Análisis de red del comportamiento de la audiencia en Internet: hacia una mejor comprensión del proceso de agenda setting
| Autor | Sílvia Majó-Vázquez |
| Cargo | PhD Candidate, Internet Interdisciplinary Institute (UOC) |
| Páginas | 61-74 |
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
www.uoc.edu/idp
Submission date: April 2015
Accepted date: May 2015
Published in: June 2015
Universitat Oberta de Catalunya
Abstract
By constructing the network of media audience, this study sheds light on the predominant modes of
exposure to online political information in Spain. Novelty data from a panel of thirty thousand individuals
is used for the research. The preliminary results bring evidences for reviewing the line of reasoning that
advocates for the prevailing fragmentation of the public sphere. More notably, the results contribute to
proving that a substantial level of audience concentration still remains in the web. The highest levels of
audience overlapping are found in those media outlets that are driving the media agenda in the offline
sphere. Therefore the study proffers evidence that the structure of the online public sphere might
guarantee the necessary shared informational experiences for a deliberative democracy.
The implications of the current networked audience behaviour for the study of the agenda setting process
are discussed along with the chances for a shared public agenda in Spanish society. Observational
methods and content analysis have been used in the study of the agenda setting process so far. However,
the current communication environment characterized by unlimited, decentralized and abundant
sources of political information prompts the application of new analytical approaches. Networks are at
the heart of online communication and network science allows for analyzing its structure. It provides the
ARTICLE
A Network Analysis of Online
Audience Behaviour:
Towards a Better Comprehension
of the Agenda Setting Process*
Sílvia Majó-Vázquez
PhD Candidate
Internet Interdisciplinary Institute (UOC)
Sílvia Majó-Vázquez
61
* This brief of research is part of an ongoing broader project conducted with professors Ana S. Cardenal and Sandra
González-Bailón. Jointly they are mapping the online media networks and audience flow in Spain. Additional
acknowledgement to Joshua Becker from DiMeNet group at Annenberg Penn for useful input to this preliminary
analysis
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
62
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
62
affordances to map and study audience aggregated behaviour when searching for political information.
In doing in so, it also unveils the mechanisms that might still guarantee a public agenda in the digital age.
Keywords
online audience behaviour, Social Network Analysis, modes of exposure, fragmentation, agenda setting
Topic
political communication, political science
Análisis de red del comportamiento de la audiencia
en Internet: hacia una mejor comprensión del proceso
de agenda setting
Resumen
Este estudio construye la red de audiencias de los medios de comunicación en España con el objetivo
de identificar los modos prevalentes de exposición a la información política en Internet. Para ello, se
utilizan datos inéditos procedentes de un panel de treinta mil individuos. Los resultados preliminares
que se obtienen aportan evidencias que sugieren una revisión de los argumentos a favor de que Internet
fragmenta la esfera pública. Es más, los resultados muestran que todavía existe un nivel substancialde
concentración de las audiencias en los medios de comunicación en Internet. Los niveles más altos decon-
centración se localizan en las webs de los medios que tienen un rol predominante en el procesode
agenda setting fuera de la red. En este sentido, el estudio presenta evidencias de que la estructura
de la esfera pública en Internet podría garantizar los niveles necesarios de información compartida por
los ciudadanos para el buen funcionamiento de una democracia deliberativa.
Asimismo, en este trabajo se analizan las consecuencias que puede tener el actual comportamiento de
los individuos que consumen información política en la red, para el estudio del proceso de agenda setting.
Además se discuten las posibilidades de que exista una agenda pública compartida a la luz de la estructura
de la actual red de audiencias compartidas. Hasta el momento, los análisis de datos observacionales
así como de contenido han sido los métodos predominantes en el estudio de la formación de la agenda
pública. Sin embargo, el entorno actual de comunicación, caracterizado por la abundancia de fuentes de
información descentralizadas y la ilimitada oferta de noticias políticas, apela a la utilización de nuevas
metodologías. Sin lugar a dudas, las redes conforman la base de la comunicación en línea y por lo tanto, el
análisis de redes ofrece las técnicas y herramientas necesarias para identificar el impacto de su estructura
en el comportamiento de la audiencia. Esta metodología permite estudiar sus movimientos durante la
búsqueda de información política y con ello, supone un avance para la identificación de los mecanismos
y estructuras que pueden seguir garantizando una agenda pública en la era digital.
Palabras clave
comportamiento del público en internet, análisis de redes sociales, formas de exposición, fragmentación,
agenda setting
Tema
comunicación política, ciencia política, análisis de redes
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
63
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
63
1. Audience behaviour
and networked public sphere
Mapping the network of media audiences offers a useful
lens in investigating to what extent there is a fragmented
online public sphere. The aggregate behaviour of people
consuming political information on the web is a proxy to
know if citizens have a range of common informational
experiences.
According to many prominent political and social theorist,
democracy depends extensively on an informed citizenry
regarding the most important problems in their societies
(Converse, 1964; Habermas, 1994; Katz, 1996; Rawls,
2009). This requirement implies that political information
is a vehicle for engagement in the democratic process and
citizens should enjoy access to civic space for public affairs
dialogue (Baum, 2012, p. 268).
Traditional media have been guaranteeing this shared public
sphere until now by filtering the great amount of information
available, producing political news and ultimately organizing
the public debate through the agenda setting process
(Lippmann, 1922; M. McCombs and Shaw, 1972; Norris,
2001). Their function has received many labels. Whereas
Sunstein named it social glue (2009), Baum referred to it as
a media commons (2012). Embedded in their reasoning is the
claim that citizens tend to share media definition of what is
important in a society as several recent investigations have
proved (Arsenault and Castells, 2006; Chaques-Bonafont et
al., 2015; Maxwell McCombs, 2013; Palau and Davesa, 2013).
Now some of those scholars claim that in the place of a
collective shared agenda fragmented and competing media
agendas have emerged (Althaus and Tewksbury, 2002; Shaw
and Hamm, 1997; Tewksbury, 2005) because everyday media
experience is becoming more individualized (Chaffee and
Metzger, 2009). People tend to not to consume similar
media diets but instead they have a “Daily me”. This is a
term coined by Negroponte (1995) that many other scholars
have echoed to warn against the trend towards a decline
in common experiences and a system of individualized
information filtering provided by the web. More recently,
Pariser (2011) also proposed the concept “filter bubble”
to define the power of the web to personalize information
and tailoring people’s media consumption. This scenario
though might compromise two general constitutional ideals,
deliberative democracy and public forum (Sunstein, 2009)
Nonetheless, evidences to prove that the Internet has caused
a more fragmented public sphere have been theoretically
ambiguous. Two lines of reasoning have discussed the potential
consequences of the Internet for the public sphere so far. The
earliest approaches to this debate contended that Internet
would democratize the public sphere (Negroponte, 1995;
Rheingold, 1994; Rogers, 2004) by increasing the number
of issues considered in the public debate and bypassing the
filtering function of the traditional media. On the contrary, a
second line of reasoning have argued that filtering affordances
brought by Internet as well as an unlimited number of
sources of information prompt rising trend toward not only
the end of mass audiences (Castells, 2009; Napoli, 2011) but
unlimited audience fragmentation (Baum and Groeling, 2008;
Baum, 2012; Joseph Turow, 1998). Such situation might be
problematic for the functioning of the democracy because
again, it potentially leaves people underinformed about
central issues facing a nation (Katz, 1996; Tewksbury, 2005).
All those preceding views lay out contradictory scenarios
and they have been feeding the aftermath theoretical
research in this field. However, not until recently have a
few researchers taken advantage of network analysis to
bring empirical evidence to assess the feasibility of these
theories (Ksiazek, 2011; Taneja and Wu, 2014; Webster and
Ksiazek, 2012; González-Bailón, 2009). The key aspect of
this research is that they do find a greater level of audience
fragmentation on the web, but in coexiste with a predominant
level of audience concentration. Their results, as Hindman
posed, seemed more the model of winner take-all patterns
where a small number of outlets dominate the web mirroring
patterns found in the traditional media (2008, p. 273).
This study contributes to this latter line of research by
providing novel empirical evidence based on the case of Spain,
which certainly has not been studied before. Our preliminary
results prove that audience concentration still remains on the
web. More noteworthy, by mapping the aggregated audience
behaviour and identifying traditional and new media outlets,
the study brings evidence that those media driving the agenda-
setting process offline are still at the core of the audience
flow. The user control afforded by the new communication
technology might not spell the end of a shared public sphere.
2. Data and Methods
To build the audience network of Spanish news media, this
study uses data from comScore measurement tool, Media
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
64
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
64
Metrix. This company provides digital analytics for 43
worldwide countries and it is the official source of record for
online measurement in Spain since 2011. In this country, they
have an online panel of 30,000 individuals, representative
of the Spanish population.1 It is worth noting that comScore
data is unique in several ways. Firstly, their data comes
from electronic meter instead of self-report audience
recall. They use biometric data to identify who is using
the tracked device and avoid gathering data from people
outside the panel. Secondly, this company has a unified
digital measurement system consisting on a panel-centric
hybrid solution that provides digital audience measurement
by bridging panel and server based approaches (comScore,
2013). Furthermore, to account for the site’s audience in the
case of the news media industry, comScore combines the
online data from its panelists with data from Estudio General
de Medios in Spain (AIMC, 2014). Finally, data is reported
monthly and includes socio-demographic information on
the panelist. Unfortunately though, Media Metrix solution
for Spain does not provide political leaning information
of their panel members. Thus this study will not measure
levels of cross-cutting exposure (Mutz and Martin, 2001) or
exposure to dissimilar political views. That said, among the
news media industry comScore is a widely renowned and
used source of information for the audience measurement.
Quite relevant, this is the first time that comScore data
for Spain is used in for the study of modes of exposure to
political information, at least to the best of our knowledge.
The total number of sites included in our sample is 113 and
the data for this study was collected on December 2014.
comScore requires a minimum of 16 panelists to visit a site
within a given month in order for the basic statistics to be
reported. To get the sample though, we have to take into
consideration several additional conditions. On one hand,
sites must provide political information either way: Producing
that information or aggregating or curating information
provided by other primary sources. Noteworthy is the fact
that, we also consider within our sample social networks
and infotainment (Brants, 1998), online media outlets. The
frontiers between entertainment and information blur
offline as well as online and infotainment outlets have a
growing role in the media diets of citizens when seeking
for political information (Bennett, 1992; Edelman, 1988;
Thussu, 2008).
On the other hand, we know that the distribution of
audiences in internet are shaped in the a form of a long
tail (Albert et al., 1999; Anderson, 2006; Elberse, 2008;
Webster and Ksiazek, 2012). Consequently, if one just brings
into the sample the head of the distribution, which equals
to those media outlets that concentrate the majority of the
audience, we would not capture, if they do exist, the levels
of audience overlapping among all type of media outlets
reporting about the Spanish public sphere. On the contrary,
we might only see firstly, substantial levels of audience
concentration corresponding to those traditional media that
already have great levels of visibility offline; and secondly,
few big audience hubs corresponding to those nodes of
popular social networks. In this vein, and not surprisingly,
previous analysis reveals that the richer the organization
publishing a site is, the more people flock to it (González-
Bailón, 2009; Napoli, 2011). Hence aiming to study audience
overlapping among media outlets to evaluate the level of
fragmentation of public sphere, we must get a sample
that includes media located at the tail of the distribution
too. To meet this purpose, we add to the sample top read
media outlets categorized under the label “new media” by
comScore (ie, niche news media sites, blogs and aggregators).
Furthermore, we included traditional news media sites that
reach a minimum of 0,5% of the total Internet audience in
Spain2 (Ksiazek, 2011; Webster and Ksiazek, 2012). As a result
of these conditions, the sample studied largely represents
the most widely visited traditional as well as new media sites.
However, to test the accuracy of comScore data in providing
the ranking of the most visited media websites in Spain,
we also used Alexa.com (Alexa Internet, 2015). Several
researches have relied on Alexa data to obtain traffic
rankings of different Internet properties by countries (see
Ennew et al., 2005; Price and Grann, 2012; Reay et al.,
2013; Wu and Ackland, 2014). Unfortunately, Alexa’s open
information is limited to the first 500 most visited entities
per country. Among them, we found 44 media outlets and
by running a correlation we compared their position in Alexa
ranking with the one they occupy in comScore ranking. The
calculation yields a pretty high score of 0.906, which means
the two different sources provide almost identical audience
measures. For those entities beyond the first 44 positions in
the ranking of most visited media in Spain, we solely relied
on the data provided from comScore.
1. Data is weighted using enumeration study data to ensure it is representative of the overall Spanish online population.
2. According to comScore Spanish Internet audience is 26.601.217 million people.
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
65
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
65
These previous criteria yield a final sample of 113 sites
including traditional news media sites – television, radio,
newspaper or news agency- along with new media sites
–blogs, social networks, aggregators, online news media (see
the Annex 1 for Table 1 to find the complete list of the media
and their reach audience).
Building on the empirical framework from Ksiazek (2011)
we apply a network analytical approach to understand
online audience behaviour among the sites of our sample.
Afterwards, we examine the structural characteristics of the
network that audience forms. As said before, we argue that
audience behaviour accounts for the attention that each
media outlet receives and hence, it captures the informational
experiences of people. Likewise, relaying on the abundant
scientific evidences that started with the pioneering work
by M. McCombs and Shaw (1972), we assume that the public
agenda tends to match the media agenda (Arsenault and
Castells, 2006; Bonafont and Palau, 2011; Chaques-Bonafont
et al., 2015; Jones and Baumgartner, 2005). Thus, this
analysis can shed light on the debate of whether or not there
is a shared public agenda in the online sphere by assessing
the audience behaviour -base on audience overlapping- and
identifying,
if it does exist, common informational experiences.
3
Noteworthy that the nodes in this network are media sites
and a link is defined when the observed audience duplication
between two outlets differs from the expected duplication
between those two outlets. In that situation a tie is considered.
To further illustrate, the matrix for this network is directed and
binary. Thus 1 indicates the existence of a tie between two nodes.
Consequently, a tie primarily shows that two media outlets
share a portion of their audience. Besides, it defines that the
audience of outlet i is likely to attend the audience of outlet j as
compared with the overall audience. Figure 1 offers an example
of this latest explanation. There, we can see that the audience
of ElPais.es is also attending the information reported on
Eldiario.es and hence a tie is send from the former to the latter.
To identify this relationship, we used the proposed
measurement system by Ksiazek (2011), more precisely,
the measure named Deviation-from-Random-Duplication
(DfRD). This measure helps to rule non-significant
connections out of our sample. As previous research has
proved there is always room for random ties to arise and
hence it is necessary to apply a method in order to preserve
only those nodes that are significant. In this case, firstly, we
use primary duplication between every pair of nodes within
our sample. That is the percentage of the audience of outlet i
that it is also exposed to outlet j. It is a weighted and directed
measure that informs about the degree to which 2 media
outlets share audience members. Secondly, we calculate
the expected duplication among those media outlets using
the reach percentage of each outlet in any given pair and
3. Internet is a large and rapidly growing portion of Spanish media diet. Currently 34% of the population uses it to find information about politics
and society almost every day, according to data collected from Centro de Investigaciones Sociológica (see www.cis.es for more information).
Yet, there is still a 52% of the population that never uses the web for political information purposes. Differences across ages and different
levels of education are broader. Thus the evidences brought by this study apply to political interested citizens who get their media diet online.
Fig. 3.1. Network relationship between media outlets
ElPais.es ElDiario.es
New Media
Traditional Media
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
66
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
66
computing the formula by Ksiazek (2011). Finally, we subtract
the expected duplication from the observed.
4
When the
result yields a positive value, it indicates a non-random
tendency to attend to a given outlet. Thus a tie is defined.
On the contrary, a negative value is considered a tendency
to avoid a given outlet. We worked out the DfRD for all
13,110 possible pairs within the sample and we obtained
4,570 significant connections. Finally, we dichotomized our
matrix
5
to map the network of audiences and compute the
descriptive measures that we explain following.
3. Results
This research brief is part of an extensive study, currently
underway. Results presented are therefore preliminary
and solely focused on the structural characteristics of the
Spanish online audience network. Assessing the structure
of a network helps to reduce its complexity, thereby giving
us primary properties that matter for understanding modes
of exposure to political information.
4. The logic behind subtracting expected from observed, is that when an observed value is much greater than the expected value (i.e. the
duplication is much greater than expected by random chance), the calculation will yield a large, positive DfRD value. Conversely, when the
observed is only slightly greater than the expected, it will result in yield a small value; and if the observed is less than the expected, it will
produce a negative value (Ksiazek, 2011, p. 242)
5. Before dichotomizing the matrix, all loops have been removed to avoid ties in which both ends connect to a single node (Kolaczyk and
Csárdi, 2014).
Notes: This is a directed network that represents the overlapping audience behaviour among 113 Spanish media outlets. A tie is defined as
indicating a non-random tendency in the attendance of a given outlet. Nevertheless, there is no tie among a pair of media when the audience of
i tend to avoid j. As a result 4,570 connections or ties are defined. We use the Fruchterman and Reingold layout method to visualize the graph.
It is a spring-embedder method of graph drawing (Kolaczyk and Csárdi, 2014).
Fig. 4.1. Spanish Audience Network
Spanish Media Outlets
Ties
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
67
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
67
To this end, we construct the network of the Spanish media
audience (see Figure 2) and examine its centralization
degree. It is a general method for calculating a graph-level
centrality score based on node-level centrality measure
(Kolaczyk and Csárdi, 2014). According to Freeman’s formula
(1979), network centralization varies between 0 and 1. A
network with a level of centralization 1 means that one
node completely dominates the network (Freeman, 1979,
p. 228). That node is connected to each one of the others,
and it is also connected to all of the others. This is the most
centralized and unequal type of network with a shape of a
perfect “star”
6
(Hanneman and Riddle, 2005).
In the case of our network though, its centralization
degree measures the level of inequality or variance in the
overlapping audience distribution between the media outlets.
The degree of centralization of the audience network is 0.33.
Not surprisingly, in a competitive media market as it is the
case of Spain, there is not a system of perfect concentration
of audiences –the star pattern network. But certainly, the
degree of centralization of the network proves that there is
a substantial level of audience concentration and it comes
from a few nodes situated at the core of the network. To
further support this, we compute the in-degree and out-
degree distributions of nodes (see the Annex 1 for Figure
A1). The analysis supports this claim by showing that the
network studied is bimodal. In short, a small group of media
outlets (n=42) placed at the core of the network, receives the
greatest levels of online audience attention. Consequently
the largest number of ties leads to this small group of nodes.
Furthermore, we examine the cohesion of the network. This
measure is extremely consequential in understanding the
audience flow due to the fact that it identifies differences
in the level of connectivity in a network. In our case, the
small group of media receiving the greatest levels of
online audience forms the strongly connected component
of the network. The centralization degree of this strongly
connected component is .003, which supports the claim
that the levels of audience overlapping among its media
outlets are extremely high (see on Figure 3 the numerical
identification of the nodes in the main component and see
for Annex 1 Table 2 for their label identification and the
percentage of audience reach).
6. A star network has a level of centralization 1 and represents the maximum inequality in the degree distribution.
Fig. 4.2. New Media and Traditional Media in the Spanish Audience Network and main component
New Media
Traditional Media
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
68
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
68
The analysis of connectivity also detects that the network
is weakly connected.
7
Thus audiences do not only duplicate
among the nodes on the main component, but rather weak
ties also facilitate overlapping audiences between nodes
beyond this group. This type of ties bridge the gap between
the most visited media and small or niche online media in
the network (see them in the periphery of Figure 3) and in
doing so reduce the level of fragmentation in the online
public sphere (Prior, 2008) by providing more opportunities
for shared informational experiences.
In congruence with previous research, by building the
audience network and identifying its main component, we
can see that several media outlets dominating the online
audience flow are those which also receive the greatest
levels of audience in the offline sphere (Napoli, 2011). This
is the case of major newspapers such as El País, El Mundo,
ABC or La Vanguardia. Noteworthy scholars have focused
their attention on newspapers when assessing the agenda
setting process. They play a leading role above all types
of media outlets in organizing and prioritizing issues for
the public debate (Baumgartner and Chaques-Bonafont,
2015; M. E. McCombs and Shaw, 1972; Palau and Davesa,
2013; Vargo et al., 2014). In the case of Spain, El País and
El Mundo are considered the main drivers of the agenda
setting process by the most comprehensive study in the
field
8
and they are also at the core of audience flow in our
audience network.
4. Discussion
This study is anchored in the open debate about the future
of the public sphere in the digital age. The preliminary and
descriptive results presented allow for debunking of claims
that equate fragmentation of audiences with nonexistence
of shared informational experiences and consequently
with the demis of the public sphere. The evidences prove
that concentration and fragmentation coexist in the
online sphere. There are substantial levels of audience
overlapping among media outlets. We have found a core
of media that dominates the audience flow. Besides, the
patterns of audience overlapping detected show that people
simultaneously seek information on niche and small media.
Equally important, we can find among the media outlets,
receiving the greatest amount of audience flow and audience
overlapping, the traditional media that play a major role in
organizing the public agenda in the offline sphere.
Networks are used in many branches of science as a way
of representing the connection patterns between the
components of complex systems (Newman, 2012). Here we
have applied this network analytical approach to the study of
audience behaviour. However, we have limited the analysis to
structural characteristics of the network. The existence of an
online public sphere is not proven by the descriptive analysis
reported here, of course, but the evidences presented are
in line with the condition that must exist for common
informational experiences to take place in a fragmented
media environment. To further investigate our claims, we
will apply additional analysis to disentangle the drivers of
this network audience behaviour in our future research.
7. For a directed graph, our case study, two variations of the notion connectedness are possible. A network is strongly connected when every
node is reachable form every other by a direct tie. A network is weakly connected if the underlying graph (the result of stripping away the
tail and the head of the ties) is connected (Kolaczyk and Csárdi, 2014).
8. Spanish Policy Agenda Project .ub.edu/spanishpolicyagendas/>.
5. Bibliography
AIMC. (2014). Estudio General de Medios, resumen abril 2013 a marzo 2014. Retrieved from http://www.
aimc.es/
ALBERT, R.; JEONG, H.; BARABÁSI, A.-L. (1999). “Internet: Diameter of the world-wide web”. Nature, vol.
401, no. 6749, pp. 130–131. dx.doi.org/10.1038/43601>
ALEXA INTERNET, I. (2015). Alexa. Top Sites in Spain. [Accessed: 20 January 2015]
com/topsites/countries/ES>
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
69
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
69
ALTHAUS, S. L.; TEWKSBURY, D. (2002). “Agenda Setting and the “New” News: Patterns of Issue Importance
Among Readers of the Paper and Online Versions of the New York Times”. Communication Research,
vol. 29, no. 2, pp. 180–207. doi.org/10.1177/0093650202029002004>
ANDERSON, C. (2006). The long tail: Why the future of business is selling less of more. Hyperion.
ARSENAULT, A.; CASTELLS, M. (2006). “Conquering the minds, conquering Iraq: The social production
of misinformation in the United States--a case study”. Information, Communication & Society, vol.
9, no. 3, pp. 284–307. dx.doi.org/10.1080/13691180600751256>
BAUM, M. (2012). “Media, Public Opinion, and Presidential Leadership”. In: A. J. BERINSKY (ed.). New
Directions in Public Opinion (pp. 258–270). Routledge.
BAUM, M. A.; GROELING, T. (2008). “New Media and the Polarization of American Political Discourse”.
Political Communication, vol. 25, no. 4, pp. 345–365.
BAUMGARTNER, F. R.; CHAQUES-BONAFONT, L. (2015). “All News is Bad News : Newspaper Coverage of
Political Parties in Spain News Coverage in a Partisan Environment”. Political Communication, vol.
33, no. 2. dx.doi.org/10.1080/10584609.2014.919974>
BENNETT, W. L. (1992). The governing crisis: Media, money, and marketing in American elections. Palgrave
Macmillan.
BONAFONT, L. C.; PALAU, A. M. (2011). “Assessing the Responsiveness of Spanish Policymakers to the
Priorities of their Citizens”. West European Politics, vol. 34, no. 4, pp. 706–730.
80/01402382.2011.572388>
BRANTS, K. (1998). “Who’s afraid of infotainment?” European Journal of Communication, vol. 13, no. 3,
pp. 315–335. g/10.1177/0267323198013003002>
CASTELLS, M. (2009). “The Culture of Real Virtuality: The Integration of Electronic Communication, the
End of the Mass Audience, and the Rise of Interactive Networks”. In: The Rise of Interactive Networks (p.
406). Oxford, UK: WILEY-BLACKWELL PUBLISHING, INC. g/10.1002/9781444319514.ch5>
CHAFFEE, S. H.; METZGER, M. J. (2009). “The end of mass communication?” Mass Communication &
Society, vol. 4, no. 4, pp. 365–379.
CHAQUES-BONAFONT, L.; PALAU, A., BAUMGARTNER, F. R. (2015). Agenda Dynamics in Spain. Palgrave
Macmillan.
Comscore, I. (2013). MMX User Guide 2013. comScore, Inc.
Analytics/Media-Metrix>
CONVERSE, P. E. (1964). The nature of believe system in mass public. In: D. APTER (ed.). Ideology and
Discontent.
EDELMAN, M. J. (1988). Constructing the political spectacle. University of Chicago Press.
ELBERSE, A. (2008). “Should you invest in the long tail?” Harvard Business Review, vol. 86,no. 7/8, p. 88.
ENNEW, C.; LOCKETT, A.; BLACKMAN, I. [et al.] (2005). “Competition in Internet retail markets: The
impact of links on web site traffic”. Long Range Planning, vol. 38, no. 4, pp. 359–372.
org/10.1016/j.lrp.2005.04.009>
FREEMAN, L. C. (1979). “Centrality in social networks conceptual clarification”. Social Networks, 1(3), pp.
215–239. org/10.1016/0378-8733(78)90021-7>
GONZÁLEZ-BAILÓN, S. (2009). “Traps on the Web”. Information, Communication & Society, vol. 12, no.
8, pp. 1149–1173. ttp://doi.org/10.1080/13691180902767265>
HABERMAS, J. (1994). “Three Normative Models of Democracy”. Costellations, vol. 1, no. 1, pp. 1–10.
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
70
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
70
HANNEMAN, R. A.; RIDDLE, M. (2005). Introduction to Social Network Methods.
edu/~hanneman/nettext/C10_Centrality.html>
HINDMAN, M. (2008). “What is the online public sphere good for?” In: J. TUROW and L. TSUI (eds.). The
hyperlinked society: questioning connections in the Digital Age (pp. 268–288). Michigan: University
of Michigan Press Ann Arbor.
JONES, B. D.; Baumgartner, F. R. (2005). The Politics of Attention How Government Prioritizes Problems.
Chicago: University of Chicago Press.
KATZ, E. (1996). “And Deliver Us from Segmentation”. Annals of the American Academy of Political and
Social Science, no. 546, pp. 22–33.
0002716296546001003>
KOLACZYK, E. D.; CSÁRDI, G. (2014). Statistical analysis of network data with R (Vol. 65). Springer.
dx.doi.org/10.1007/978-1-4939-0983-4>
KSIAZEK, T. B. (2011). “A Network Analytic Approach to Understanding Cross-Platform Audience Behavior”.
Journal of Media Economics, vol. 24, no. 4, pp. 237–251.
LIPPMANN, W. (1922). Public opinion. New York: Macmillan.
McCOMBS, M. (2013). Setting the agenda: The mass media and public opinion. John Wiley & Sons.
McCOMBS, M. E.; SHAW, D. L. (1972). “The Agenda-Setting Function of Mass Media”. The Public Opinion
Quarterly, vol. 36, no. 2, pp. 176–187.
dx.doi.org/10.1086/267990>
McCOMBS, M.; SHAW, D. (1972). “The agenda-setting function of mass media”. Public Opinion Quarterly,
no. 36, pp. 176–187. < http://dx.doi.org/10.1086/267990>
MUTZ, D. C.; MARTIN, P. S. (2001). “Facilitating communication across lines of political difference: The
role of mass media”. American Political Science Review, vol. 95, no. 1, pp. 97–114.
NAPOLI, P. M. (2011). Audience evolution: New technologies and the transformation of media audiences.
Columbia University Press.
NEGROPONTE, N. (1995). Being Digital. New York: Knopf.
NEWMAN, M. E. J. (2012). “Communities, modules and large-scale structure in networks”. Nature Physics,
vol. 8, no. 1, pp. 25–31. g/10.1086/267990>
NORRIS, P. (2001). “El impacto de las comunicaciones políticas en las democracias post-industriales”.
Revista Española de Ciencia Politica, no. 4, pp. 7–33.
PALAU, A. M.; DAVESA, F. (2013). “The Impact of Media Coverage of Corruption on Spanish Public Opinion”.
Revista Española de Investigaciones Sociológicas, no. 144, pp. 97–124.
org/10.5477/cis/reis.144.97>
PARISER, E. (2011). The filter bubble: How the new personalized web is changing what we read and how
we think. Penguin.
PRICE, A.; GRANN, V. R. (2012). “Portrayal of complementary and alternative medicine for cancer by
top online news sites”. The Journal of Alternative and Complementary Medicine, vol. 18, no. 5, pp.
487–493. dx.doi.org/10.1089/acm.2011.0110>
PRIOR, M. (2008). “Are hyperlinks “weak ties”?” In: J. TUROW and L. TSUI (eds.). The hyperlinked
society: questioning connections in the Digital Age (pp. 227–249). Michigan: University of Michigan
Press Ann Arbor.
RAWLS, J. (2009). A theory of justice. Harvard university press.
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
71
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
71
REAY, I.; BEATTY, P.; DICK, S., Miller, J. (2013). “Privacy policies and national culture on the internet”.
Information Systems Frontiers, 15(2), pp. 279–292.
RHEINGOLD, H. (1994). The virtual community. London: Minerva.
ROGERS, R. (2004). Information politics on the Web. Cambridge, Massachusetts: MIT Press.
SHAW, D. L.; Hamm, B. J. (1997). “Agendas for a public union or for private communities? How individuals
are using media to reshape American society”. Communication and Democracy, pp. 209–230.
SUNSTEIN, C. R. (2009). Republic. com 2.0 (second). Princeton University Press.
TANEJA, H.; WU, A. X. (2014). “Does the Great Firewall really isolate the Chinese? Integrating access
blockage with cultural factors to explain Web user behavior”. The Information Society, vol. 30, no.
5, pp. 297–309.
TEWKSBURY, D. (2005). “Seeds of Audience Fragmentation: Specialization in the Use of Online News
Sites”. The. J. Broad. & Elec. Media, no. 49, p. 332.
THUSSU, D. K. (2008). News as entertainment: The rise of global infotainment. Sage.
TUROW, J. (1998). Breaking up America: Advertisers and the new media world. University of Chicago Press.
VARGO, C. J.; GUO, L.; McCOMBS, M. [et al.] (2014). “Network issue agendas on Twitter during the 2012
US presidential election”. Journal of Communication, vol. 64, no. 2, pp. 296–316.
WEBSTER, J. G., KSIAZEK, T. B. (2012). “The Dynamics of Audience Fragmentation: Public Attention in
an Age of Digital Media”. Journal of Communication, vol. 62, no. 1, pp. 39–56. org/10.1111
/j.1460-2466.2011.01616.x>
WU, L.; ACKLAND, R. (2014). “How Web 1.0 fails: the mismatch between hyperlinks and clickstreams”.
Social Network Analysis and Mining, vol. 4, no. 1, pp. 1–7.
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
72
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
72
6. Appendix
TABLE A7.1. Reach of Spanish media sites studied
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
73
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
73
TABLE A7.2. List and reach of Spanish media sites of network main component
Figure A1. In-degree and out-degree of the audience network
020 40 60 80 100 1 20
degree nodes mode in
Frequency
0 10 20 30 40 50 60 70
Node In-degree
28 30 32 34 36 38 40 42
degree nodes mode out
0 10 20 30 40 50
Frequency
Node Out-degree
IDP no 20 (June, 2015) I ISSN 1699-8154 Journal promoted by the Law and Political Science Department
Eloi PuigEloi Puig
Jose R. Agustina
www.uoc.edu/idp
Universitat Oberta de Catalunya
74
A Network Analysis of Online Audience Behaviour…
Eloi Puig
Eloi Puig
Sílvia Majó-Vázquez
74
Recommended citation
MAJÓ-VÁZQUEZ, Sílvia (2015). “A Network Analysis of Online Audience Behaviour: Towards a Better
Comprehension of the Agenda Setting Process”. IDP. Revista de Internet, Derecho y Política. No. 20,
pp. 61-74. UOC [Accessed: dd/mm/yy]
The texts published in this journal, unless otherwise indicated, are subject to a Creative
Commons Attribution-NoDerivativeWorks 3.0 Spain licence. They may be copied, distributed
and broadcast provided that the author, the journal and the institution that publishes
them (IDP Revista de Internet, Derecho y Política; UOC)are cited. Derivative works are not
permitted. The full licence can be consulted onhttp://creativecommons.org/licenses/by-
nd/3.0/es/deed.en.
About the author
Sílvia Majó-Vázquez
majosilvia@uoc.edu
PhD Candidate
IN3, Internet Interdisciplinary Institute (UOC)
http://in3.uoc.edu/opencms_portalin3/opencms/es/investigadors/list/majo_vazquez_silvia
Av. Carl Friedrich Gauss, 5
Parc Mediterrani de la Tecnologia
08860 Castelldefels
(Barcelona)
