Content uploaded by Patricia Delponti
Author content
All content in this area was uploaded by Patricia Delponti on May 24, 2023
Content may be subject to copyright.
Full Terms & Conditions of access and use can be found at
https://www.tandfonline.com/action/journalInformation?journalCode=witp20
Journal of Information Technology & Politics
ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/witp20
Can information literacy increase political
accountability? Linking information evaluation
with obstinate partisanship via social media
political homophily
Alberto Ardèvol-Abreu, Carmen Costa-Sánchez & Patricia Delponti
To cite this article: Alberto Ardèvol-Abreu, Carmen Costa-Sánchez & Patricia Delponti (2023):
Can information literacy increase political accountability? Linking information evaluation with
obstinate partisanship via social media political homophily, Journal of Information Technology
& Politics, DOI: 10.1080/19331681.2023.2173699
To link to this article: https://doi.org/10.1080/19331681.2023.2173699
Published online: 20 Feb 2023.
Submit your article to this journal
View related articles
View Crossmark data
Can information literacy increase political accountability? Linking information
evaluation with obstinate partisanship via social media political homophily
Alberto Ardèvol-Abreu , Carmen Costa-Sánchez , and Patricia Delponti
ABSTRACT
Scholars have recently sought to explain why some voters remain loyal to “their” parties or
candidates “no matter what they do” – an attitude that has been labeled as obstinate partisanship
(OP) and limits electoral accountability. We argue that the development of OP may have to do, at
least partly, with people’s (in)ability to critically evaluate political information and their tendency to
isolate themselves in online bubbles of congenial information and interactions. Building on this
framework, we use two-wave panel survey data (N
W1
= 1,259; N
W2
= 982) to explore direct and
indirect associations between information evaluation – a key component of information literacy –
and OP. We nd that information evaluation is negatively associated with OP in cross-sectional and
autoregressive regression models. Analyses also support an indirect relationship between informa-
tion evaluation and OP through political homophily in social media and online. We discuss possible
implications for democracy and information literacy initiatives.
KEYWORDS
Information evaluation;
political homophily in social
media and online; obstinate
partisanship; political
accountability; information
literacy; social media use
Introduction
In representative democracies, informed citizens are
expected to control the performance of politicians
and parties and punish them when they do not
respond to voters’ demands (Adserà, Boix, &
Payne, 2003; Snyder & Strömberg, 2010).
Regretfully, electoral accountability is not always
effective and there are many instances where the
evidence of poor performance and even corrupt
practices goes relatively unpunished (see Cobb &
Taylor, 2015; Eggers, 2014; Fernández-Vázquez,
Barberá, & Rivero, 2016). Among the several reasons
why (some) voters fail to hold political actors
accountable for their actions is a blind loyalty to
their preferred party, an attitude that previous litera-
ture has labeled as obstinate partisanship (OP)
(Ardèvol-Abreu & Gil de Zúñiga, 2020). Obstinate
partisans (OPs) remain loyal to “their” party “no
matter what they do,” “both when they are doing
well and not so well,” and “even when they make a
mistake” (Ardèvol-Abreu & Gil de Zúñiga, 2020, p.
330; see also Goyanes, Borah, & Gil de Zúñiga, 2021).
The reasons why some voters become OPs and
support their party at any cost are less clear. The
seminal study of OP proposes that this attitude
results, at least in part, from reduced exposure to
discussion disagreement (with discussion network
size and discussion disagreement being negatively
associated with OP). The related literature on the
electoral impact of scandals suggests some alterna-
tive explanations. Thus, the tradeoff hypothesis
assumes that voters may overlook some political
misdeeds if they perceive other advantages that out-
weigh the corruption costs (e.g., the incumbent is
perceived as an efficient public manager, see Pereira
& Melo, 2015). Another conceivable reason is that
some voters may lack information about an incum-
bent’s or a political party’s misconduct (information
hypothesis, see Dunning et al., 2019). This latter
account is however not fully satisfactory for the
case of OPs because they remain loyal even when
they are aware that their party “made a mistake.”
With this previous knowledge in mind, we seek
to examine people’s (reduced) ability to evaluate
information and its sources critically as a potential
antecedent of OP. Information evaluation has been
proposed as a key component of information lit-
eracy, and its performance indicators refer to the
capacity to “use criteria for information evaluation,
source evaluation and evaluation of information
gathering methods” as well as to determine the
“usefulness of information” (Dolničar, Podgornik,
Bartol, & Šorgo, 2020, p. 6).
CONTACT Alberto Ardèvol-Abreu aardevol@ull.edu.es Departamento de Ciencias de la Comunicación y Trabajo Social, Universidad de La Laguna,
Camino La Hornera, 37, 38200 La Laguna, Spain
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS
https://doi.org/10.1080/19331681.2023.2173699
© 2023 Taylor & Francis
In the current saturated media environment,
citizens need basic information evaluation skills to
separate facts from opinions, independent from
interested sources of information, quality news
from propaganda and disinformation, etc. In the
political realm, people also need information eva-
luation skills to assess whether a certain informa-
tion (e.g., a political scandal) is important enough
to reconsider their support for a certain candidate
or party. OP, we argue, may be a convenient mental
shortcut for some people who lack information
evaluation skills and therefore cannot properly
understand the subtleties and implications of the
good or bad performance of a given political actor.
It is not necessarily that they are unaware of the
good and bad actions of the politicians (such as in
the information hypothesis), but they are less
equipped to properly assess the wider meaning of
those actions. This would result in an increased
tendency to disregard performance indicators and
support “their” party, “no matter what they do.”
Besides this proposed direct relationship
between information evaluation and OP, the cur-
rent investigation builds on the literature of selec-
tive exposure (Iyengar & Hahn, 2009) and
homogeneous discussion networks (Huckfeldt,
Mendez, & Osborn, 2004) to explore the possible
mediating role of political homophily in social
media and online in the relationship between infor-
mation evaluation and OP. Our measure of political
homophily in social media and online refers to
people’s tendency to isolate themselves in a self-
bubble of politically congenial information and
conversations on the internet and avoid disagree-
able viewpoints.
Previous work suggests that discussing politics
within large and ideologically dissimilar networks
“softens” OP (Ardèvol-Abreu & Gil de Zúñiga,
2020), that exposure to heterogeneous messages
makes citizens more likely to incorporate positive
and negative assessments of political candidates
(Huckfeldt et al., 2004), and that selective exposure
to likeminded media predicts increased polariza-
tion in political attitudes (Kim, 2015). We argue
that people with strong information evauation
skills are probably more aware of the importance
of considering different points of view to get a
complete and accurate picture of any situation,
particularly in the political arena. They are also
more likely to understand that the media environ-
ment includes biased sources and malicious actors
and that social media algorithms tend to feed them
with attitude-congruent information. Information
evaluation-skilled individuals may therefore strive
to create a more open information and discussion
environment that includes people and sources they
disagree with (i.e., reduce their political homophily
in social media and online), which will in turn
decrease their level of OP. To investigate these
possible direct and indirect connections between
information evaluation and OP, we take advantage
of an original, two-wave panel survey conducted on
a sample of 1,259 residents in Spain.
Connecting social media political homophily
with obstinate partisanship
The tendency of some individuals to interact with
people they agree with is not a new phenomenon.
Already in the early Columbia studies, Berelson,
Lazarsfeld, and McPhee (1954) noted that voters
tend to discuss politics with those holding similar
political orientations. There are several comple-
mentary explanations for the formation of homo-
geneous networks of political discussion, including
the higher availability of like-minded individuals in
one’s social circles (Stroud & Collier, 2018), peo-
ple’s general preference for similar others (Huber &
Malhotra, 2017; McPherson, Smith-Lovin, & Cook,
2001), or the distaste for cognitively dissonant
information and the consequent search for congru-
ent messages (see Colleoni, Rozza, & Arvidsson,
2014; Guo, Rohde, & Wu, 2018).
The related idea that people seek out content that
support their own views is not recent either.
Selective exposure was widely explored in the
1960s, but the lack of clear empirical support for
its principles at the time made it lose part of its
appeal (Stroud, 2008; see also Sears & Freedman,
1967). Most of the initial studies on partisan selec-
tive exposure were conducted in the United States,
where people were exposed to political news from
professional media sources that commonly adhered
to the canonical point-counterpoint journalism
(Iyengar & Hahn, 2009). In this “overtly neutral
media environment,” it was difficult for people to
find “partisan messages or messengers” (Iyengar &
Hahn, 2009, p. 21; see also Mutz & Martin, 2001).
2A. ARDÈVOL-ABREU ET AL.
But the proliferation of digital and social media
has brought renewed attention to the creation of
self-bubbles of political content and discussion,
especially in connection with the echo chamber
hypothesis (Garrett, 2009a). There are several pos-
sible mechanisms that may lead to the formation of
these bubbles of political homogeneity. One of
them is selective affiliation: Since online and social
media increase users’ options to select content and
conversation partners, some people may simply
seek a more congenial media environment (e.g.,
via searching, friending, following, liking, replying,
etc.; see Bond & Sweitzer, 2022). The complemen-
tary process would be an active avoidance of opi-
nion challenges (e.g., via unfriending, unfollowing,
or hiding, see Skoric, Zhu, & Lin, 2018). The latter
seems to be a less common strategy (Garrett,
2009b), although may still play a noticeable role in
some cases. Another mechanism has to do with the
way automatized systems select and rank the stories
users are exposed to and the “Whom to follow”
suggestions. This involuntary mechanism for selec-
tive exposure is frequently referred to as algorith-
mic curation and provides the foundation for the
“filter bubble argument” (see, for example,
Cardenal, Aguilar-Paredes, Galais, & Pérez-
Montoro, 2019; see also Gil de Zúñiga, Cheng, &
González-González, 2022 for details on the opera-
tionalization of this concept).
The extent to which social media and, more
generally, the internet, facilitate cross cutting
exchange and exposure to diverse ideas or, on the
contrary, catalyze like-minded interactions and the
reinforcement of previous attitudes is still an issue
of academic debate (Bond & Sweitzer, 2022; see also
Cinelli, Morales, Galeazzi, Quattrociocchi, &
Starnini, 2021). In any case, we can reasonably
contend that there is variance across individuals
in their levels of political homophily – in other
words, in the extent to which they are trapped in
an online self-bubble of content and interactions or
strive to scape it. We argue that these differences
may influence the extent to which individuals
declare unconditional loyalty to “their” party or,
on the contrary, make their support conditional to
its performance. Those with high levels of political
homophily may be aware of “their” party’s bad
performance or even scandals, but they: a) perceive
lower level of criticism toward the wrongdoings of
“their” party and a higher level of criticism toward
the wrongdoings of other parties; b) are exposed to
less detail about “their” party’s bad or poor perfor-
mance and its broader implications for society; c)
are more likely to encounter messages that reframe
the bad sides of “their” party in a less negative way
(e.g., it is an isolated case and not indicative of a real
problem). This is partly because they are more
exposed to online content from hyperpartisan
media and interested sources, rumors, and misin-
formation, which disseminate more easily within
echo chambers (Recuero et al., 2022; Törnberg,
2018; see also Choi, Chun, Oh, Han, & Kwon,
2020). This differential exposure to like-minded
ideas, “us versus them” dynamics, and alternative
framing of political events may make those with
high levels of political homophily more likely to
“dig in their heels” and increase their level of OP.
More formally:
H1: Political homophily in social media and online
has a positive effect on OP.
Information evaluation and social media
political homophily
The amount and complexity of the information
people are exposed to in the contemporary media
environment can be overwhelming. In the political
and news media arena, the challenge is not only the
massive flow of information, but also the wide
circulation of inaccurate, biased, misleading, and
false content that can make it harder to gain a
clear picture of the political developments. In this
context, information literacy is “no longer a nice-
to-have proficiency but a requirement” for critically
engaging with news and ensuring informed parti-
cipation (Machete & Turpin, 2020).
Since its emergence in the mid-1970s in the field
of library and information science, the term infor-
mation literacy and its competency standards have
been defined and updated in many ways. The
American Library Association (ALA) defined infor-
mation literacy as the set of abilities that enable
individuals “to recognize when information is
needed” and “locate, evaluate, and use effectively
the needed information” (American Library
Association, 1989). More recent accounts of the
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS 3
construct emphasize its intersections with critical
thinking and implications for (political) expression
and participation. Thus, the UK’s Library and
Information Association defines information lit-
eracy as “the ability to think critically and make
balanced judgements about any information we
find and use. It empowers us as citizens to reach
and express informed views and to engage fully
with society” (Information Literacy Group, 2018,
p. 3). Information literacy skills are commonly
developed during the educational process and
incorporated into the curriculum “at both, upper-
primary and secondary school levels” (Dolničar et
al., 2020, p. 2)
Based on a synthesis of contemporary perspec-
tives on information literacy, Dolničar and collea-
gues (2020) developed an information literacy
competence framework consisting of seven stan-
dards such as “formulate information needs” or
“handle information ethically” (p. 6). Under this
framework, information literacy is a relatively
stable set of skills “that today’s adolescents need to
successfully navigate the abundance of informa-
tion” and become “informed and responsible citi-
zens” (Dolničar et al., 2020, pp. 1–4). As part of this
work, they also designed an information literacy
knowledge test comprised of 48 multiple-choice
questions with one correct answer from 4 alterna-
tives that tests actual knowledge instead of per-
ceived abilities.
One of the standards proposed in this framework
is information evaluation, understood as the abil-
ities to “use criteria for information evaluation,
source evaluation and evaluation of information
gathering methods” and to determine the “useful-
ness of information” (Dolničar et al., 2020, p. 6).
Information evaluation-related abilities are closely
associated with critical thinking and involve keep-
ing an open mind to new information and perspec-
tives while assessing the quality of the evidence
behind each claim. For example, fact-checking stra-
tegies to verify online information such as learning
more about the sources of the information, corro-
borating their claims against other evidence, or
finding the original version of the information
(see “lateral reading” in Brodsky et al., 2021) are
only effective when people are willing to confront
their own biases and look beyond their congenial
perspectives. Information evaluation is therefore at
odds with dogmatic adherence to previous ideas
and, we argue, with the confinement in a commu-
nity of like-minded voices. We suggest that those
with higher levels of information evaluation are
more likely to take action to expand their political
self-bubble and actively search for diverse sources
and points of view, thus reducing their political
homophily. Stemming from these ideas, we propose
our second hypothesis:
H2: Information evaluation has a negative effect on
political homophily in social media and online.
Direct and indirect paths from information
evaluation to obstinate partisanship
As detailed on the previous sections, we antici-
pate a negative association between information
evaluation and political homophily in social
media and online, such that those more skilled
in evaluating the expertise and credibility of a
given source and determining the usefulness of
the information (i.e., with adequate information
evaluation skills) will be interested in gaining
broader perspectives and strive to escape their
self-bubble. Furthermore, we propose a positive
relationship between political homophily in
social media and online and OP: Those with a
higher tendency to rely on their self-bubble of
homophilic content and interactions will show
more “obstinate” political attitudes. Accordingly,
it also seems pertinent to explore an indirect
relationship between information evaluation
and OP through political homophily in social
media and online. The association between
information evaluation and political homophily
(H2) represents the a path of the proposed med-
iation model, while the connection between poli-
tical homophily and OP (H1) constitutes the b
path. Hence, we propose our third hypothesis:
H3: Information evaluation has an indirect effect
on obstinate partisanship via political homophily in
social media and online. Put differently, those with
higher levels of information evaluation will tend to
reduce their levels of political homophily and, in
turn, their degre of OP.
4A. ARDÈVOL-ABREU ET AL.
Finally, it might also be possible that some other
mechanisms, not considered in this study, mediate
the relationship between information evaluation
and OP. If this were the case, the direct path from
information evaluation to OP after political homo-
phily has been controlled (c’ path) would be signif-
icantly different from zero, and the proposed
mediation would be partial. We therefore formulate
a research question:
RQ: Is there a direct effect of information evalua-
tion on OP after controlling for the indirect effect
of information evaluation on OP through political
homophily in social media and online (suggesting a
partial mediation)?
Methods
Sample
This study is part of a broader research project that
uses two waves of survey data to examine the atti-
tudinal and behavioral impact of specific uses of
mobile and internet technologies (record IDs #
2020/1833 and 2021/0001994). For the purposes
of data collection, we contracted the services of
the international online research company
Netquest, which administers an online panel in
Spain with more than 119,000 panelists. The team
developed the questionnaires, entirely in Spanish
and uploaded them to the survey site Qualtrics.
Between July 5 and 8, 2021, Netquest sent the
survey links to a sample of 3,458 panelists with a
demographic composition reflective of the Spanish
population in terms of gender, age, and educational
level. Respondents received compensation for their
time. From this first wave (W
1
), we received 1,259
valid responses, each with a unique ID – but with-
out any personally identifying information. The
sample was quite balanced in terms of gender
(47.9% females) and had a mean age of
45.66 years (SD = 14.99). The age distribution was
as follows: younger than 25 years, 10.9%; 25–
34 years, 14.9%; 35–44 years, 22.3%; 45–64 years,
38.3%; and 65 years or over, 13.6%. The median
education level of the sample was high school grad-
uate (M = 3.79; SD = 2.09 in a scale ranging from
1 = primary education to 7 = postgraduate and
doctoral studies; 14.3% had only primary education;
37.2% had some university education or more).
Three and a half months later, Netquest recon-
tacted all respondents who completed the question-
naire in W
1
. The survey was open between October
18 and 28, 2021. In this second wave (W
2
), 982 of
the original respondents completed the question-
naire, for a retention rate of 78%. Judged by the
standards of other published research in the field
that use panel survey data, our retention rate in the
second wave is rather high. As a result of this
dropout, the W
2
sample was slightly older and
better educated than in W
1
: W
2
age, M =
47.02 years, SD = 14.54; W
2
education, M = 3.91;
SD = 2.09.
Variables of interest
Information evaluation
Our key independent variable measures one of the
several information literacy standards that guide
people “toward becoming informed and responsi-
ble citizens” (Dolničar et al., 2020, p. 1). To mea-
sure this dimension of information literacy, we
used the seven multiple-choice items that comprise
the information evaluation subscale in Dolničar et
al. (2020) knowledge test. All seven items are asso-
ciated with one of the two performance indicators
of information evaluation: a) “knows how to use
criteria for information evaluation, source evalua-
tion and evaluation of information gathering meth-
ods;” or b) “determines usefulness of information”
(Dolničar et al., 2020, p. 6). Items include questions
such as “Which of the following organizations
represents the highest authority in assessing the
reliability of information on new scientific achieve-
ments?” (Dolničar et al., 2020, p. 10). The last item
of the subscale (“I do not know the proper
Slovenian word for the English term hashtag
[. . .]”) was adapted to make it relevant to the
Spanish context. Correct responses were coded as
1, while incorrect or missing ones where coded as 0.
We generated an additive index of information
evaluation ranging from 0 to 7 (W
1
: M = 3.50‚
SD = 1.42).
Political homophily in social media and online
Building on previous studies on selective exposure
(Iyengar & Hahn, 2009) and homogeneous
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS 5
discussion networks (Huckfeldt et al., 2004), we
used three items to capture respondents’ propensity
to interact with like-minded people and content on
the internet and avoid disagreeable viewpoints (Gil
de Zúñiga, Cheng, & González-González, 2022):
“When you are online or on social media, you
tend to consume news and information that are
aligned with your viewpoints,” “When you are
online or on social media, you tend to avoid expo-
sure to news and information that are not aligned
with your viewpoints,” “You live in your own ‘bub-
ble’ online or on social media, mostly connecting
with people like you and looking for opinions you
agree with” (averaged scale, 1 = strongly disagree to
5 = strongly agree; W
1
Cronbach’s α = .58
1
; M =
3.07; SD = 0.78; W
2
Cronbach’s α = .64; M = 3.06;
SD = 0.79).
Obstinate partisanship
Based on previous operationalizations of the con-
struct (Ardèvol-Abreu & Gil de Zúñiga, 2020), our
main dependent variable asked respondents about
their level of agreement with the following three
statements: “You will always vote for the same
political party, no matter what they do,” “You sup-
port your political party, even when they make a
mistake,” and “Being loyal to your party is impor-
tant, both when they are doing well and not so well”
(averaged 5-point disagree-agree scale); W
1
Cronbach’s α = .87; M = 2.10; SD = 1.06; W
2
Cronbach’s α = .87; M = 2.05; SD = 1.06).
Control variables
All our regression models include two blocks of
control variables measured in W
1
, namely demo-
graphics (age, gender, education, and family house-
hold income) and political/media variables
(strength of partisanship, social media use for
news, and professional news media use).
Demographic controls were measured with single
items (see description of the sample above). Family
household income was measured at ordinal level
(W
1
median = 4 [€1,801 to 2,700]). Strength of
partisanship was assessed by combining two survey
items. The first one is a dichotomous yes/no ques-
tion that asked respondents whether they consider
themselves to be “close to any political party.” “No”
responses to this first question were recoded as 0
(weakest partisanship), while those who replied “yes”
were asked another question to explore whether they
feel “very close to this party, fairly close or merely a
sympathizer” (see similar question wordings in
Schmitt, Hobolt, Van der Brug, & Popa, 2020).
“Very close” responses were recoded as 3 (strongest
partisanship), “fairly close” as 2, and “merely a sym-
pathizer” as 1 (range 0 to 3, W
1
: M = 0.57; SD = 0.84).
Social media use for news was measured with five
Likert-type items about respondents’ frequency of
news consumption – in the last month – on
“Facebook,” “Twitter,” “Instagram,” “YouTube,”
and “social media other than mentioned above”
(averaged scale, 1 = never to 5 = all the time; W
1
Cronbach’s α = .69; M = 2.45; SD = 0.84). Finally, we
also asked respondents about their frequency of pro-
fessional news media use, including “TV news,”
“national newspapers,” “local or regional newspa-
pers,” and “radio” (averaged 4-item scale, W
1
Cronbach’s α = .67; M = 2.63; SD = 0.82).
Statistical analyses
To test H1 and H2, we ran a series of cross-sectional
and autoregressive ordinary least squares regression
models. Cross-sectional models include W
1
mea-
sures of both the predictor and the outcome variable.
Autoregressive models “strengthen cause-and-effect
interpretations” and consider the “temporal
sequence whereby the hypothesized causal influence
precedes the emergence of the hypothesized effect
over some relevant period of time” (Roth &
MacKinnon, 2012, pp. 186–187). To do so, we
included a W
2
measure of the dependent variable
in the model, but also its baseline (W
1
) score, which
serves as covariate for the criterion variable.
Following a similar strategy, our autoregressive
mediation models include as covariates W
1
mea-
sures of the mediator and the outcome variable
(see a similar approach in Selenko & Batinic, 2013;
Wakefield, Bowe, Kellezi, Butcher, & Groeger, 2019).
All analyses were performed with SPSS, version
25, with the aid of two macros: HCREG, to imple-
ment heteroskedasticity-consistent standard error
estimators (HC0) instead of assuming homoske-
dasticity (see Hayes & Cai, 2007); and PROCESS,
to perform the mediation analyses (Hayes, 2017).
6A. ARDÈVOL-ABREU ET AL.
Results
Our first hypothesis postulated a positive effect of
political homophily in social media and online on
OP. We found empirical support for this prediction.
Both regression models in Table 2 show that political
homophily in social media and online predicts OP in
the first wave of the study (β = .215, p < .001, Model
C) and in the autoregressive approach (β = .057, p =
.047, Model D). This relationship holds in the media-
tion models in Figure 1A,B: β = .216, p < .001 (cross-
sectional); β = .131, p < .001 (W
1
-W
2
-W
2
mediation
model). Among the control variables, age and
strength of partisanship were the most robust predic-
tors of OP (Table 2, Model D). Compared to the W
1
overall sample, the subsample of respondents scoring
high (4 or more) on OP was older (M = 52.11 years,
SD = 15.63), included more men (64.6%), had a lower
education level (M = 2.76, SD = 1.67), and expressed a
higher level of partisanship (M = 0.99, SD = 1.12)
H2 stated that information evaluation is a negative
predictor of political homophily in social media and
online. Cross-sectional and autoregressive OLS
regression models (A and B) in Table 1 provide
empirical support for this second hypothesis: infor-
mation evaluation is concurrently and longitudinally
associated with reduced levels of political homophily
(β = −.067, p = .029; β = −.072, p = .033, respectively).
Models in Figures 1A,B) provide further support for
H2: information evaluation negatively predicts politi-
cal homophily in the cross-sectional (Model E, a path,
β = −.064, p = .039) and autoregressive (Model F, a
path, β = −.073, p = .036) simple mediation models.
H3 hypothesized an indirect effect of information
evaluation on OP via reduced levels of political homo-
phily in social media and online. Those scoring high
on information evaluation will tend to lower their
levels of political homophily and, in turn, reduce
their OP attitudes. This indirect association was sta-
tistically significant cross-sectionally (Model E and
Table 3, ab path, completely standardized point esti-
mate = −.014, bootstrapped SE = .007, CI 95% [−.028,
−.001]) and longitudinally (Model F and Table 3, ab
path, completely standardized point estimate = −.010,
bootstrapped SE = .005, CI 95% [−.022, −.001]). This
provides empirical support for H3 and suggests that
information evaluation does help reduce levels of OP
through political homophily. This effect, however,
seems to be only partially mediated by online political
homophily (RQ): In both mediation models, the
Table 1. OLS regression models predicting political homophily in
social media and online.
Political Homophily in Social Media and Online
Cross-Sectional
(W
1
-W
1
, Model A)
Autoregressive
(W
1
-W
2
,
Model B)
Block 1: Autoregressive Control W
1
Political Homophily in SM and Online – .387***
∆R
2
– 16.3%
Block 2: Demographics W
1
Age .028 −.083#
Gender (Female) .071* .046
Education .059 .002
Income −.046 −.019
∆R
2
0.6% 0.6%
Block 3: Political and
Media Variables W
1
Strength of Partisanship .137*** .029
Social Media Use for News .059# .021
Professional News Media Use .012 .075*
∆R
2
2.4% 0.6%
Block 4: Variable of Interest W
1
Information Evaluation −.067* −.072*
∆R
2
0.4% 0.5%
Total R
2
3.4% 17.9%
Note. Sample sizes: Cross-sectional model n = 1,079. Autoregressive n = 839.
Missing cases were excluded using listwise deletion, which is preferable
when the data is missing not at random (Pepinsky, 2018). Standardized
regression coefficients reported. Significance tests were computed using
the Huber-White robust method (HC0, see Hayes & Cai, 2007). # p < .10; *
p < .05; ** p < .01; *** p < .001 (two-tailed). W
1
= Wave 1. W
2
= Wave 2.
Table 2. OLS Regression Models Predicting Obstinate Partisanship
Obstinate Partisanship
Cross-Sectional (W
1
-
W
1
, Model C)
Autoregressive (W
1
-
W
2
Model D)
Block 1: Autoregressive
Control W
1
Obstinate Partisanship — .510***
∆R
2
— 34.4%
Block 2: Demographics W
1
Age -.007 .079*
Gender (Female) -.066* -.018
Education -.119*** .007
Income -.040 .025
∆R
2
3.2% 0.5%
Block 3: Political and
Media Variables W
1
Strength of Partisanship .198*** .144***
Social Media Use for
News
.087** .050
Professional News Media
Use
.064* -.049
∆R
2
7.4% 2.2%
Block 4: Variables of Interest
W
1
Political Homophily in
SM and Online
.215*** .057*
Information Evaluation -.206*** -.074*
∆R
2
8.8% 0.8%
Total R
2
19.4% 37.9%
Note. Sample sizes: Cross-sectional model n = 1,074. Autoregressive n = 831.
Missing cases were excluded using listwise deletion, which is preferable
when the data is missing not at random (Pepinsky, 2018). Standardized
regression coefficients reported. Significance tests were computed using
the Huber-White robust method (HC0, see Hayes & Cai, 2007). # p < .10; * p
< .05; ** p < .01; *** p < .001 (two-tailed). W
1
= Wave 1. W
2
= Wave 2.
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS 7
direct effect of information evaluation on OP remains
significant after controlling for the indirect effect of
information evaluation on OP through political
homophily (cross-sectional Model E, c’ path, β =
−.206, p < .001; autoregressive Model F, c’ path,
β −.066, p = .036; see Figure 1A,B).
2
Discussion and conclusions
Common views of democratic accountability
emphasize the importance of voters’ critical
engagement based on their evaluations of politi-
cians, parties, and governments’ performance. In
consonance with the retrospective voting thesis,
citizens are expected to judge whether politicians
and government institutions have done a good job
or not, and act accordingly. For this model to work,
citizens have to 1) be equipped with the skills to
gather, evaluate, and understand political informa-
tion, and 2) be willing to punish and relinquish
their loyalty when “their” party makes a mistake
or does not perform well (Ardèvol-Abreu & Gil de
Zúñiga, 2020). Considering the steady increase of
information, misinformation, and disinformation
in digital environments, as well as the potential of
online media to isolate their users in political self-
(a)
(b)
Figure 1. (A) (Model E). Cross-sectional model exploring the indirect relationship between information evaluation and obstinate
partisanship through political homophily in social media and online. (B) (Model F). Autoregressive model exploring the indirect
relationship between information evaluation and obstinate partisanship through political homophily in social media and online.
Sample sizes: Cross-sectional model (1A), n = 1,074. Autoregressive model (1B), n = 827. Models were computed using model 4 of the
PROCESS macro for SPSS (version 3.5) (Hayes, 2017). Grey arrows represent model covariates and include demographic (age, gender,
education, and income) and political and media variables (strength of partisanship, social media use for news, and professional news
media use). Although not represented for simplicity, the macro also estimates (for Model F) the association between W
1
“political
homophily in social media and online” and W
2
“obstinate partisanship,” on the one hand, and between W
1
“obstinate partisanship”
and W
2
“political homophily in social media and online,” on the other. Significance tests use the Huber-White robust method (HC0).
Path entries are standardized coefficients. Models include indirect effects (reported in Table 3). * p < .05; *** p < .001 (two-tailed).
8A. ARDÈVOL-ABREU ET AL.
bubbles, this study investigates information lit-
eracy-related skills and their possible influence on
the way citizens use online media and question
their political loyalties.
Thus, our analyses focused on people’s capacity
to “use criteria for information evaluation, source
evaluation and evaluation of information gathering
methods” and determine the “usefulness of infor-
mation” (information evaluation, see Dolničar et
al., 2020, p. 6) and its potential role in reducing
political homophily in social media and online and
attitudes of OP. Our findings suggest a positive role
of information evaluation in reducing political
homophily. This may indicate that those with
higher information evaluation skills strive to create
a more open and diverse online environment, prob-
ably because they are more aware of the circulation
of interested, biased, and even blatantly false infor-
mation online. Those scoring high on information
evaluation may also understand the complexities
and subtleties of today’s political issues and may
therefore be more willing to consider a variety of
perspectives when building their own views. On the
contrary, those who lack basic information evalua-
tion skills may be less prone to actively try to
“break” their online self-bubble, where they see
news that is aligned with their views and connect
with people they agree with. Algorithm curation
mechanisms will probably reinforce the vicious cir-
cle by feeding them with more opinion-friendly
perspectives.
Our results also support the prediction that poli-
tical homophily in social media and online has a
positive effect on OP. Those with higher levels of
political homophily are more exposed to congenial
information and reinforcing discussions, which, we
argue, leaves them more vulnerable to biased fram-
ing, polarized ideological discourse, change of
focus, and other (manipulative) discursive strate-
gies (see Soares & Recuero, 2021). For example,
when a political actor is caught doing something
wrong, hyperpartisan media and other tendentious
sources are more likely to reframe the facts and
create alternative versions of the events (Soares &
Recuero, 2021) that circulate more easily in bubbles
of political homophily. Soares and Recuero (2021)
found that hyperpartisan outlets during the 2018
Brazilian presidential campaign used strategies of
biased framing and “us versus them” polarized
ideological discourse “to reframe the events to cre-
ate an ‘alternative’ narrative” (p. 11). In Spain,
political corruption scandals have been long played
down using arguments such as “this was an isolated
case,
3
” “all politicians/parties are the same,
4
” “other
candidates/parties have stolen more,
5
” etc. Those
who live in a bubble of political homophily may
perceive that “their” party’s wrongdoings are not as
bad as other parties’ wrongdoings, or that they are
mere exceptions that are irrelevant to the overall
picture. With these perceptions in mind, people
scoring high on political homophily in social
media and online may be more prone to rely in
their partisan identity and to not react to any
external event. They will therefore be more likely
to support “their” party no matter what they do,
when they are doing not so well, and even when
they make a mistake.
Our study also explored political homophily as a
potential mediating mechanism through which
information evaluation relates to OP. The main
idea is that information evaluation-skilled indivi-
duals tend to have reduced levels of political homo-
phily, and that their openness to more diverse
political ideas and discussions in turn diminishes
their OP attitudes. Our findings also support this
interpretation and suggest that the ability to evalu-
ate information, the reliability of the sources, and
the usefulness of its content (i.e., information eva-
luation, see Dolničar et al., 2020, p. 6) has strong
democratic implications. Information evaluation-
related skills may help citizens to step out of their
self-bubble of political homophily online and
become more open to new ideas, even if they dis-
confirm their previous beliefs. A more
Table 3. Indirect effects models of information evaluation on obstinate partisanship.
Indirect Effects Paths Point Estimate Boot. 95% C. I.
Information Evaluation W
1
→ Political Homophily SM and Online W
1
→ Obstinate Partisanship W
1
−.014 −.028 to −.001
Information Evaluation W
1
→ Political Homophily SM and online W
2
→ Obstinate Partisanship W
2
−.010 −.022 to −.001
Notes: Indirect effects are based on models E and F in Figures 1A,B (cross-sectional and autoregressive models, respectively). Completely standardized
coefficients reported. Models were bootstrapped 10,000 iterations (seed: 31,216).
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS 9
heterogeneous online environment may, in turn,
make people more responsive to the political per-
formance of politicians and parties such that they
are more likely to hold them accountable if they
perform poorly. In this respect, we see OP attitudes
as an obstacle to political accountability and a par-
tial explanation for why, at times, poorly perform-
ing – or even corrupt – politicians are reelected.
Our findings therefore contribute to the growing
realization that improving citizens’ information
evaluation and, more generally, information lit-
eracy skills, can be a suitable strategy to move
toward more open, critical, and less polarized socie-
ties in sociopolitical terms. The use of media-based
manipulative strategies is already taking place
(through, for example, change of focus, biased
framing, polarized ideological discourse, or fabri-
cated information; see Soares & Recuero, 2021),
and increasing evidence suggests that people in
online self-bubbles are particularly vulnerable to
them. On top of this, the current picture is one in
which ordinary citizens have limited time and skills
to deal with the flood of information and opinions
that circulate over the internet and social media.
This also applies to younger citizens: Previous
research indicates that some of the labels placed
on younger generations (e.g., “digital natives,”
“Net Generation,” or “Generation@”) do not corre-
spond with their enhanced skills to navigate the
digital world, but rather to their greater need to
acquire them – in light of their increased activity
on online platforms and social media (Bennett,
Maton, & Kervin, 2008; Buckingham, 2015;
Coombes, 2009; Ng, 2012). The present study is
therefore in line with earlier calls on the urgency
to develop a comprehensive strategy for informa-
tion literacy and related skills that empowers citi-
zens to access, evaluate, and effectively use political
information.
Interestingly, our indirect effects models suggest
partial mediations because, in both the cross-sec-
tional and autoregressive approaches, the c’ path is
different from zero: There seems to be a direct,
negative effect of information evaluation on OP
even after controlling for the indirect effect through
political homophily. We interpret this finding as an
indication that there could be additional mechan-
isms (not explored in this study) that explain the
negative effect of information evaluation on OP.
Further studies should theorize additional explana-
tions and perform parallel mediation analyses
including different intervening variables. For exam-
ple, those with lower scores on information evalua-
tion may be more prone to engage in partisan
motivated reasoning, a cognitive bias that is used
to protect one’s partisan identification (Bolsen,
Druckman, & Cook, 2014). Motivated reasoning
may therefore function as an additional mediating
mechanism between information evaluation
and OP.
While this article makes substantive contribu-
tions, it is not without limitations. One of them
comes from the nature of our longitudinal data.
We collected two waves of surveys, which allowed
us to control for wave 1 levels of the outcome
variable in our regression analyses and therefore
account for the temporal sequence between the
hypothesized cause and the hypothesized effect.
However, our mediation analysis only incorporates
the temporal dimension in the a path of the media-
tion (i.e., the information evaluation- political
homophily path). If we had had a third wave of
panel data, we could have modeled an indirect
effects model that included a wave 1 measure of
information evaluation (independent variable), a
wave 2 measure of political homophily (mediator),
and a wave 3 measure of OP (dependent variable).
To improve causal inference, future studies should
collect three or more waves of data and incorporate
the temporal dimension in both a and b paths of the
mediation. Another caveat is that our sample is
essentially a convenience sample derived from a
panel of respondents that are regularly asked to
participate in different studies. Our respondents
are probably not fully representative of the
Spanish population, but the sample is demographi-
cally diverse and includes respondents from across
the country, which we see as sufficient for the
explanatory purposes of this study.
All in all, the present study suggests that informa-
tion evaluation may play a significant role in shaping
citizens’ behaviors online and minimizing obstinate
attitudes of partisanship that could undermine poli-
tical accountability. These findings can be regarded
as an opportunity in the current context of polarized
politics, information overload, and widespread dis-
semination of disinformation. In this vein, the pro-
motion of information evaluation skills in and
10 A. ARDÈVOL-ABREU ET AL.
outside educational settings could help unleash the
potential of the internet and social media for infor-
mation diffusion and verification, public debate, and
insightful exchange of ideas. This may play a positive
role in political life, creating a more savvy, less
“obstinate” electorate that is ready to punish or
reward their politicians based on their performance.
Notes
1. The Cronbach’s alpha coefficient for political homo-
phily in social media and online is somewhat low in
both waves. Although there is no undisputed interpre-
tation of what an acceptable alpha value is, a commonly
used rule of thumb suggests values of .70 or higher.
However, it is also known that the alpha coefficient is
sensitive to the number of items in a test, and with short
scales – our measure of “political homophily” is com-
posed of only three items – it is common to find low
values of alpha. We performed a principal axis factoring
(PAF) with Varimax rotation on the six items of poli-
tical homophily in social media and online and obsti-
nate partisanship. This analysis produced two factors
that explained more than 67% of the total variance and
fit our two constructs. The political homophily in social
media and online items loaded together on a single
factor that suits our theoretical approach. Considering
the face validity of the measure and the result of the
PAF, we decided to use this three-item scale.
2. We hypothesized a single direction of causality, from
information evaluation to OP via political homophily in
social media and online. We see information evaluation
as a set of skills primarily grounded on education and
therefore relatively stable over time. We find it less
convincing to argue that people’s obstinate partisanship
can undermine their ability to know “how to use criteria
for information evaluation, source evaluation and eva-
luation of information gathering methods” or deter-
mine the “usefulness of information” (see Dolničar et
al., 2020, p. 6). However, following the suggestions of
one of the reviewers of this article, we also tested the
reverse causality hypothesis – from OP to information
evaluation via political homophily in social media and
online. These analyses indicate that OP does not predict
political homophily in social media and online once we
control for the W
1
measure of political homophily
(autoregressive model, β = .029, p = .388). Political
homophily in social media and online does not seem
to mediate the relationship between OP and informa-
tion evaluation either (indirect effect autoregressive
model, point estimate = −.001, 95% CI = −.004 to
.003). In contrast, there is a direct, negative association
between W
1
levels of OP and W
2
information evalua-
tion (autoregressive model, β = −.118, p <.001).
3. In Spain, former PM Rajoy and other politicians repeatedly
reframed the corruption schemes that involved political
parties throughout the country as a few “isolated cases”
or “cases from many years ago” (see Díez & Mateo, 2018).
4. A Spanish journalist and TV panelist, well-known
for his right-wing political sympathies, was asked in
a TV show in late 2017 about political corruption in
Spain. He reasoned as follows: “Regarding my vote, I
do what I want. And everyone in this country is
aware that they steal a lot from us . . . a lot of
money. But I do decide who steals from me. And a
communist will never steal from me” (La Gaceta de
la Iberosfera, 2017).
5. See the practice of whataboutism (y tú más, in Spanish)
in Pérez Curiel, Jiménez-Marín, and Pulido Polo (2021).
Disclosure statement
No potential conflict of interest was reported by the author(s).
Funding
The first autho’s contribution was funded by the “Viera y Clavijo”
Program from the Agencia Canaria de Investigación, Innovación
y Sociedad de la Información and the Universidad de La Laguna
[2019/0000093]; Data collection for the study was funded by the
Universidad de La Laguna [grants 2021/0001994 and 2020/1833].
Notes on contributors
Dr. Alberto Ardèvol-Abreu, PhD in Communication, is
Associate Professor (Profesor Titular de Universidad) at the
Department of Communication Science and Social Work at
the University of La Laguna (ULL, Spain). He is also the
principal investigator of ‘Laboratorio de Investigación sobre
Medios y sus Efectos’ (LIME) research group. His research
focuses on the effects of digital and social media on civic and
democratic life. He is also interested the broad phenomenon of
disinformation and its impact on democracy.
Dr. Carmen Costa-Sánchez, PhD in Communication and
Journalism, is Associate Professor (Profesora Titular de
Universidad) of Corporate Communication at the
Department of Sociology and Communication Sciences at
the Universidade da Coruña (UDC, Spain). She is also the
principal investigator for the ‘Culture and Interactive
Communication’ research group (UDC). Her research inter-
ests include corporate communication, health communica-
tion, transmedia storytelling, media literacy, and social media.
Dr. Patricia Delponti, PhD in Communication, is Assistant
Professor (Profesora Contratada Doctora) at the Department
of Communication Science and Social Work (ULL). She is also
a member of the LIME research group. For more than twenty
years, she worked as a journalist and a PR professional at
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS 11
public and private institutions. Her research interests focus on
news and broadcast media, corporate communication, and the
use of ICTs in education.
ORCID
Alberto Ardèvol-Abreu http://orcid.org/0000-0001-8722-
5226
Carmen Costa-Sánchez http://orcid.org/0000-0001-8154-
9537
Patricia Delponti http://orcid.org/0000-0001-9694-867X
References
Adserà, A., Boix, C., & Payne, M. (2003). Are you being
served? Political accountability and quality of government.
The Journal of Law, Economics, and Organization, 19(2),
445–490. doi:10.1093/jleo/ewg017
American Library Association (1989). Presidential committee
on information literacy. Final report. American Library
Association. Accessed 9 November, 2021. https://www.ala.
org/acrl/publications/whitepapers/presidential
Ardèvol-Abreu, A., & Gil de Zúñiga, H. (2020). “Obstinate
partisanship”: Political discussion attributes effects on the
development of unconditional party loyalty. International
Journal of Communication, 14(22), 324–345.
Bennett, S., Maton, K., & Kervin, L. (2008). The ‘digital
natives’ debate: A critical review of the evidence. British
Journal of Educational Technology, 39(5), 775–786. doi:10.
1111/j.1467-8535.2007.00793.x
Berelson, B. R., Lazarsfeld, P. F., & McPhee, W. N. (1954).
Voting: A study of opinion formation in a presidential elec-
tion. Chicago, IL: University of Chicago Press.
Bolsen, T., Druckman, J. N., & Cook, F. L. (2014). The influ-
ence of partisan motivated reasoning on public opinion.
Political Behavior, 36(2), 235–262. doi:10.1007/s11109-
013-9238-0
Bond, R. M., & Sweitzer, M. D. (2022). Political homophily in
a large-scale online communication network.
Communication Research, 49(1), 93–115. doi:10.1177/
0093650218813655
Brodsky, J. E., Brooks, P. J., Scimeca, D., Todorova, R., Galati,
P., Batson, M., . . . Caulfield, M. (2021). Improving college
students’ fact-checking strategies through lateral reading
instruction in a general education civics course. Cognitive
Research: Principles and Implications, 6(1), 1–18. doi:10.
1186/s41235-021-00291-4
Buckingham, D. (2015). Defining digital literacy. What do
young people need to know about digital media? Nordic
Journal of Digital Literacy, 10, 21–34. doi:10.18261/1891-
943X-2006-04-03 Jubileumsnummer.
Cardenal, A. S., Aguilar-Paredes, C., Galais, C., & Pérez-
Montoro, M. (2019). Digital technologies and selective
exposure: How choice and filter bubbles shape news
media exposure. The International Journal of Press/
Politics, 24(4), 465–486. doi:10.1177/1940161219862988
Choi, D., Chun, S., Oh, H., Han, J., & Kwon, T. (2020). Rumor
propagation is amplified by echo chambers in social media.
Scientific Reports, 10(1), 1–10. doi:10.1038/s41598-019-
57272-3
Cinelli, M., Morales, G. F., Galeazzi, A., Quattrociocchi, W., &
Starnini, M. (2021). The echo chamber effect on social
media. Proceedings of the National Academy of Sciences,
118(9). 10.1073/pnas.2023301118
Cobb, M. D., & Taylor, A. J. (2015). An absence of malice: The
limited utility of campaigning against party corruption.
American Politics Research, 43(6), 923–951. doi:10.1177/
1532673x15570470
Colleoni, E., Rozza, A., & Arvidsson, A. (2014). Echo chamber
or public sphere? Predicting political orientation and mea-
suring political homophily in Twitter using big data.
Journal of Communication, 64(2), 317–332. doi:10.1111/
jcom.12084
Coombes, B. (2009). Generation Y: Are they really digital
natives or more like digital refugees? Synergy, 7(1), 31–40.
Díez, A., & Mateo, J. J. (2018, May 24). Rajoy define como
“casos aislados” las tramas de corrupción que afectan al PP.
El País. https://elpais.com/politica/2018/05/24/actualidad/
1527146690_840422.html
Dolničar, D., Podgornik, B. B., Bartol, T., & Šorgo, A. (2020).
Added value of secondary school education toward devel-
opment of information literacy of adolescents. Library &
Information Science Research, 42(2), 1–18. doi:10.1016/j.lisr.
2020.101016
Dunning, T., Grossman, G., Humphreys, M., Hyde, S. D.,
McIntosh, C., Nellis, G., . . . Sircar, N. (2019). Voter infor-
mation campaigns and political accountability: Cumulative
findings from a preregistered meta-analysis of coordinated
trials. Science Advances, 5(7), eaaw2612. doi:10.1126/sciadv.
aaw2612
Eggers, A. C. (2014). Partisanship and electoral accountability:
Evidence from the UK expenses scandal. Quarterly Journal
of Political Science, 9(4), 441–472. doi:10.1561/100.
00013140
Fernández-Vázquez, P., Barberá, P., & Rivero, G. (2016).
Rooting out corruption or rooting for corruption? The
heterogeneous electoral consequences of scandals. Political
Science Research and Methods, 4(2), 379–397. doi:10.1017/
psrm.2015.8
Garrett, R. K. (2009a). Echo chambers online?: Politically
motivated selective exposure among Internet news users.
Journal of Computer-Mediated Communication, 14(2), 265–
285. doi:10.1111/j.1083-6101.2009.01440.x
Garrett, R. K. (2009b). Politically motivated reinforcement
seeking: Reframing the selective exposure debate. Journal
of Communication, 59(4), 676–699. doi:10.1111/j.1460-
2466.2009.01452.x
Gil de Zúñiga H, Cheng Z and González-González P. (2022).
Effects of the news finds me perception on algorithmic news
attitudes and social media political homophily. Journal of
Communication, 72(5), 578–591. 10.1093/joc/jqac025
12 A. ARDÈVOL-ABREU ET AL.
Goyanes, M., Borah, P., & Gil de Zúñiga, H. (2021). Social
media filtering and democracy: Effects of social media news
use and uncivil political discussions on social media
unfriending. Computers in Human Behavior, 120, 106759.
doi:10.1016/j.chb.2021.106759
Group, I. L. (2018). CILIP Definition of Information Literacy.
Accessed 25 November, 2021. https://infolit.org.uk/
ILdefinitionCILIP2018.pdf
Guo, L., Rohde, J. A., & Wu, H. D. (2018). Who is responsible
for Twitter’s echo chamber problem? Evidence from 2016 U.
S. election networks. Information, Communication & Society,
23(2), 234–251. doi:10.1080/1369118X.2018.1499793
Hayes, A. F. (2017). Introduction to mediation, moderation,
and conditional process analysis: A regression-based
approach. New York, London: Guilford Press.
Hayes, A. F., & Cai, L. (2007). Using heteroskedasticity-con-
sistent standard error estimators in OLS regression: An
introduction and software implementation. Behavior
Research Methods, 39(4), 709–722. doi:10.3758/bf03192961
Huber, G. A., & Malhotra, N. (2017). Political homophily in
social relationships: Evidence from online dating behavior.
The Journal of Politics, 79(1), 269–283. doi:10.1086/687533
Huckfeldt, R., Mendez, J. M., & Osborn, T. (2004). Disagreement,
ambivalence, and engagement: The political consequences of
heterogeneous networks. Political Psychology, 25(1), 65–95.
doi:10.1111/j.1467-9221.2004.00357.x
Iyengar, S., & Hahn, K. S. (2009). Red media, blue media:
Evidence of ideological selectivity in media use. Journal of
Communication, 59(1), 19–39. doi:10.1111/j.1460-2466.
2008.01402.x
Kim, Y. (2015). Does disagreement mitigate polarization?
How selective exposure and disagreement affect political
polarization. Journalism & Mass Communication
Quarterly, 92(4), 915–937. doi:10.1177/1077699015596328
La Gaceta de la Iberosfera. (2017, December 4). Ojeda: ‘Yo
decido quien me roba y un comunista no me va a robar’
[video file]. Retrieved from https://www.youtube.com/
watch?v=Kl6v2RmFQ6M
Machete, P., & Turpin, M. (2020). The use of critical thinking
to identify fake news: A systematic literature review. In M.
Hattingh, M. Matthee, H. Smuts, I. Pappas, Y. K. Dwivedi,
& M. Mäntymäki Eds., Responsible design, implementation
and use of information and communication technology (Vol.
2, 235–246). Cham: Springer. Accessed 3 January, 2022.
McPherson, M., Smith-Lovin, M., & Cook, J. M. (2001). Birds
of a Feather: Homophily in Social Networks. Annual Review
of Sociology, 27, 415–444. doi:10.1146/annurev.soc.27.1.415
Mutz, D. C., & Martin, P. S. (2001). Facilitating communica-
tion across lines of political difference: The role of the mass
media. American Political Science Review, 95, 97–114.
doi:10.1017/S0003055401000223
Ng, W. (2012). Can we teach digital natives digital literacy?
Computers & Education, 59(3), 1065–1078. doi:10.1016/j.
compedu.2012.04.016
Pepinsky, T. B. (2018). A note on listwise deletion versus
multiple imputation. Political Analysis, 26(4), 480–488.
doi:10.1017/pan.2018.18
Pereira, C., & Melo, M. A. (2015). Reelecting corrupt incum-
bents in exchange for public goods: Rouba mas faz in Brazil.
Latin American Research Review, 50(4), 88–115. doi:10.
1353/lar.2015.0054
Pérez Curiel, C., Jiménez-Marín, G., & Pulido Polo, M. (2021).
Political corruption, leadership and influence on Twitter.
An analysis of public transparency in the context of the 28th
April elections in Spain. Mediterranean Journal of
Communication, 12(2), 1–18. Accessed 23 December,
2021. https://www.doi.org/10.14198/MEDCOM000029
Recuero, R., Soares, F. B., Vinhas, O., Volcan, T., Hüttner, L. R., &
Silva, V. (2022). Bolsonaro and the far right: How disinforma-
tion about COVID-19 circulates on Facebook in Brazil.
International Journal of Communication, 16, 148–171.
Roth, D. L., & MacKinnon, D. P. (2012). Mediation analysis
with longitudinal data. In J. T. Newson, R. N. Jones, & S. M.
Hofer (Eds.), Longitudinal data analysis: A practical guide
for researchers in aging, health, and social sciences (pp. 181–
216). New York: Routledge.
Schmitt, H., Hobolt, S. B., Van der Brug, W., & Popa, S. A.
(2020). European parliament election study 2019, voter
study. GESIS Data Archive, Cologne. doi:10.4232/1.13473
Sears, D. O., & Freedman, J. L. (1967). Selective exposure to
information: A critical review. Public Opinion Quarterly, 31
(2), 194–213. doi:10.1086/267513
Selenko, E., & Batinic, B. (2013). Job insecurity and the benefits of
work. European Journal of Work and Organizational
Psychology, 22(6), 725–736. doi:10.1080/1359432x.2012.
703376
Skoric, M. M., Zhu, Q., & Lin, J. H. T. (2018). What predicts
selective avoidance on social media? A study of political
unfriending in Hong Kong and Taiwan. American Behavioral
Scientist, 62(8), 1097–1115. doi:10.1177/0002764218764251
Snyder, J. M., Jr, & Strömberg, D. (2010). Press coverage and
political accountability. Journal of Political Economy, 118
(2), 355–408. doi:10.1086/652903
Soares, F. B., & Recuero, R. (2021). Hashtag wars: Political disin-
formation and discursive struggles on Twitter conversations
during the 2018 Brazilian presidential campaign. Social Media
+ Society, 7(2), 1–13. doi:10.1177/20563051211009073
Stroud, N. J. (2008). Media use and political predispositions:
Revisiting the concept of selective exposure. Political
Behavior, 30(3), 341–366. doi:10.1007/s11109-007-9050-9
Stroud, N. J., & Collier, J. R. (2018). Selective exposure and
homophily during the 2016 presidential campaign. In B. R.
Warner, D. G. Bystrom, M. S. McKinney, & M. C. Banwart
(Eds.), An unprecedented election: Media, communication,
and the electorate in the 2016 campaign (pp. 21–39). Santa
Barbara, CA: Praeger.
Törnberg, P. (2018). Echo chambers and viral misinformation:
Modeling fake news as complex contagion. PloS ONE, 13
(9), 1–21. doi:10.1371/journal.pone.0203958
Wakefield, J. R. H., Bowe, M., Kellezi, B., Butcher, A., & Groeger,
J. A. (2019). Longitudinal associations between family identifi-
cation, loneliness, depression, and sleep quality. British Journal
of Health Psychology, 25(1), 1–16. doi:10.1111/bjhp.12391
JOURNAL OF INFORMATION TECHNOLOGY & POLITICS 13
