antecedents of online content creation.
SCT posits that environmental factors, personal factors, and behavior form a causal model of
‘triadic reciprocity’. Within this model (a) an individual’s environment influences personal dispo-
sitions (such as cognitions and affect), which in turn shape its choice of environment; (b) personal
dispositions influence behavior, which in turn influences these personal factors; and (c) behavior
affects the environment, which in turn impacts behavior (Bandura, 1977, 1986). This framework
has been applied to ICT use, showing that social and environmental factors – such as training and
ICT access – impact personal dispositions, which in turn affect use behavior (Ambrose & Chir-
avuri, 2010; Compeau & Higgins, 1995; Wei et al., 2011).
Basing an analysis of the participation divide on SCT allows for a consideration of the med-
iating role of cognitive factors in the effect of sociodemographics on use behavior. According to
SCT, these cognitive factors are affected by environmental influences and, in turn, significantly
affect behavior. Given this relationship, SCT provides a helpful explanation of why sociodemo-
graphic variables affect the use of new media: These variables can be associated with specific
environmental influences and thereby affect the development of personal dispositions (i.e. self-
efficacy or privacy concerns) and ultimately behavior. Whether or not a sociodemographic vari-
able signifies distinct environmental influences is largely dependent on exogenous factors, such as
economic (i.e. access), institutional (i.e. education), technological (i.e. affordances), or cultural
(i.e. discrimination) conditions. For example, the impact of gender on Internet use should
depend on the cultural context, as it may have bearing on women’s exposure to ICT relative to
that of men (Doney & Canon, 1997; Frenkel, 1990; Wei et al., 2011).
Foremost among the cognitive factors considered in SCT is self-efficacy, users’ perceptions,
or judgments of their own ability to perform a specific behavior (Bandura, 1977; Compeau &
Higgins, 1995). Self-efficacy has variously been shown to drive users’ willingness or ability to
use ICT (Compeau et al., 1999; Venkatesh & Bala, 2008; Venkatesh, Morris, Davis, & Davis,
2003). It has previously been considered in digital divide research, and has been shown to be
helpful in explaining the effect of sociodemographic variables on use behavior (Hsieh et al.,
2011; Wei et al., 2011). The literature provides for a number of conceptualizations of self-efficacy,
such as computer self-efficacy or Internet self-efficacy – depending on the ICT considered for use
(Compeau & Higgins, 1995; Compeau et al., 1999).
 Information, Communication & Society 7

While self-efficacy contributes to a willingness to use ICTs as well as an open, positive, or

playful attitude toward ICT, the reverse is the case for user anxiety or concerns. The more con-
cerns regarding an ICT, the less willing individuals are to use it (Compeau et al., 1999; Venkatesh
& Bala, 2008). One concern variously discussed in the information systems literature and shown
to inhibit Internet use is a user’s concern for privacy online (Smith, Dinev, & Xu, 2011). Both self-
efficacy and concerns have been shown to be affected by sociodemographic variables. In a
number of studies, younger, male, and higher SES users have reported higher levels of self-effi-
cacy and lower levels of ICT anxiety or concerns than their older, female, and lower SES counter-
parts (Gefen & Straub, 1997; Venkatesh & Morris, 2000; Wei et al., 2011).
Our study sets out to contribute to the current debate on the antecedents of online content cre-
ation by analyzing the mediating role of two cognitive factors: self-efficacy and privacy concerns.
Schradie (2011) acknowledges that a playful attitude toward ICT is a key driver of creative use,
while user concerns inhibit content creation. She calls for a consideration of the mediating effect
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

of these cognitive factors in participation divide research (Schradie, 2013). Previous studies have
found that use skills or perceived skills moderate some effects of sociodemographic variables on
content creation (Blank, 2013a; Correa, 2010; Hargittai & Walejko, 2008). Correa (2010) finds
that skills have no significant effect on use behavior when perceived competence is taken into
consideration. Self-efficacy goes beyond a mere estimation of use skills, as it conceptualizes a
user’s judgment of his or her ability to achieve certain goals given a specific skill set
(Bandura, 1986; Compeau et al., 1999).
By analyzing the mediating role of self-efficacy and privacy concerns, we hope to contribute
to the understanding of why certain sociodemographic variables affect specific types of online
content creation, while others do not.

2.5. Research models

We will show that the consideration of cognitive factors contributes to our understanding of online
content creation by comparing and analyzing three models. These models are based on Blank’s
(2013a) typology of online content creation. In all the three models, we consider age, gender,
and education as independent sociodemographic variables (cf. Blank, 2013a; Correa, 2010; Har-
gittai & Walejko, 2008; Schradie, 2011). As discussed in the previous segment, we also consider
two cognitive variables: self-efficacy and privacy concerns. Figures 1–3 show the competing
research models of this article. The first model most closely corresponds to the (full) regression
models used in previous studies. Here, the cognitive constructs are included as additional inde-
pendent variables.
Since we employ SEM as opposed to regression analyses, we are able to more closely
examine the mediating role of the two cognitive factors: Accordingly, the second model concep-
tualizes the cognitive constructs as mediators, as suggested by SCT as well as some previous
studies (Blank, 2013a; Schradie, 2013). In order to gauge the mediating role of the cognitive
factors, it – somewhat boldly – assumes that the effects of the sociodemographic variables are
fully mediated by the cognitive concepts. Considering this model will allow us to compare the
model fit and explanatory power of the unmediated and the fully mediate models.
Finally, the third, partially mediated model allows for examining the direct effects of the
sociodemographic variables. Since it not only takes the mediating effect of both the cognitive con-
cepts into consideration, but also allows for further effects not captured by these mediators, we
expect this model to display the highest goodness of fit – superior to the first and second models.
As to the mediating effects, we propose that age, gender, and education affect self-efficacy and
privacy concerns with younger, male, and highly educated users reporting higher self-efficacy and
lower privacy concerns (Gefen & Straub, 1997; Venkatesh & Morris, 2000; Wei et al., 2011). We
 8 C.P. Hoffmann et al.
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Figure 1. Only direct model.

further propose that self-efficacy positively affects the creation of all three forms of online
content, while privacy concerns negatively affect content creation. Self-efficacy has previously
been shown to reduce privacy concerns, so we will also take account of this effect (Compeau
et al., 1999). Finally, we will control for an effect of age and gender on education. The former
accounts for education expansion, that is, the fact that today, young people tend to attend

Figure 2. Fully mediated model.

 Information, Communication & Society 9
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Figure 3. Partially mediated model.

school longer and face a wider variety of educational options, while the latter attempts to capture
different educational paths between men and women.

3. Methodology
3.1. Data and measures
The survey sample was recruited from a pool of German Internet users demographically represen-
tative of the German adult population, provided by a leading market research institute. During
August 2013, 4089 users were invited to participate in the survey by e-mail. Participants were
offered a small monetary incentive. A total of 1488 users responded (response rate 36%). The
overall sample composition is summarized in Table 2. A gender, age, and regional distribution
composition of the sample representative of the German general population was ensured by defin-
ing quotas on these attributes. Income and marital status were not assessed in the survey, but
respondents were asked whether they had children and which occupational field they were in.
Table 2 displays the distribution of these variables.
As to choosing Germany as the context of our analysis, we recognize that online participation
is dependent on the social and cultural context (Calenda & Meijer, 2009; George, 2005). Yet, little
research has been directed at sociocultural contextual influences on online participation. Most
studies have focused on the US population with few country-by-country comparisons (cf.
Blank, 2013a; Correa, 2010; Hargittai & Walejko, 2008; Schradie, 2011). We find that
Germany does not differ much in overall Internet access and use from other Western countries
(EUROSTAT, 2013). In terms of civic engagement and political participation, Germany also
closely resembles other European countries: Online political participation in Germany is
limited to a small proportion of the population, whereas various forms of offline participation
are more prevalent (Emmer, Wolling, & Vowe, 2012). Social and entertainment-oriented uses
of the Internet enjoy greater popularity than online political participation (ARD/ZDF, 2013).
 10 C.P. Hoffmann et al.

Table 2. Demographic structure of the sample.

Variable Category N % Missings
Age 18–29 years 231 16.9
30–39 years 201 14.1
40–49 years 301 19.8
50–59 years 271 17.7
60 years and above 484 31.5
Total 1488 100 0
Gender Female 746 51.1
Male 742 48.9
Total 1488 100 0
Education Primary 195 13.2
Secondary 797 53.8
Tertiary 490 33.0
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Total 1482 100 6

Children Yes 873 59.8
No 587 40.2
Total 1460 100 28
Occupational Field In school 94 6.3
Agriculture/mining 7 0.5
Automobile/aviation/navigation 29 2.0
Construction/infrastructure 45 3.0
Consulting/law 38 2.6
ICT/electronic devices 87 5.9
Energy/environment/recycling 24 1.6
Finance/insurance/real estate 60 4.0
Health care/social care 116 7.8
Wholesale and retailing 83 5.6
Communication/media 8 0.5
Machine engineering 31 2.1
Medicine/pharma/chemistry 29 2.0
Food & drinks 12 0.8
Public service 179 12.1
Tourism/hospitality/culture/sports 33 2.2
Transport/logistics 38 2.6
Other 570 38.4
Total 1483 100 5

Compared those of the United States, German citizens tend to use the Internet for political pur-
poses somewhat less frequently (Köcher & Bruttel, 2011; Smith, 2013).
The questionnaire addressed the participants’ media consumption, Internet use, online partici-
pation, and several questions on offline civic engagement and political participation. We
measured participants’ online self-efficacy based on three items, each addressing their judgments
of their own ability to publish different forms of content online. This measure was based on pre-
vious measures and focused on online tools that allow for content creation (cf. Compeau et al.,
1999; Venkatesh & Bala, 2008; Venkatesh et al., 2003). Privacy concerns were measured with
three items from Malhotra, Kim, and Agarwal’s (2004) global information privacy concern
measure.
The measures of the three types of online content creation were derived from Blank’s (2013a)
typology. Skilled content creation encompasses publishing texts and comments on the Internet
and commenting on the writings of others. Social and entertainment content creation covers
the use of SNS and the posting of photos or videos online. Political content creation covers
 Information, Communication & Society 11

not only commenting on political issues, but also participating in political discussions online. Our
measure of political content creation, therefore, is somewhat broader than that proposed by Blank
(2013a). The wording of all items is reported in Appendix 1.

3.2. Methods
We relied on SEM to address the research questions. In contrast to normal regression analysis,
SEM can address indirect effects and latent variables. Furthermore, it takes into account measure-
ment errors in the specification of latent constructs and gives out global goodness-of-fit measures,
such as RMSEA (root mean square error of approximation), SRMR (standardized root mean
square residual), CFI (comparative fit index), and TLI (Tucker–Lewis index), to allow for
model comparisons. Previous studies on online content creation relied either on (stepwise) logis-
tic regression (Blank, 2013a; Hargittai & Walejko, 2008; Schradie, 2011) or linear regression
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

(Correa, 2010), complicating the test for indirect effects.

We used MPlus Statistical Software (Version 7) to carry out the analyses, relying on robust
maximum-likelihood estimation to account for non-normality and other sources of distortion,
such as heteroscedasticity and non-normal distribution of error terms (Byrne, 2012). The com-
plete measurement model of all latent constructs is reported in Appendix 2. It satisfies the necess-
ary conditions (Bollen, 1989; Netemeyer, Bearden, & Sharma, 2003); that is, it has convergent
and discriminant validity (Fornell & Larcker, 1981; see Appendix 3). The only exceptions are
the AVE (average variance extracted) of privacy concerns, which falls slightly short of the
threshold value, and the correlation between skilled and social and entertainment content. We
opted for retaining the privacy concerns measure with three items because it is derived from a
well-tested scale. Given the large (squared) correlation between skilled content on the one
hand and social and entertainment content on the other hand, discriminant validity cannot be
assumed for these two constructs. We chose to retain the three categories, though, because of con-
ceptual reasons and to replicate Blank’s (2013a) typology of online content.

4. Results and discussion

Overall, participants report only little enthusiasm for creating content online. Only about 10%
create skilled content often or very often. About half of the sample never publishes texts on
the Internet, or comments or discusses in online communities. For social and entertainment
content, these numbers are markedly higher. On average, users are active on SNS and online
media platforms several times a month. Yet, over three-thirds of the German online population
never or only rarely post pictures or videos. Finally, political content creation is even less
common than the other types considered. Both the average values and the relative frequencies
point to low participation rates.
Figure 4 shows the results of the first, direct model. This model most closely resembles the
analysis of previous studies, and, indeed, we find results closely in line with those reported by
Blank (2013a). We find that older users are less likely to publish skilled and social and entertain-
ment content, while there is no significant effect for political content – presumably because the
higher activity levels of younger users are offset by the higher political interest of older users.
While Blank (2013a) did not find any effect of gender on content production, we find that political
content is less likely to be produced by female users. The only direct effect of education on
content production is a negative effect on social and entertainment content production, replicating
Blank’s (2013a) finding in this regard. We find that self-efficacy drives all three types of content
creation, while privacy concerns reduce the production of social and entertainment and, surpris-
ingly, weakly increases that of political content, while not affecting skilled content creation.
 12 C.P. Hoffmann et al.
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Figure 4. Only direct model, significant effects only (thin line p < .05, fat line p < .001).

The fully mediated model (Figure 5) shows strong effects of self-efficacy on all three types of
content production and a negative effect of privacy concerns on social and entertainment content
production only. As to the effects of the sociodemographic variables on the cognitive factors, we
find that self-efficacy is markedly higher for younger users, and also significantly higher for male
and highly educated users. At the same time, older users report higher levels of privacy concerns.
These findings are in line with previous studies of self-efficacy (Gefen & Straub, 1997; Venkatesh
& Morris, 2000; Wei et al., 2011). In fact, the limited effect of privacy concerns on actual use
behavior has been observed in previous studies, too (Compeau et al., 1999). We also find that
the level of education is lower for older and female users. Overall, our findings confirm the prop-
osition derived from SCT that sociodemographic variables indicate distinct learning experiences
that affect ICT attitudes and use behavior.
Figure 6 shows the partially mediated model. This model presents the most complex and
refined description of the direct and indirect effects of sociodemographic variables on online
content production. When considering the mediating effect of the cognitive factors, we still
find significant direct effects of age and education: younger users are more likely and highly edu-
cated users are less likely to create social and entertainment and skilled content. We do not find a
direct effect of gender on content creation. Also, political content creation is not directly affected
by any of the sociodemographic variables. Self-efficacy, again, strongly and significantly drives
all three types of content creation, while privacy concerns negatively affect social and entertain-
ment content creation.
Again, privacy concerns are positively associated with political content creation. Since the
only sociodemographic antecedent of privacy concerns is age, we would again argue that age
is associated with political interest which drives political content creation. Another possible
explanation for the somewhat counterintuitive effect of privacy concerns might reside in the
 Information, Communication & Society 13
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Figure 5. Fully mediated model, significant effects only (thin line p < .05, fat line p < .001).

fact that privacy protection constitutes a politically contested topic; thereby, politically interested
users may be more critical in regard to their online privacy protection. Privacy concerns also show
a negative effect on social and entertainment content creation in the partially mediated model.
Again, we find that the effect of sociodemographic variables is clearly mediated by cognitive
factors, as suggested by SCT. We find that younger users report significantly higher levels of
online self-efficacy. Education also contributes to self-efficacy, with older and female users
reporting lower educational levels. Female users report significantly lower levels of self-efficacy,
which cannot be explained by differing educational levels alone.

Figure 6. Partially mediated model, significant effects only (thin line p < .05, fat line p < .001).
 14 C.P. Hoffmann et al.

Table 3. Goodness-of-fit model comparison.

Model 1: direct Model 2: indirect Model 3: direct and indirect
χ2
737 924 633
df 130 135 126
CFI 0.95 0.94 0.96
TLI 0.94 0.92 0.95
RMSEA 0.06 0.06 0.05
SRMR 0.04 0.05 0.04
AIC 69,925 70,123 69,818
BIC 70,333 70,503 70,249
Notes: N = 1488.
AIC, Akaike information criterion; BIC, Bayesian information criterion.
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Looking at the explained variances in all the three models, we find some noteworthy ten-
dencies. Model 3 exhibits the highest R 2 for social and entertainment and skilled content creation,
with Model 1 showing better results than Model 2. R 2 for political content creation is quite low,
overall, indicating that there are other important drivers for this form of participation not captured
by the models. Also, this value barely varies across models. R 2 for social and entertainment
content creation varies most across the models. These findings demonstrate that it is important
to consider both direct and indirect effects of sociodemographic variables on content creation.
We also find that the sociodemographic antecedents explain a substantial part in the variance
of online self-efficacy, a finding that marks the prevalence of a digital divide, even in 2013
and in a developed country like Germany.
Finally, Table 3 shows the goodness-of-fit values for the three models. Model 3 performs best
and exceeds the threshold value specified in the literature for SEM goodness-of-fit measures (Hu
& Bentler, 1999). Model 1 performs better than Model 2 with a similar number of degrees of
freedom. However, given that Model 2 is based on the assumption that the effects of sociodemo-
graphic variables on content creation are fully mediated by cognitive factors, it performs quite
well – almost as well as Model 1. This result shows the importance of taking indirect effects
into account when exploring the participation divide. It also speaks in favor of extending the
methodology applied to the issue beyond regression models alone.

5. Conclusion
5.1. Summary and implications
Our findings contribute to the current debate on the scope of the so-called participation divide by
highlighting the mediating role of cognitive factors, allowing for a more differentiated under-
standing of why and how sociodemographic variables affect online content creation. Derived
from SCT, our analysis focuses on two cognitive factors, in particular online self-efficacy and
privacy concerns.
In line with previous calls for a consideration of ‘cognitive pathways’ in the analysis of online
participation (Bimber, 2001), we find that the two analyzed cognitive factors significantly affect
online content creation. Online self-efficacy, especially, has a strong, positive impact on the cre-
ation of social, skilled, and political content. The effect of online privacy concerns is less pro-
nounced, showing a weak negative effect on the creation of social content, and a weak
positive effect on the creation of political content. The latter might be explained by privacy con-
cerns being more pronounced among older Internet users who are more interested in politics, or
with political interest increasing the awareness of online privacy risks. Overall, we find that the
 Information, Communication & Society 15

limited effect of privacy concerns on use behavior is in line with previous findings (Compeau
et al., 1999).
Analyzing the mediating role of the two cognitive factors does allow for a more nuanced
understanding of the participation divide: SCT implies that sociodemographic variables may indi-
cate environmental influences, such as access or training. These influences shape cognitive dis-
positions, which in turn affect use behavior. We find that online self-efficacy decreases with
age, increases with education, and is more pronounced among male users. These findings hold
even when controlling for the fact that older, female users exhibit lower levels of educational
attainment. Given the strong positive effect of online self-efficacy on all three forms of content
creation, our analysis contributes to a theoretical explanation for why previous studies found
younger, educated, and male users to be more active creators of online content (Correa, 2010;
Hargittai & Walejko, 2008; Schradie, 2011).
As our analysis differentiates direct and indirect effects of sociodemographic variables on
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

content creation, we are able to contribute to the current debate on the scope of the participation
divide: By basing our study on the differentiation of content creation suggested by Blank (2013a),
we find that established findings do apply even if types of content creation are differentiated –
counter to the suggestion that sociodemographic variables might only affect specific types of
content production. Self-efficacy is a key driver of all forms of content creation, and it clearly
differs by age, gender, and education. Thus, a participation divide exists for all types of
content creation.
Also, we are able to confirm a key finding of Blank (2013a): There is a significant negative
effect of education on the production of social and entertainment and skilled content, even after
taking age and indirect education effects into consideration. Thereby, education is in fact nega-
tively associated with some types of content creation. We interpret this finding as highly educated
users being less interested in online interactions, especially for social or entertainment purposes.
This effect holds even though highly educated users report higher self-efficacy, which generally
drives content production on the Internet.
Some important implications can be derived from the insights gained by considering the med-
iating role of cognitive factors, based on SCT: We do find evidence for persisting social inequal-
ities when it comes to online content creation (cf. Schradie, 2013). Age and education clearly
affect user self-efficacy, and thereby content creation. Ensuring access and training opportunities
might help ameliorate these differences as they have been shown to affect user self-efficacy. It is
noteworthy that gender also impacts self-efficacy. Additionally, we find a compounding indirect
effect with female users reporting lower educational levels. These findings indicate a need for
further examinations of gender differences in socialization experiences, especially as they
pertain to the use of and familiarity with new media.
Differentiating the forms of content creation reveals that previous research might have over-
estimated the scope of the participation divide by focusing heavily on political content creation
(Lutz et al., 2014). Yet younger, less educated users appear more prone to create and share
social and skilled content. Of course, our analysis sheds little light on the question of whether
there are distinct user groups, continuously engaging in distinct social domains – or whether
there may be a dynamic component at play, with experiences gained in one domain being trans-
ferred to another over time. In other words, could younger, less educated users gain use experi-
ence and increase their self-efficacy by creating social or skilled content, ultimately facilitating the
creation of other content, such as political? Of course, this question is related to a normative
evaluation of different types of content creation: Should we consider the creation of political
content more valuable or desirable than the creation of skilled or social content?
Finally, it is noteworthy that the predominant sociodemographic effect on content creation is
that of age – both a direct effect and one mediated through cognitive factors. Thereby, our analysis
 16 C.P. Hoffmann et al.

shows that the participation divide is largely caused by an age gap in online content creation. We
find that younger users are clearly more geared toward interactive, social, and entertainment uses
of the Internet, and they report significantly higher levels of online participation. This effect
remains strong even after taking self-efficacy and educational attainment into consideration.
We would conclude that a better understanding of the specific socio-technical socialization of
younger Internet users should provide a bigger contribution to our understanding of the partici-
pation divide than analyzing the effect of users’ SES.

5.2. Limitations and suggestions for future research

This research has some limitations, which limit its scope and provide avenues for future research.
First, the study is based on an online survey; participants were recruited through e-mail invita-
tions. Self-selection unavoidably affects user participation in these kinds of survey. We tried to
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

counter this effect by defining quotas on critical sociodemographic variables. Also, the chosen
method excludes the segment of the population without Internet access (in Germany about
15%). Thus, our results are only generalizable to the online population and not the overall popu-
lation. Future research could also consider offliners to compare their profiles with online partici-
pants and online nonparticipants.
Second, some authors have noted that online participation depends on the social and cultural
context (Calenda & Meijer, 2009; George, 2005). Our study was conducted among German Inter-
net users, which might have an effect on its findings. We find that the German online population is
very similar to that of other Western countries, both in access to the Internet, use frequency, and
preference for online participation. Our findings, accordingly, appear to be closely in line with
those derived from US or UK studies. Yet, more research into the sociocultural determinants
of online participation would certainly be helpful to facilitate cross-national comparative
analyses.
Third, the collected data only cover one point in time. Thus, inferences across time are not
possible and the issue of isolating different causal effects (e.g. of the cognitive constructs on
content creation) remains. Future research on online content creation, especially the migration
across various forms of online creation, could use panel designs to describe and explain
changes over time. Fourth, additional explanatory factors should be included in the research
model. Future studies might include broader indicators of users’ social and cultural background,
such as income and ethnicity.
Fifth, additional forms of online content could be considered in future research. We replicated
Blank’s (2013a) typology to add to a cumulative research agenda on online content creation.
However, other forms of content not originally considered by Blank (2013a) should not be neg-
lected. This includes health-related content or educational content. Sixth and finally, we had to
rely on self-reported data. Such data are subject to challenges such as memory bias and social
desirability. Online participation research should, therefore, combine different data sources,
including observational data.
Overall, this study contributes to the research of online content creation by solidifying its
theoretical basis, expanding its geographical scope, and advancing the applied methodology.
Yet, it also demonstrates that research into the participation divide is still relatively recent and pro-
vides ample opportunities for further investigation.

Disclosure statement
No potential conflict of interest was reported by the authors.
 Information, Communication & Society 17

Notes on contributors
Christian Pieter Hoffmann is an assistant professor at the Institute for Media and Communications Manage-
ment, University of St. Gallen, Switzerland, and a lecturer at the University of Applied Sciences in Business
Administration Zurich. His research is focused on online communication, social media, and political com-
munication. [email: christian.hoffmann@unisg.ch]
Christoph Lutz is a Ph.D. student in media and communication at the University of St. Gallen (Switzerland).
His dissertation focuses on online participation and his other research interests include social media in
science and public administration, online privacy and trust, and digital serendipity. [email: christoph.
lutz@unisg.ch]
Miriam Meckel is a professor for communication management and director of the Institute for Media and
Communications Management, University of St. Gallen, Switzerland. Her research focuses on online com-
munication, social media, corporate communications, and the transformation of journalism. [email: miriam.
meckel@unisg.ch]
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

References
Ambrose, P. J., & Chiravuri, A. (2010). A socio-cognitive interpretation of the potential effects of downsiz-
ing on software quality performance. Information Systems Journal, 20(3), 239–265.
ARD/ZDF. (2013). ARD/ZDF-Onlinestudie 2013. Retrieved August 19, 2014, from http://www.ard-zdf-
onlinestudie.de/index.php?id=392
Bakker, T. P., & de Vreese, C. H. (2011). Good news for the future? Young people, Internet use, and political
participation. Communication Research, 38(4), 451–470.
Bandura, A. (1977). Social learning theory. Englewood Cliffs, NJ: Prentice-Hall.
Bandura, A. (1986). Social foundations of thought and action. Englewood Cliffs, NJ: Prentice-Hall.
Best, S. J., & Krueger, B. S. (2005). Analyzing the representativeness of Internet political participation.
Political Behavior, 27(2), 183–216.
Bimber, B. (2001). Information and political engagement in America: The search for effects of information
technology at the individual level. Political Research Quarterly, 54(1), 53–67.
Blank, G. (2013a). Who creates content? Information, Communication & Society, 16(4), 590–612.
Blank, G. (2013b). Social stratification and content production: A response. Information, Communication &
Society, 16(6), 999–1004.
Blank, G., & Reisdorf, B. C. (2012). The participatory web. Information, Communication & Society, 15(5),
537–554.
Bollen, K. A. (1989). Structural equations with latent variables. Hoboken, NJ: Wiley Interscience.
Boulianne, S. (2009). Does Internet sse affect engagement? A meta-analysis of research. Political
Communication, 26(2), 193–211.
Bridges, F., Appel, L., & Grossklags, J. (2012). Young adults’ online participation behaviors: An exploratory
study of web 2.0 use for political engagement. Information Polity, 17(2), 163–176.
Byrne, B. (2012). Structural equation modeling with Mplus. New York, NY: Routledge.
Calenda, D., & Meijer, A. (2009). Young people, the Internet, and political participation. Information,
Communication & Society, 12(6), 879–898.
Compeau, D., Higgins, C. A., & Huff, S. (1999). Social cognitive theory and individual reactions to comput-
ing technology: A longitudinal study. MIS Quarterly, 23(2), 145–158.
Compeau, D. R., & Higgins, C. A. (1995). Application of social cognitive theory to training for computer
skills. Information Systems Research, 6(2), 118–143.
Correa, T. (2010). The participation divide among ‘online experts’: Experience, skills and psychological
factors as predictors of college students’ web content creation. Journal of Computer-Mediated
Communication, 16(1), 71–92.
Dahlgren, P. (2011). Young citizens and political participation: Online media and civic cultures. Taiwan
Journal of Democracy, 7(2), 11–25.
Di Gennaro, C., & Dutton, W. (2006). The Internet and the public: Online and offline political participation
in the United Kingdom. Parliamentary Affairs, 59(2), 299–313.
DiMaggio, P., Hargittai, E., Celeste, C., & Shafer, S. (2003). From unequal access to differentiated use: A
literature review and agenda for research on digital inequality. Princeton, NJ: Princeton University.
Doney, P. M., & Cannon, J. P. (1997). An examination of the nature of trust in buyer-seller relationships.
Journal of Marketing, 61(2), 35–51.
 18 C.P. Hoffmann et al.

Dutta-Bergman, M. J. (2006). Community participation and Internet use after September 11:
Complementarity in channel consumption. Journal of Computer-Mediated Communication, 11, 469–
484.
Emmer, M., Wolling, J., & Vowe, G. (2012). Changing political communication in Germany: Findings from
a longitudinal study on the influence of the internet on political information, discussion and the partici-
pation of citizens. Communications – European Journal of Communication Research, 37(3), 233–252.
Epstein, Y. M., Rosenberg, H. S., Grant, T. V., & Hemenway, N. (2002). Use of the Internet as the only outlet
for talking about infertility. Fertility and Sterility, 78(3), 507–514.
EUROSTAT. (2013). Information society statistics. Retrieved August 20, 2014, from http://epp.eurostat.ec.
europa.eu/portal/page/portal/information_society/data/database
Fornell, C., & Larcker, D. F. (1981). Evaluating structural equation models with unobservable variables and
measurement error. Journal of Marketing Research, 18(1), 39–50.
Frenkel, K. A. (1990). Women and computing. Communications of the ACM, 33(11), 34–46.
Gefen, D., & Straub, D. W. (1997). Gender differences in the perception and use of e-mail: An extension to
the technology acceptance model. MIS Quarterly, 21(4), 389–400.
George, C. (2005). The internet’s political impact and the penetration/participation paradox in Malaysia and
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

Singapore. Media, Culture & Society, 27(6), 903–920.

Gibson, R. K., Lusoli, W., & Ward, S. (2005). Online participation in the UK: Testing a ‘contextualised’
model of Internet effects. The British Journal of Politics and International Relations, 7(4), 561–583.
Gil de Zúñiga, H., Veenstra, A., Vraga, E., & Shah, D. (2010). Digital democracy: Reimagining pathways to
political participation. Journal of Information Technology & Politics, 7(1), 36–51.
Gui, M., & Argentin, G. (2011). Digital skills of internet natives: Different forms of digital literacy in a
random sample of northern Italian high school students. New Media & Society, 13(6), 963–980.
Hargittai, E. (2002). Second-level digital divide: Differences in people’s online skills. First Monday, 7(4).
Retrieved from http://firstmonday.org/article/view/942/864
Hargittai, E. (2010). Digital na(t)ives? Variation in Internet skills and uses among members of the net gen-
eration. Sociological Inquiry, 80(1), 92–113.
Hargittai, E., & Hinnant, A. (2008). Digital inequality: Differences in young adults’ use of the Internet.
Communication Research, 35(5), 602–621.
Hargittai, E., & Shafer, S. (2006). Differences in actual and perceived online skills: The role of gender. Social
Science Quarterly, 87(2), 432–448.
Hargittai, E., & Walejko, G. (2008). The participation divide: Content creation and sharing in the digital age.
Information, Communication & Society, 11(2), 239–256.
Helsper, E. (2010). Gendered internet use across generations and life stages. Communication Research, 37
(3), 352–374.
Hoffman, L. (2012). Participation or communication? An explication of political activity in the Internet age.
Journal of Information Technology & Politics, 9(3), 217–233.
Hsieh, J. J. P.-A., Rai, A., & Keil, M. (2011). Addressing digital inequality for the socioeconomically dis-
advantaged through government initiatives: Forms of capital that affect ICT utilization. Information
Systems Research, 22(2), 233–253.
Hu, L., & Bentler, P. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional
criteria versus new alternatives. Structural Equation Modeling – A Multidisciplinary Journal, 6(1),
1–55.
Jenkins, H. (2006). Confronting the challenges of participatory culture: Media education for the 21st
century. MacArthur Foundation Report. Retrieved from http://www.macfound.org/media/article_pdfs/
JENKINS_WHITE_PAPER.PDF
Jugert, P., Eckstein, K., Noack, P., Kuhn, A., & Benbow, A. (2013). Offline and online civic engagement
among adolescents and young adults from three ethnic groups. Journal of Youth and Adolescence, 42
(1), 123–135.
Köcher, R., & Bruttel, O. (2011). 1. Infosys-Studie. Social Media, IT & Society 2011 (p. 84). Retrieved from
http://www.infosys.com/de/newsroom/press-releases/documents/social-media-it-society2011.pdf
Li, N., & Kirkup, G. (2007). Gender and cultural differences in Internet use: A study of China and the UK.
Computers & Education, 48(2), 301–317.
Livingstone, S., & Bober, M. (2004). Taking up online opportunities? Children’s uses of the Internet for edu-
cation, communication and participation. E-Learning, 1(3), 395–419.
Livingstone, S., Bober, M., & Helsper, E. (2005). Active participation or just more information? Information,
Communication & Society, 8(3), 287–314.
 Information, Communication & Society 19

Lutz, C., Hoffmann, C. P., & Meckel, M. (2014). Beyond just politics: A systematic literature review of
online participation. First Monday, 19(7). Retrieved from http://firstmonday.org/ojs/index.php/fm/
article/view/5260/4094
Malhotra, N. K., Kim, S. S., & Agarwal, J. (2004). Internet users’ information privacy concerns (IUIPC): The
construct, the scale, and a causal model. Information Systems Research, 15(4), 336–355.
Moy, P., Manosevitch, E., Stamm, K., & Dunsmore, K. (2005). Linking dimensions of internet use and civic
engagement. Journalism and Mass Communication Quarterly, 82(3), 571–586.
Netemeyer, R. G., Bearden, W. O., & Sharma, S. (2003). Scaling procedures. Issues and applications.
London: Sage.
Ono, H., & Zavodny, M. (2003). Gender and the Internet. Social Science Quarterly, 84(1), 111–121.
Park, H. M., & Perry, J. L. (2008). Do campaign web sites really matter in electoral civic engagement?
Empirical evidence from the 2004 post-election Internet tracking survey. Social Science Computer
Review, 26(2), 190–212.
Pasek, J., More, E., & Romer, D. (2009). Realizing the social Internet? Online social networking meets
offline civic engagement. Journal of Information Technology & Politics, 6(3–4), 197–215.
Rice, R. E., & Fuller, R. P. (2013). Theoretical perspectives in the study of communication and the Internet,
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

2000–2009. In W. Dutton (Ed.), The Oxford handbook of internet studies (pp. 353–377). Oxford: Oxford
University Press.
Rojas, H., & Puig-i-Abril, E. (2009). Mobilizers mobilized: Information, expression, mobilization and par-
ticipation in the digital age. Journal of Computer-Mediated Communication, 14(4), 902–927.
Sandaunet, A.-G. (2008). The challenge of fitting in: Non-participation and withdrawal from an online self-
help group for breast cancer patients. Sociology of Health & Illness, 30(1), 131–144.
Schradie, J. (2011). The digital production gap: The digital divide and Web 2.0 collide. Poetics, 39(2),
145–168.
Schradie, J. (2012). The trend of class, race, and ethnicity in social media inequality. Information,
Communication & Society, 15(4), 555–571.
Schradie, J. (2013). The digital production gap in Great Britain. Information, Communication & Society, 16
(6), 989–998.
Shah, D. V., Kwak, N., & Holbert, R. L. (2001). ‘Connecting’ and ‘disconnecting’ with civic life: Patterns of
Internet use and the production of social capital. Political Communication, 18(2), 141–162.
Smith, A. (2013). Civic engagement in the digital age. Pew Internet American Life Project (p. 59). Retrieved
from http://pewinternet.org/Reports/2013/Civic-Engagement.aspx
Smith, H. J., Dinev, T., & Xu, H. (2011). Information privacy research: An interdisciplinary review. MIS
Quarterly, 35(4), 989–1015.
Towner, T., & Dulio, D. (2011). The web 2.0 election: Does the online medium matter? Journal of Political
Marketing, 10(1–2), 165–188.
Van Deursen, A., & Van Dijk, J. (2011). Internet skills and the digital divide. New Media & Society, 13(6),
893–911.
Van Dijk, J. (2005). Skills access. In J. Van Dijk (Ed.), The deepening divide: Inequality in the information
society (1st ed., pp. 71–94). London: Sage.
Van Dijk, J. (2006). Digital divide research, achievements and shortcomings. Poetics, 34(4–5), 221–235.
Venkatesh, V., & Bala, H. (2008). Technology acceptance model 3 and a research agenda on interventions.
Decision Sciences, 39(2), 273–315.
Venkatesh, V., & Morris, M. G. (2000). Why don’t men ever stop to ask for directions? Gender, social influ-
ence, and their role in technology acceptance and usage behavior. MIS Quarterly, 24(1), 115–139.
Venkatesh, V., Morris, M. G., Davis, G. B., & Davis, F. D. (2003). User acceptance of information technol-
ogy: Toward a unified view. MIS Quarterly, 27(3), 425–478.
Verba, S., Schlozman, K. L., & Brady, H. E. (1995). Voice and equality: Civic voluntarism in American poli-
tics. Cambridge, MA: Harvard University Press.
Vitak, J., Zube, P., Smock, A., Caleb, T., Carr, C., & Ellison, N. (2011). It’s complicated: Facebook users’
political participation in the 2008 election. Cyberpsychology, Behavior and Social Networking, 14(3),
107–114.
Wang, S.-I. (2007). Political use of the Internet, political attitudes and political participation. Asian Journal of
Communication, 17(4), 381–395.
Ward, S., Gibson, R., & Lusoli, W. (2003). Online participation and mobilisation in Britain: Hype, hope and
reality. Parliamentary Affairs, 56(4), 652–668.
Wei, K.-K., Teo, H.-H., Chan, H. C., & Tan, B. C. Y. (2011). Conceptualizing and testing a social cognitive
model of the digital divide. Information Systems Research, 22(1), 170–187.
 20 C.P. Hoffmann et al.

Woodly, D. (2007). New competencies in democratic communication? Blogs, agenda setting and political
participation. Public Choice, 134(1–2), 109–123.
Zhang, W., Johnson, T. J., Seltzer, T., & Bichard, S. L. (2010). The revolution will be networked: The influ-
ence of social networking sites on political attitudes and behavior. Social Science Computer Review, 28
(1), 75–92.
Zillien, N., & Hargittai, E. (2009). Digital distinction?: Status-specific types of Internet usage. Social Science
Quarterly, 90(2), 274–291.

Appendix 1. Questionnaire

Construct Item Wording (Scale)

Skilled Content Creation (SCC) SCC1 I comment on content that other people have published.
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

SCC2 I publish my own texts and comments on the Internet.

SCC3 I actively participate in discussions in online communities.
Social and Entertainment Content ECC1 Social network sites (e.g. Facebook, XING).
Creation (ECC) ECC2 Media platforms (e.g. Youtube, Flickr).
ECC3 I share on the Web photos and videos I have created.
Political Content Creation (PCC) PCC1 I like and share political content on the Internet.
PCC2 I publish commentaries about political topics on the Internet.
PCC3 I try to persuade others online to become politically active.
PCC4 I actively participate in a political online group or online
community
Privacy Concerns (PC) PC1 All things considered, the Internet could cause serious privacy
risks.
PC2 Compared to others, I am more sensitive about the way online
services handle my personal information.
PC3 To me, it is the most important thing to protect my privacy from
online services.
Online Self-efficacy (OSE) How well do you think you are able to …
OSE1 … publish information on a blog or on Twitter?
OSE2 … publish a video on the Internet (e.g. on Youtube)?
OSE3 … create or edit an article on Wikipedia?
 Information, Communication & Society 21

Appendix 2. Measurement model of the latent constructs

Construct Item Standardized loading R2 α CR AVE

Skilled Content Creation (SCC) SCC1 .89 .80 .89 .89 .74
SCC2 .88 .78
SCC3 .80 .64
Entertainment Content Creation (ECC) ECC1 .70 .48 .71 .79 .55
ECC2 .68 .46
ECC3 .84 .52
Political Content Creation PCC1 .77 .60 .91 .91 .72
PCC2 .93 .87
PCC3 .87 .75
PCC4 .82 .67
Privacy Concerns (PC) PCC1 .52 .27 .72 .78 0.47
PCC2 .75 .57
Downloaded by [Universitaet St Gallen] at 03:15 19 January 2015

PCC3 .76 .58

Online Self-efficacy (OSE) OSE1 .86 .74 .88 .88 .71
OSE2 .89 .79
OSE3 .77 .60
Criterion ≥ 0.5 ≥ 0.4 ≥ 0.7 ≥ 0.6 ≥ 0.5
Note: CR, composite reliability.

Appendix 3. Fornell–Larcker criteria of the latent constructs (discriminant validity)

No. of items AVE SCC ECC PCC PC

SCC 3 .74
ECC 3 .55 .83
PCC 4 .72 .47 .38
PC 3 .47 .02 .07 .00
OSE 3 .71 .30 .42 .14 .03
Note: Squared correlations between constructs.

View publication stats