News consumption proles facing disinformation: exposure and
responses to suspect information
Perles de consumo informativo ante la desinformación: exposición y reacciones
ante una información sospechosa
doxa.comunicación |
nº 43, pp. 347-371
| 347
July-December of 2026
ISSN: 1696-019X / e-ISSN: 2386-3978
How to cite this article:
Pastora Estebanez, P. and García-Faroldi, L. (2026). News consumption proles facing disinformation: exposure and responses to suspect
information.
Doxa Comunicación
, 43, pp. 347-371.
https://doi.org/10.31921/doxacom.3519
Pablo Pastora Estebanez.
PhD student at the University of Málaga. He holds a BA in Sociology from the University
of Granada and an MA in Applied Sociology from the University of Málaga. He previously worked as a researcher on
the Andalusian project
“Impact of Disinformation in Andalusia: A Cross-Sectional Analysis of Audiences, Journalistic
Routines, and Media Agendas (Desinfoand)”
. He is currently a researcher in the national project
“Journalistic Applications
of AI to Reduce Disinformation: Trends, Uses, and Perceptions among Professionals and Audiences (DESINFOPERIA)”
.
His research interests focus on disinformation, articial intelligence and its impact on journalism and audiences,
polarization, and conspiracy theories. He approaches these topics from the perspective of computational sociology,
integrating data analysis and computational models to better understand complex social phenomena.
.
University of Malaga, Spain
[email protected]
ORCID:
0000-0003-4629-4860
Livia García-Faroldi.
Associate professor in Sociology at the University of Málaga. Her research focuses on public
opinion, using both national and international surveys. She has studied support for the European integration process,
how that support interacts with dierent political identities, and how anti-European discourse is linked to xenophobia.
Her most recent lines of research analyse the digital divide in political participation, as well as public opinion regarding
fake news and the spread of disinformation among citizens. She has led, together with Laura Teruel, the Excellence
Project “Impact of disinformation in Andalusia: cross-sectional analysis of audiences and journalistic routines
and agendas. DisinfoAND” (ProyExcel_00143, PAIDI 2020) and edited the open access book
Mass Media Facing
Disinformation: Articial Intelligence, Hate Speech, Conspiracy eories, and Fact-Checking
, published by Tirant.
.
University of Malaga, Spain
[email protected]
ORCID:
0000-0001-7816-7562
is content is published under Creative Commons Attribution Non-Commercial License. International License CC BY-NC 4.0
Recibido: 23/12/2025 - Aceptado: 15/04/2026 - En edición: 22/06/2026 - Publicado: 01/07/2026
Resumen:
El presente artículo analiza los perles de consumo informativo y su
relación con la sospecha de haber recibido información falsa y con
las reacciones generadas por dicha sospecha. A partir de una encues
-
ta representativa de la población andaluza (n=1.550), se identican
Received: 23/12/2025 - Accepted: 15/04/2026 - Early access: 22/06/2026 - Published: 01/07/2026
Abstract:
is article analyzes patterns of news consumption and their
relationship with the perception of having been exposed to false
information, as well as with the responses triggered by such perceptions.
Based on a representative survey of the Andalusian population
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
348 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
1. Introduction
e increasingly widespread use of digital devices, especially smartphones with
continuous internet connectivity, has brought
about profound changes in recent years in how information is accessed, interpreted, and valued in contemporary society.
News consumption is now fragmented, information disengagement is growing, along with distrust of both traditional media,
social media, and online platforms. is paper addresses the dierent proles of news consumption that exist today and how
they relate to the phenomenon of disinformation.
2. eoretical framework
2.1. e transformation of how we consume news
News consumption has undergone profound changes in recent years. Since the internet burst onto the scene in daily life,
causing the digitization of traditional media, the expansion of social media to ever wider parts of the population has brought
about further transformations in the news landscape. ey can be summarized in three points: (i) the decentralization and
personalization of consumption; (ii) the erosion of media professionals’ roles
;
and (iii) the growth of news avoidance and
incidental news consumption.
Regarding the decentralization and personalization of news consumption, the latest Spanish edition of the
Digital News Report
(2025, hereinafter DNR) conrms that all news sources have lost ground, even social media, and that online consumption is
now predominant (seven out of ten Spaniards have used an online channel). Television remains the leading source of news
tres perles de consumo mediante análisis de conglomerados. Poste
-
riormente, se estiman modelos de regresión logística para analizar la
probabilidad (1) de sospechar haber recibido desinformación, (2) de
reaccionar a través de la consulta de medios de comunicación o (3) de
vericación, o (4) la probabilidad de no hacer nada. Los resultados
muestran que los perles con mayor consumo digital presentan una
mayor probabilidad de sospecha, pero no necesariamente una mayor
reacción. El interés político emerge como el principal predictor de las
respuestas activas, mientras que la pasividad se concentra especial
-
mente en perles digitales de bajo interés informativo. Estos hallazgos
sugieren que la gestión de la desinformación recae sobre los sectores
más politizados de la ciudadanía, lo que refuerza las dinámicas de
desigualdad en la esfera pública digital.
Palabras clave:
Noticias incidentales; desconexión informativa; IAG; redes sociales;
agencias de vericación.
(n=1,550), three consumption proles are identied through cluster
analysis. Subsequently, logistic regression models are estimated to
examine (1) the likelihood of suspecting exposure to disinformation,
(2) the likelihood of reacting by consulting news media, (3) engaging
in verication practices, or (4) taking no action. e results show that
proles with higher levels of digital consumption are more likely to report
suspicion of disinformation, but not necessarily more likely to respond
to it. Political interest emerges as the main predictor of active responses,
whereas passivity is particularly concentrated among digitally oriented
proles with low informational interest. ese ndings suggest that the
management of disinformation is disproportionately borne by the most
politically engaged segments of the population, thereby reinforcing
existing inequalities within the digital public sphere.
Keywords:
Incidental news consumption; news avoidance; generative articial
intelligence; social media; fact-checking.
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 349
(54%), followed by social media (46%), the digital versions (websites and apps) of traditional newspapers (25%), and lastly,
print newspapers (19%) and radio (18%).
e generational gap is evident in news consumption patterns, linked to the younger generations’ massive and constant use
of social media. According to the CIS
Study on Social Media Audiences (II)
(2025, study 3,511), 56.2% of those aged 18 to 24
usually get their news from television, compared to 70.4% of the population as a whole. In contrast, television is mentioned by
more than 80% of Spaniards aged 55 and over. At the other end of the spectrum, among those aged 18 to 24, the use of social
media as a news source is overwhelming (87.3%). Conversely, its use is much lower
–
though far from negligible
–
among those
aged 65 to 74 (38.7%) and, especially, among those 75 and over (14.5%).
e prominence of social media as an information source can be explained by several of its dening characteristics. According
to uses and gratications theory, audiences are motivated to choose a medium that satises their needs and provides
gratications such as surveillance, information seeking, and maintaining contact with others (Katz
et al.
, 1973, Katz
et al.
, 1974,
McQuail
et al.
, 1972). us, news fulls several needs: cognitive, aective, personal integration, and escapism or catharsis (Katz
et al.
, 1973). In a recent qualitative study, Anderson (2025) found that social and emotional gratications are more frequent
among young people and women. ese types of gratications, moreover, are easier to obtain through social media than
through traditional media. is author nds that, while established news values such as relevance, timeliness, and magnitude
are important, younger people value live content and the ability to share news through social media.
News consumption has changed not only because of diversication of the channels through which citizens access information,
but also because of the nature of consumption itself, which is more personalized and potentially allows for greater public
participation (although this does not necessarily mean that such participation will come about). According to the DNR (2025),
algorithmic access is the predominant mode today (59%), while 37% of direct access to news is through searches of journalistic
brands, social media, search engines, aggregators, and notications.
A second consequence of the changes in news consumption habits relates to the waning of the role of news professionals, whose
role as gatekeepers (ltering and selecting what is news and what is published) has diminished. is phenomenon is closely
related to the rise of novel news channels and, in particular, to the emergence of so-called Social Media Inuencers. (SMI,
Hudders
et al.
, 2020). ese are gures who can be considered opinion leaders, and who extrapolate classic communication
theories such as those of Lazarsfeld
et al.
(1944) and Katz (1957) to the digital context. Peter and Muth’s qualitative study
(2023) analyses the impact of inuencers on the forming of political opinions. Young people use inuencers as a complement
to traditional media, especially to understand complex issues (Stehr
et al.
, 2015). e interviews show that traditional media
are considered more credible sources of political information than these leaders, who inuence more by reinforcing prior
opinions than by altering them.
e latest signicant change in the communications sphere concerns the growth of news avoidance and incidental news
consumption. e latest DNR report (2025) states that 37% of Spaniards say they often or sometimes avoid the news, especially
the young and those who express disinterest in politics or hold extreme views. e perceived information level (measured
by the degree to which individuals consider themselves informed about what is happening in the country) is 68.5% in Spain,
according to data from CIS study 3.511 (2025) mentioned above. at shows a decrease of 4.5 percentage points from two
350 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
years earlier (Feriouni & Ahmed-Mohamed, 2024). In both cases, young people consider themselves less informed than the
population as a whole, although their average rose in 2025 from 59% to 60.7%. e group that considers itself the best informed
has indeed changed. In 2023, it was the 55-64 age group (79% self-identied as such), while in 2025 it is the 45-54 year-olds
(73.4%). Feriouni and Ahmed-Mohamed (2024) point to three reasons why young people have this lower self-perception:
1) their interests are not aligned with current events, which are heavily focused on politics and economics, as they prefer
news related to entertainment and sports (a trend corroborated in the study by Catalina-García
et al.
, 2021); 2) their swift,
fragmented consumption of news through social media; and 3) their mostly incidental (unintentional) exposure, which leads
them to perceive their knowledge as more supercial. However, the young population is hardly uniform, and dierences can
be found in how news is consumed according to age and educational level (Catalina-García
et al.
, 2021).
As Boczkowski
et al.
(2017) point out, incidental news consumed through social media represents a break from the structured
way traditional media was consumed, with its structured consumption times. Incidental consumption blurs the line between
news and other content arriving via social media; news access is constant, but clicks on news are sporadic, and little time is
spent reading it.
Alongside the rise in news avoidance and the consumption of incidental news, a reduction in news diet has been observed
in recent years. According to data from the DNR (2025), 9% do not consult any news brands in Spain, 21% use one or two
news outlets per week, and 49% consult ve or more outlets. As the report points out, the data has both positive and negative
implications: on the positive side, it implies brand loyalty among some consumers, but on the other hand, the risk of cognitive
biases and information isolation has risen, reinforced by the growing personalization of content.
2.2. e spread of disinformation in digital environments and its consequences
Disinformation, dened by the European Commission as “veriably false or misleading information that is created, presented
and disseminated for economic gain or to intentionally deceive the public, and may cause public harm” (2018, p. 5), has become
widespread as one of the dening phenomena of contemporary society, with its increasingly evident interference in politics
both nationally and internationally. So much so that the World Economic Forum, in its Global Risks Report 2025, identied
false information
–
whether intentional (disinformation) or unintentional (misinformation)
–
as the primary risk for the next
two years. ese phenomena are linked to others, such as conspiracy theories and hate speech, and have consequences both
in the communication sphere and in the social and political spheres. e spread of false narratives inuences social dynamics
such as the discrediting of the media, social polarization, the weakening of institutional trust, and the erosion of scientic
knowledge.
Shu
et al.
(2020) categorize the challenges posed by disinformation into two types: content-related and user-related. e former
type is often sensationalist and emotional, aecting the news consumer more and prompting greater interaction (Shu
et al.
,
2017). erefore, posts containing false content often go viral and trend on social media (Vosoughi
et al.
, 2018). Furthermore,
the low cost of creating sources of disinformation and the ease of using software-controlled social media bots to help spread it
contribute to the problem (Shao
et al.
, 2017).
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 351
Regarding user-related challenges, those who get their information through social media are more vulnerable to disinformation.
A recent study (Montiel Torres
et al.
, 2025) observed that young people (15-24 years old) report greater diculty in identifying
false content, while, at the other extreme, the over-75s (the main consumers of news through traditional media) show greater
condence in their ability to do so. Pariser (2011) warns that an important aspect of social media platforms to consider is
the existence of “bubbles” or “echo chambers” created as a result of the platforms’ recommender systems. Users are thus
exposed to content that aligns with their own beliefs, reinforcing them in the process, while keeping users unaware of opposing
narratives and beliefs. Such “echo chambers” or “bubbles” make the task of detecting and mitigating disinformation especially
dicult because, as Shu
et al.
note (2020), repeated exposure to the same viewpoint only reinforces pre-existing beliefs and
increases resistance to changing one’s opinion, even if the information is proven false. In turn, such news “bubbles” produce
an increasing distance between groups of citizens with diering opinions, hindering consensus and fostering polarization.
Recent academic debate focuses on digital platforms’ role in this process. While some studies argue that “echo chambers” and
“lter bubbles” (Terren & Borge, 2021) isolate individuals in homogeneous news environments, other research nuances that
view, noting that incidental exposure to diering perspectives still exists (Dubois & Blank, 2018). Such intentional, selective
exposure explains why, beyond the control of algorithms, users tend to seek out and reinforce content that conrms their
previous beliefs, while actively rejecting opposing views (Tucker
et al.
, 2018).
However, disinformation does not circulate solely through social media. Given the low cost of creating and publishing online
content and the enormous reach of social media platforms, several alternative media sources (so-called
pseudo-media
) have
recently emerged that often disseminate false and/or highly biased claims. When mainstream media outlets report on such
claims to debunk them, they actually increase their reach
–
initially limited
–
and can consequently lead to increased distrust of
legitimate media sources among audiences (Haller & Holt, 2019).
e combination of a diet of news consumed through social media and the spread of disinformation in those settings has
negative consequences for trust in the media. ree studies analysing the Spanish situation are highlighted here. Hameleers
et al.
(2022) conducted a survey in ten European countries and concluded that the perception that disinformation and
misinformation exist reduces trust in the news and is associated with greater consumption of news from non-journalistic
sources. e association between lower trust in the news and the use of social media and other non-journalistic channels
had been conrmed in the analysis by Fletcher and Park (2017) in 11 countries, using data from the DNR (2015); Park and
colleagues (2020) reached the same conclusion using DNR data from 2016 and 2019 for 26 countries.
e latest Digital News Report (2025) identies four population proles in Spain based on their interest in news and the
credibility they accord it. e category that has grown most in the last ve years is the so-called “news neets”, meaning those
who both declare little interest in the news and bestow it little credibility. While they represented a quarter of the population in
2021, by 2025 they accounted for 38%. is highlights the increasingly widespread information disengagement among citizens.
At the other end of the spectrum, those who show great interest in the news and considerable trust in it are decreasing in
number, falling from 29% in 2021 to 20% in 2025, almost half of the rst group.
However, trust in the news is not a uniform phenomenon, and signicant dierences exist depending on the news source.
Studies show greater trust when people are asked about news they consult personally than when they are asked about trust
352 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
in general (40% in the rst case and 31% in the second, according to DNR data for Spain, 2025). Various surveys have also
revealed that people trust news from news brands more than that coming from social media. Data from CIS study 3,511 (2025)
shows that over half of the population trusts radio and print media news, one in four television news, and one in six for social
media news. Interestingly, young people’s trust in all these news sources is higher than that of the general population, except
for radio. e dierence is particularly pronounced with regard to social media
–
rstly
–
and the press
–
secondly. 29.7% of
young people trust news posted on social media, compared to 63.1% for that published in the press. is greater trust in
the press reinforces the trend found in the previous edition of the survey (Feridouni & Ahmed-Mohamed, 2024). Similarly,
Catalina-García
et al.
(2021) surveyed young people aged 17-24 in Madrid and found that they accord greater credibility to
traditional than to social media, even though the latter is their primary channel.
Finally, this section on the rise of disinformation in the hybrid media system cannot conclude without mentioning the recent
emergence of Generative Articial Intelligence (GAI) and its role in this context. ere is a general consensus among the
public that it helps to produce (fake images, deepfakes) and distribute disinformation (through bots
and recommendation
algorithms), although a majority also believe that it can help ght hoaxes through tools that facilitate their detection (García-
Faroldi
et al.
, 2025).
2.3. Research questions and hypotheses
After a review of the literature, this study aims to answer the following research questions:
1.
What are the news consumption proles that characterize the population? What sociodemographic, political, and
informational characteristics are they related to?
2.
Are there dierences in the probability of suspecting having encountered false information and reactions to that
suspicion depending on news consumption proles?
3.
What sociodemographic, political, and informational variables are related to the probability of suspecting having
encountered false information and to the reactions to that suspicion?
With regard to the rst research question, it is expected that news consumption proles will be mainly related to greater or
lesser access to news through social media (H1).
Regarding the second research question, signicant dierences are expected in the probability of having encountered false
news, this being more likely among those who inform themselves more through social media (H2a). Signicant dierences are
also expected based on news consumption proles in the reactions to such suspicion (H2b).
Finally, it is expected that informational and political variables will be of greater importance than sociodemographic variables
in estimating the probability of suspecting having received false news and of the reactions to that suspicion (H3).
3. Methodology
Data from a representative survey of the Andalusian population aged 15 and over was used to answer the research questions
and verify the hypotheses proposed. is survey was conducted within the framework of the Excellence Project “Impact of
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 353
disinformation in Andalusia: a cross-sectional analysis of audiences and journalistic routines and agendas (DesinfoAND)”,
part of the Andalusian Plan for Research, Development and Innovation (PAIDI 2020, ProyExcel_00143). e survey was carried
out in December 2023 by a polling rm contracted specically for the purpose. e sample comprised 1,550 people: 1,250 from
online surveys drawn from a panel of over 9,000 people, and 300 telephone interviews conducted with the population aged 60
and over
1
. is mixed method ensured that the older population were adequately represented in the overall sample, as online
surveys tend to under-represent that age group. e sample was stratied by size of residence using data from the 2021 Census,
and quotas were established for sex and age using data from the Municipal Register (
Padrón Continuo
) (2021). e margin of
error for a 95.5% condence level was ±2.55% for the overall sample, assuming simple random sampling
2
.
e study was conducted in accordance with the Declaration of Helsinki. Informed consent was obtained from all participants,
and data anonymity was ensured. e study was approved by the Ethics Committee of the University of Málaga (protocol
160-2024-H).
Five questions were chosen to create the dependent variables.
Firstly, the item concerning the suspicion of having received false information is as follows:
inking about the last 7 days, have you received any information or news through any channel (through the mass media, social
media, instant messaging, video platforms, etc.) that makes you suspect it may be false?
Yes / No / I don’t know, I’m not sure.
For the purposes of the analysis, only the rst two responses were taken into account to perform the logistic regressions, and
those who indicated that they were not sure were considered missing values (496 cases).
Secondly, the respondent’s reaction on receiving information of this type is included in a multiple-choice question:
If you had any doubts about whether information from social media (Facebook, Twitter/X, Instagram, TikTok, etc.) or messaging
apps (WhatsApp, Telegram, etc.) was true or false, how would you react?
Select all that apply.
–
I would nd out where the message came from.
–
I would cross-check the information with traditional media (press, radio, television).
–
I would consult specialized fact-checking sources.
–
I would ask people around me who are informed about current aairs.
–
I wouldn’t do anything.
–
I would stop following the individual or digital media outlet that had disseminated it.
1
e individuals participating in the online panel were recruited by the polling rm via the internet, with nancial incentives for their participation. e telephone
survey of people aged 60 and over was conducted using randomly selected phone numbers.
2
It is standard practice in public opinion studies to calculate the margin of error for simple random sampling. While the introduction of quotas for sex and age
means that the sampling does not meet the randomness requirement, thus increasing the margin of error, stratication decreases the variance and, therefore, the
error. e nal sample matches the designed sample in terms of age, sex, and habitat size.
354 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
–
I would report it as false news.
–
Not applicable. I never use the internet, or I don’t use any social media or instant messaging.
–
I don’t know.
Due to space limitations, only the second (check with media outlets), the third (consult fact-checking sources), and the fth
(do nothing) responses were chosen for analysis in this study. ese elicit dichotomous responses (Yes/No), if the respondent
selected the option.
Regarding the independent variables, the following question was used to create the news consumption proles:
How often
do you use the following sources for news
?
Television/ Print newspapers/ Digital newspapers/ Radio/ Social media and video
platforms.
e respondent could answer on a Likert scale of 1 to 5 for each source, the data was then recoded so that the highest values
corresponded to the greatest frequency of use: Every day (5), Once or more times a week (4), A few times a month (3), Hardly
ever (2), Never (1), Don’t know.
e remaining independent variables were divided into two groups: those of a sociodemographic and sociopolitical nature
and those related to information.
a) Sociodemographic and political variables:
Included here are sex (categories male/female); age (divided into 15 to 24 years/ 25 to 44 / 45 to 59 / 60 and over); educational
level (three categories: up to compulsory secondary studies/ non-compulsory secondary studies/ university); subjective social
class (the ve original responses have been recoded into three: high (sum of high and upper-middle classes)/ middle/ working
class (sum of lower-middle and working class)); political ideology (the original scale of 0 to 10 was recoded into ve groups:
extreme left (values 0-2), left (3-4), centre (5), right (6-7) and extreme right (8-10)); and the respondent’s declared interest in
politics (the original variable with four responses was recoded into two groups, those who show interest -having responded a
lot or quite a lot- and those who do not -having indicated little or no interest-).
b) Informational variables:
is group includes questions related to trust in dierent sources and interest in the news.
Questions related to trust stem from the question:
–
To what extent do you trust the news and information you access or receive through…? A lot/ Quite a bit/ Little/ Not at all/ I
don’t use it to access news/ Don’t know
–
Television
–
Print newspapers
–
Digital newspapers (online)
–
Radio
–
Social media
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 355
–
Instant messaging applications (WhatsApp, Telegram, Facebook Messenger, etc.)
–
Friends, family, and coworkers
Traditional media outlets, including digital newspapers, have been recoded as a single variable that captures the average
trust of all of them. If a respondent indicated that they did not use a certain channel, that channel was not included in the
calculation of the average. Only 7 people said that they did not use any of the proposed channels for news.
To simplify the analysis, the categories of variables that measure trust (in the media, social media, messaging apps, and people
close to them) were recoded into two: trustful (for those who chose the options “a lot” or “quite a bit”) or distrustful (for those
who chose “little” or “not at all”). is dichotomization, and subsequent ones, is based on criteria of analytical parsimony and
ease of interpretation, since the intermediate categories on the scales show limited substantive dierences between them, and
therefore, dichotomization allows for more stable estimates in logistic regression models.
Finally, interest in the news is reected in the question:
To what extent would you say you are interested in the news?
–
I’m very interested in it.
–
I’m quite interested in it.
–
I’m not very interested in it.
–
I’m not interested in it at all.
–
I don’t know.
Similarly to how interest in politics was handled, the original variable was recoded to dichotomize it: those who show interest
–having responded a lot or quite a bit– and those who do not –responses of little or not at all.
e analytical strategy followed in this study consists of three parts:
1.
Cluster analysis to establish news consumption proles.
2.
Descriptive analysis to examine how proles are distributed in the dierent social categories studied.
3.
Logistic regression analysis with dichotomous dependent variables to estimate the probabilities related to each
independent variable of having received news suspected to be false and of reacting by consulting traditional media,
specialized fact-checkers, or by deciding to do nothing.
1. Analysis of clusters of news consumption proles:
e process to create the news consumption proles was as follows. A cluster analysis was performed based on the frequency
of use of dierent news sources. is procedure allowed individuals to be classied into internally homogeneous groups that
diered from each other, based on their media consumption patterns. e analysis was based on ve variables that capture the
frequency of use of the main news sources, as explained above.
e study was conducted in three complementary phases, following standard methodological approaches for multivariate
exploratory techniques. A hierarchical cluster analysis was applied in the rst phase using Ward’s method, employing the
356 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
squared Euclidean distance as a measure of dissimilarity to minimize within-group variance (Ward, 1963). Secondly, by
visualizing a dendrogram and using the Calinski-Harabasz test (1974), the three-group cluster was identied as the most
robust, maximizing the ratio of between-group to within-group variance (Milligan & Cooper, 1985). Finally, a non-hierarchical
cluster analysis (k-means clustering) was performed to optimize the partitioning based on the centroids obtained in the
previous hierarchical solution (Punj & Stewart, 1983). is combined strategy improved the stability of the classication and
validated the robustness of the three news consumption proles identied.
Having obtained the nal three-cluster solution, statistical validation and characterization of the proles was carried out using
analysis of variance (one-way ANOVA). e results of the analysis conrmed the existence of statistically signicant dierences
between the three clusters in all the consumption dimensions analysed (p < .001 in all cases).
2. Descriptive analysis:
Once the three clusters were identied, they were descriptively characterized. Descriptive statistics (mean and standard
deviation) were calculated for the frequency of use of each news channel within each group. Based on the results, the clusters
were substantively interpreted and labelled as per their dominant patterns of news consumption.
Furthermore, in order to externally validate the clusters, their association with several sociodemographic and political variables
was analysed. ese variables were examined using contingency tables and chi-square tests. e analysis conrmed that news
consumption proles are related dierently to relevant sociodemographic, trust-related, and political characteristics, thereby
reinforcing the substantive validity of the typology found.
3. Logistic regressions:
Binary logistic regression analyses were performed in the third phase with the aim of estimating the probability that individuals
suspect they have received false information and react dierently to said suspicion, based on their consumption prole, and
the other characteristics mentioned, be they sociodemographic, political, informative or related to trust.
e model’s results are presented using odds ratios, a common measure in analysis of this type due to its ease of substantive
interpretation. Odds ratios indicate how much the probability of the analysed event increases or decreases when an
independent variable increases by one unit or when a category is compared to the category of reference, all other independent
variables in the model remaining constant. Values greater than 1 indicate an increase in the probability of the event, while
values less than 1 indicate a decrease. To obtain more robust estimates against potential problems of heteroscedasticity or
deviations from the model’s classical assumptions,
robust standard errors were employed.
Finally, to aid interpretation of the results, adjusted probabilities (margins) were calculated for the dierent proles. e
estimates allow for an intuitive visualization of the dierences between groups, translating the model coecients into
predicted probabilities.
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 357
4. Results
4.1. Analysis of news consumption proles
Table 1 presents the average frequency of use (scale 1–5, where 5 = “every day” and 1 = “never”) for each news channel by
cluster. First, we nd the traditional prole (n = 601), characterized by very intensive use of traditional media, especially
television (M = 4.49) and radio (M = 4.02), with the majority of respondents reporting daily use in both cases. In contrast, social
media use is minimal (M = 1.56), concentrated in the categories “never” and “hardly ever”. Newspaper consumption, both
print and digital formats, is moderate to low, falling in the middle of the scale.
e second cluster represents a moderate digital prole (n = 406) dened by a generally low pattern of news consumption,
with the exception of social media, which constitutes their main route for accessing news (M = 4.03). Television use (M = 3.52)
is at intermediate levels, while radio (M = 1.76) and newspapers (especially print) show lower values. is prole, therefore,
combines a strong centrality of social media with a low overall news intensity, which dierentiates it from both the traditional
and omnichannel proles.
Finally, the third cluster presents the intensive omnichannel prole (n = 515), characterized by high and widespread
consumption across all news channels. is is primarily marked by intensive use of social media (M = 4.63) and television
(M = 4.60), followed by radio (M = 4.21) and digital newspapers (M = 3.77). is prole shows a high frequency of daily or
weekly consumption across virtually all media analysed, constituting a pattern of varied and sustained exposure to news.
Table 1. Media consumption frequencies by consumption proles
Consumption
prole
n
Television
(mean)
Print media
(mean)
Digital media
(mean)
Radio
(mean)
Social media
(mean)
Traditional6014.49
2.50 2.54 4.02 1.56
Moderate digital
4063.52
1.45 2.22 1.76 4.03
Intensive
omnichannel
5154.60
2.05 3.77 4.21 4.63
Total
1.5224.27
2.06 2.87 3.48 3.26
Source: Own elaboration
Note: All dierences between proles are statistically signicant (one-way ANOVA, p < .001)
358 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
As shown in Figure 1, the proles exhibit dierent sociodemographic and political characteristics, as well as distinct levels of
trust in information channels, when consuming news, with statistically signicant dierences across all the variables analysed.
e traditional prole is primarily composed of older individuals and has a lower proportion of the young. It is the only prole
where men outnumber women and has the lowest education level, whilst being in the middle in both ideology and social class.
ey show less trust in social media and digital platforms but relatively high levels of trust in traditional media and in people
close to them. e prole also exhibits intermediate levels of interest in politics and the news.
e second prole, termed moderate digital, is the youngest, with an over-representation of people between 15 and 24 years old.
It is the prole that most commonly has secondary education levels and lower socioeconomic status. ey are characterized
by lower trust in traditional media, apps, and people close to them, and by intermediate levels of trust in social media and
platforms. is prole shows the lowest levels of interest in politics and news, reinforcing their character as sporadic and
detached consumers of news.
Finally, the intensive omnichannel user prole is middle-aged, with a higher proportion of people between 25 and 44 years
old, and an over-representation of university graduates and those from middle and upper social classes. is prole shows the
highest levels of trust across all channels, including social media, digital platforms, and people close to them, and is the most
interested in both politics and news.
ese descriptive results point to the existence of structural inequalities in access, intensity, and orientation of news
consumption, which not only refer to generational dierences, but also to levels of interest, trust, and education.
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 359
Figure 1. Frequency of independent variables in news consumption proles
a)
all channels, including social media, digital platforms, and people close to them, and is the
most interested in both politics and news.
These descriptive results point to the existence of structural inequalities in access, intensity,
and orientation of news consumption, which not only refer to generational differences, but also
to levels of interest, trust, and education.
Figure 1. Frequency of independent variables in news consumption profiles
a)
b)
Source: Own elaboration in Python.
4.2.
Analysis of the suspicion of receiving false information and reactions to it:
Table 2 presents the results of the estimated logistic regression model used to analyse the
factors associated with the suspicion of having received false news, measured dichotomously.
The model is statistically significant (Wald χ² = 80.84; p < 0.001) and
explains a moderate
proportion of variance (Pseudo R² = 0.104).
Table 2. Logistic regression on the suspicion of disinformation
b)
all channels, including social media, digital platforms, and people close to them, and is the
most interested in both politics and news.
These descriptive results point to the existence of structural inequalities in access, intensity,
and orientation of news consumption, which not only refer to generational differences, but also
to levels of interest, trust, and education.
Figure 1. Frequency of independent variables in news consumption profiles
a)
b)
Source: Own elaboration in Python.
4.2.
Analysis of the suspicion of receiving false information and reactions to it:
Table 2 presents the results of the estimated logistic regression model used to analyse the
factors associated with the suspicion of having received false news, measured dichotomously.
The model is statistically significant (Wald χ² = 80.84; p < 0.001) and
explains a moderate
proportion of variance (Pseudo R² = 0.104).
Table 2. Logistic regression on the suspicion of disinformation
Source: Own elaboration in Python
360 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
4.2. Analysis of the suspicion of receiving false information and reactions to it
Table 2 presents the results of the estimated logistic regression model used to analyse the factors associated with the suspicion
of having received false news, measured dichotomously. e model is statistically signicant (Wald
χ
² = 80.84; p < 0.001) and
explains a moderate proportion of variance (Pseudo R² = 0.104).
Table 2. Logistic regression on the suspicion of disinformation
(Odds Ratios, Robust Standard Errors)
Dependent variable: Suspicion of having received false news (1 = yes)
Reference categories: Traditional prole · Age 15–24 · Male · Compulsory education or less · Distrustful · Ideologically centrist ·
Lower social class · No interest in politics · No interest in news
Variable
Odds Ratio
p
Consumption prole (Traditional)
Digital
1.85 **
0.043
Omnichannel
1.82 **
0.028
Yes
No
0.54 ***
0.001
Yes
No0.820.426
Yes
No
1.56 *
0.070
Yes
No
0.64 **
0.026
Interest in news (No)
Yes1.610.197
Age (15–24)
25–440.970.924
45–591.070.822
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 361
60+1.720.144
Sex (Male)
Female1.00
0.986
Education level (Compulsory or below)
Post-compulsory secondary
0.790.433
University1.010.973
Political ideology (Centre)
Far left0.54 **
0.043
Left0.890.660
Right0.720.232
Far right
0.750.363
Social class (Lower)
Middle0.860.451
Upper
0.41 **
0.008
Political interest (No)
Yes
3.45 ***
0.000
n=621
Pseudo R²=0,104
Wald
χ
²=80,84 ***
Source: Own elaboration
Notes: * p < 0.10 · ** p < 0.05 · *** p < 0.01
362 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Firstly, news consumption proles are one of the main explanatory factors. Compared to the traditional prole, both the digital
and omnichannel proles show a greater likelihood of suspecting they have received false news, with odds ratios
signicantly
greater than one. is result is consistent with the fact that, compared to the traditional news consumer, these proles use
more digital and social media and are more exposed to the disinformation circulating on those platforms. Conversely, classic
sociodemographic variables (age, sex, and education level) do not show statistically signicant associations with the suspicion
of disinformation once the other factors are allowed for.
Regarding political ideology, only the far-left category shows a signicant association, exhibiting a lower probability of suspicion
compared to centrist positions. Socioeconomic status also shows major dierences. Belonging to the upper social strata is
associated with a lower probability of suspicion compared to the working classes. Variables related to trust in news sources
reveal dierentiated eects depending on the channel. Trust in traditional media is associated with a lower probability of
suspicion, as is trust in people close to the individual. In contrast, trust in social media does not show a signicant relationship
with suspicion, while trust in messaging applications shows a moderately signicant positive eect. Finally, interest in politics
is one of the model’s most robust predictors. ose who are interested in politics are signicantly more likely to suspect they
have received false information. is eect is not replicated for general interest in news.
Overall, the results indicate that suspicion of disinformation is more closely linked to news consumption patterns, trust in
certain sources, and the degree of political involvement than to traditional sociodemographic characteristics.
Table 3 presents the results of three logistic regression models that analyse the factors associated with dierent reactions when
suspecting disinformation: consulting media outlets, consulting fact-checkers, and doing nothing. All models include the
same set of independent variables and share the same reference categories. For reasons of clarity, the table excludes variables
that were not statistically signicant for any of the three dependent variables.
Table 3. Factors associated with dierent reactions when suspecting disinformation
(Odds Ratios, Robust Standard Errors)
Reference categories: Traditional prole · Age 15–24 · Male · Compulsory education or less · Distrusts · Ideological centre · Lower social
class · No political interest · No interest in news
Variable
Consulting media outletsConsulting fact-checkersDoing nothing
Digital Prole
——
2.28 **
Omnichannel Prole
———
Trust in media (Trusts)
2.97 ***
——
Trust in apps (Trusts)
0.48 ***
——
Interest in news (Yes)
2.46 **
——
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 363
Age 45–59
——
3.49 **
Age 60+
0.59 *
—
4.18 **
Female
0.76 *
——
Post-compulsory secondary
——
3.64 **
University
—
0.63 *3.30 **
Right
—
0.55 **
—
Far right
—
0.31 ***3.69 ***
Upper social class
—
1.88 **
—
Political interest (Yes)
0.59 ***1.46 **0.56 *
n887887887
Pseudo R
2
0.0960.0600.117
Wald
χ
2
100.5763.5752.80
Source: Own elaboration
Notes: = not signicant (p ≥ 0.10); * p < 0.10; ** p < 0.05; *** p < 0.01. In all three models p < 0.001.
Firstly, the model explaining the likelihood of checking in traditional media shows no signicant dierences between news
consumption proles. However, relevant sociodemographic and trust eects do emerge. People aged 60 and over are less
likely to consult the media compared to the 15-24 age group, as are women, whose behaviour shows a lower propensity to
turn to the media when they have suspicions. In contrast, trust in traditional media signicantly increases the likelihood of
consulting such sources. Conversely, trust in apps or digital platforms is associated with a lower likelihood of consulting the
media, suggesting a shift in news sources. Similarly, interest in politics reduces the likelihood of consulting the media, while
interest in news signicantly increases it.
Regarding the model for consulting specialized fact-checkers, a discernible pattern emerges. News consumption proles do not
show signicant associations, but ideological, educational, and social class variables do. Individuals who identify as right-wing
or far-right are less likely to consult fact-checking sources compared to those with centrist positions. Conversely, belonging to
a higher social class and demonstrating an interest in politics signicantly increases the likelihood. As for educational level,
university graduates show a lower propensity to consult fact-checking sources.
364 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
In cases of inaction, clear dierences emerge related to news consumption patterns. ose with a moderate digital prole are
characterized by a greater likelihood of not reacting, compared to those of a traditional prole. Age is also a relevant factor:
both the 45–59 age group and, especially, the 60-plus group show a higher probability of taking no action when suspicious.
Similarly, individuals with post-compulsory secondary education and university degrees exhibit a greater propensity for
passivity. From an ideological perspective, the far right stands out for its higher probability of inaction, while political interest
is associated with less passivity, signicantly reducing the likelihood of not reacting.
Overall, the results show that reactions to disinformation do not follow a uniform pattern, but rather depend on the type
of response considered. While passivity is particularly associated with age, the low-consumption digital prole, and certain
ideological positions, active responses such as checking in media outlets and using fact-checking agencies appear to be more
closely linked to political interest, ideological stance, and, in the case of fact-checking, social class.
e adjusted probability analysis (Table 4) allows for the interpretation of logistic regressions in terms of probabilities and
facilitates a substantive comparison between news consumption proles. First, the results show clear dierences in the
probability of suspecting having received disinformation. While the estimated probability among individuals with a traditional
prole is around 48%, this increases to approximately 61% among moderate digital proles and 60% among intensive
omnichannel proles. ese results indicate that, when allowing for all variables, proles with greater exposure to digital
environments are more likely to suspect they have received potentially false content.
Table 4. Adjusted probabilities (margins) by news consumption prole
Dependent variable
Traditional
Moderate
digital
Intensive
omnichannel
n
Pseudo
R²
Suspecting disinformation
48%61%60%6210.104
Consulting media outlets
57%53%60%8870.096
Consulting fact-checkers
36%31%35%8870.060
Doing nothing
6%12%7%8870.117
Source: Own elaboration
Note: Probabilities estimated from logistic regressions. All other variables in the model remain constant
Reference category in the proles: Traditional
Regarding reactions, it was found that the likelihood of consulting traditional media is slightly higher among omnichannel
proles (60%), followed by traditional proles (57%), while it is somewhat lower among moderate digital proles (53%). As
for consulting fact-checking agencies, the likelihood is generally lower, although there are still dierences between proles.
Traditional and omnichannel proles show similar probabilities (35–36%), while moderate digital proles show a lower
likelihood of checking information through this channel (31%). e pattern points to less institutionalization of fact-checking
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 365
among those who primarily consume news on social media. Finally, passivity in the face of disinformation (measured as the
likelihood of taking no action against it) is signicantly higher among moderate digital proles. e estimated probability of
this last group taking no action reaches 12%, which is double that observed among traditional proles (6%) and also higher
than that of omnichannel proles (7%).
5. Conclusions and discussion
e results shown in the previous section provide answers to the research questions posed in this study and allow testing of
the hypotheses proposed.
First, cluster analysis reveals that there are three distinct news consumption proles: one dominated by traditional media,
another characterized by moderate consumption primarily through digital and social media, and an omnichannel prole
with high levels of consumption across all media. is nding corroborates the rst hypothesis, which assumed that news
consumption proles are distinguished mainly by the degree of access to news through social media (H1). While the second and
third proles do consume news via the internet, the intensity with which the third does so is very dierent from the second, as
is its consumption of traditional media. In addition to diversity of sources, dierences were found in all the variables analysed,
allowing each prole to be characterized. Prole one is older, with primary education, little trust in digital media, and much
more trust in traditional sources; prole two is younger and from a lower social class, completed compulsory education, and
shows little trust or interest; while number three is middle-aged, has higher education, is upper-middle class, and shows the
greatest trust and interest.
e results are consistent with previous research in showing that digital proles are more common among young people. Six
out of ten respondents aged 15 to 24 belong to the group that primarily gets its news from social media. erefore, incidental
news consumption is more common in this age group, a nding also observed by Ceballos-del-Cid
et al.
(2025). However,
cluster analysis also reveals that 37.5% of young people belong to the group that is most actively engaged with news.
Regarding the second question, it has been observed that the probability of suspecting the presence of false content is higher
among those who use the internet more frequently for news, a not unexpected result given that social media is the primary
platform for disseminating that type of content, thus corroborating hypothesis H2a. However, hypothesis H2b is only partially
conrmed, since news consumption proles are only signicant in terms of inaction when encountering suspicious news
items
–
with the probability of passivity doubling in the moderate digital user group compared to traditional and intensive
omnichannel users
–
but not in terms of conrming via other media or consulting fact-checking agencies. Such passivity could
be explained by the low level of interest in news and politics that characterizes this moderate digital user prole. Indeed,
interest in news and politics are two of the strongest predictors of active reactions, suggesting that inaction is primarily a
consequence of disengagement from the news and not a deliberate strategy for managing the digital environment.
Finally, regression analyses have shown that sociodemographic variables are worse predictors of suspicion of having received
false news and of reactions to such suspicion than informational and political variables, thus conrming hypothesis H3. is
nding suggests that the sociodemographic dierences observed in previous studies may be mediated by variables more
366 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
directly related to the news ecosystem, factors that have been taken into consideration in this work, such as consumption
patterns, trust in dierent sources, and interest in politics.
e key sociodemographic variables, although none are relevant in all three reactions studied, are age (greater passivity
among those over 45 and less checking in traditional media among the over-60s) and educational levels. e results reveal
that graduates show a lower propensity to consult fact-checkers and a greater likelihood of doing nothing. e rst nding
observed in those with university degrees suggests greater condence in their own judgment of the trustworthiness of news.
e second may be related to this group feeling less social pressure against choosing this response (only 133 people in the
sample do so), which is likely socially disapproved. Sex, for its part, is only relevant in the case of checking news in the media,
such behaviour being less frequent among women than men. Other studies agree that sex dierences in the perception of and
reaction to disinformation are subtle and non-linear (Montiel
et al.
, 2025; Turel, 2024).
Political interest is the most inuential factor determining reactions when suspecting disinformation. It is a signicant variable
in all three cases: those who express an interest in politics are more likely to verify news in specialized media and are less likely
to do nothing or to check in traditional media. Conversely, an interest in news leads to greater consultation of media outlets
for conrmation. e crucial role of the media in fact-checking is evident in that it is the most frequently cited response by the
population surveyed (52.5%) when faced with suspected disinformation, a result that coincides with the DNR report (2025).
ese ndings suggest that managing disinformation in the digital public sphere falls primarily to the most politically engaged
and socially mobilized sectors, while a signicant portion of the population chooses inaction in the face of suspicion. ese
results match those obtained by the DNR (2025), which indicate that participation in public debate is linked to interest in
current aairs and to those who get news through social media. Future studies may wish to look more closely at the factors
examined here and the dynamics that currently characterize public discourse, employing qualitative techniques combined
with research based on the extraction of large amounts of data from social media.
e increase in news consumption through incidental news
–
reected in the moderate digital prole, which represents a
quarter of the population surveyed
–
challenges journalists’ traditional role as the arbiters of whether news is important or
not (Boczkowski
et al.
, 2017). is erosion of journalism is related not only to the expansion of social media use but to the
loss of trust in traditional media, which is even called out as a generator of disinformation (Montiel Torres
et al.
, 2025). at
perception leads to less news consumption through traditional media, but also decreased consumption through social media
and even through Articial Intelligence (Rasul
et al.
, 2025). us, a phenomenon of news avoidance emerges, whereby over a
third of the Spanish population voluntarily limits their exposure to news (DNR, 2025).
It is increasingly common to learn about the latest political news not through journalistic sources but through non-specialized
websites, as highlighted by the international study conducted by Wojcieszak
et al.
(2024), who conclude that of all visits to
political content, only one third come from news sites. News consumption is an essential component in the formation of
citizenship, but not enough attention has been paid to the fact that social media can potentially play a signicant role in young
people’s political socialization, something which has consequences in the public sphere (Bode, 2016). One consequence is
the polarization of political debate, particularly clear at election time (Teruel
et al.
, 2026). It is accentuated by algorithmic
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 367
dynamics, which prioritize emotional content that is swiftly shared and leads to the creation of echo chambers in which people
of diering opinions are increasingly less likely to interact with each other or to consume news from opposing viewpoints.
Finally, it goes without saying that the emergence of Generative Articial Intelligence (GAI) in recent years has been
another disruptive factor in how news is produced and consumed, making it increasingly dicult to distinguish truth from
falsehood (Sedano
et al.
, 2025). New GAI tools have led the European Union to regulate their use in this and other social
spheres (Fernández-Torres & Cea, 2025). According to a recent survey (Pastora-Estebanez & García- Faroldi, 2025), one in
ve Andalusians uses it to search for information and news. Furthermore, their assessment of the information provided is
very positive: four out of ten consider it highly trustworthy. However, more than half have a negative view of the use of GAI by
media professionals to write news articles (
ibid.
). Media literacy is increasingly needed to combat disinformation; GAI itself is
proving to be a powerful tool for achieving this goal (Zafra
et al.
2025).
ese results allow the identication of some specic practical implications. Firstly, the media literacy actions mentioned
should be targeted primarily at the moderate digital prole, characterized by little interest in news and politics and greater
passivity in the face of disinformation. Given that political interest emerges as the strongest predictor of active reactions,
those actions would be more eective if integrated with civic education, fostering not only media literacy but also engagement
with the public sphere. Secondly, the results point to the desirability of promoting a diversied news diet, since omnichannel
proles show a greater propensity to fact-check. Finally, the low rate of consultation of fact-checking sources, even among
those who actively react to suspected disinformation, suggests the need to increase the visibility and public accessibility of
specialized fact-checking agencies.
6. Funding
is work has been funded by the Excellence Project “Impact of disinformation in Andalusia: cross-sectional analysis of
audiences and journalistic routines and agendas (DesinfoAND)”, of the Andalusian Plan for Research, Development and
Innovation (Junta de Andalucía, PAIDI 2020, ProyExcel_00143); and by the national research project PID 2023-147486OB-
I00 (State Research Agency, Ministry of Science, Innovation and Universities, “Journalistic applications of AI to reduce
disinformation: trends, uses and perceptions of professionals and audiences” (DESINFOPERIA)).
is article has been translated into English by Brian O’Halloran, to whom we are grateful for his work.
368 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
7. Specic contributions of each author
Name
Conception and design of the workPablo Pastora Estebanez and Livia García-Faroldi
Methodology
Pablo Pastora Estebanez and Livia García-Faroldi
Data collection and analysisPablo Pastora Estebanez
Discussion and conclusionsPablo Pastora Estebanez and Livia García-Faroldi
Writing, formatting, revision and approval of versionsPablo Pastora Estebanez and Livia García-Faroldi
8. Conict of interest
e authors declare that there is no conict of interest regarding the publication of this article.
9. Bibliographic references
Anderson, K. J. (2025). What is News? Exploring Dierences in How Younger and Older Cohorts Use News in Today’s Media
Environment.
Communication Research and Practice
, 11(1), 3-23.
https://doi.org/10.1080/22041451.2024.2344997
Boczkowski, P. J., Matassi, M. & Mitchelstein, E. (2017). Incidental News: How Young People Consume News on Social
Media.
Proceedings of the 50th Hawaii International Conference on System Sciences
, 1785-1792.
https://doi.org/10.24251/
HICSS.2017.217
Bode, L. (2016). Political news in the news feed: Learning politics from social media.
Mass Communication & Society
,
19
(1),
24-48.
https://doi.org/10.1080/15205436.2015.1045149
Calinski, T. & Harabasz, J. (1974). A dendrite method for cluster analysis.
Communications in Statistics-theory and Methods
,
3
(1), 1-27.
https://doi.org/10.1080/03610927408827101
Catalina-García, B., García-Jiménez, A. & Paniagua-Santamaría, P. (2021). Percepción del consumo de noticias en línea y de
las prácticas relacionadas a ellas por parte de los jóvenes de la Comunidad de Madrid (España).
Cuadernos.info
, (50), 22-44.
https://doi.org/10.7764/cdi.50.27513
Ceballos-del-Cid, Y., Gómez-Calderón, B. & Córdoba-Cabús, A. (2025). Redes sociales y hábitos de consumo informativo de
los jóvenes españoles: un análisis diacrónico (2021-2023) [Social Media and News Consumption among Young Spaniards: A
Diachronic Analysis (2021-2023)].
Revista Mediterránea de Comunicación/Mediterranean Journal of Communication, 16
(1),
e28010.
https://www.doi.org/10.14198/MEDCOM.28010
Comisión Europea (2018).
La lucha contra la desinformación en línea: un enfoque europeo
. Comisión Europea, Bruselas.
https://eur-lex.europa.eu/legal-content/ES/TXT/?uri=CELEX:52018DC0236
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 369
Dubois, E. & Blank, G. (2018). e echo chamber is overstated: the moderating eect of political interest and diverse media.
Information, Communication & Society, 21(5), 729–745.
https://doi.org/10.1080/1369118X.2018.1428656
Feridouni Solimani, A. & Ahmed-Mohamed, K. (2024). Ciudadanía Digital: Niveles, Consumo y Conanza en la Información
de los Jóvenes Españoles.
Media & Jornalismo
, 45(9), 1-21.
https://doi.org/10.14195/2183-5462_45_9
Fernández-Torres, M.J. & Cea, N. (2025): “Europa ante el desafío de la desinformación en tiempos de inteligencia articial”.
En: L. Teruel & L. García Faroldi (Eds).
Los medios de comunicación ante la desinformación: inteligencia articial, discursos de
odio, teorías de la conspiración y vericación
”. Tirant lo Blanch, pp. 363-380.
https://riuma.uma.es/xmlui/handle/10630/39110
Fletcher, R. & Park, S. (2017). e impact of trust in the news media on online news consumption and participation.
Digital
Journalism
,
5
(10), 1281–1299.
https://doi.org/10.1080/21670811.2017.1279979
García-Faroldi, L., Teruel, L. & Blanco, S. (2025). Unmasking AI’s Role in the Age of Disinformation: Friend or Foe?
Journalism
and Media
,
6
(1), 19.
https://doi.org/10.3390/journalmedia6010019
Haller, A. & Holt, K. (2019). Paradoxical populism: How PEGIDA relates to mainstream and alternative media.
Information,
Communication & Society
, 22(12), 1665–1680.
https://doi.org/10.1080/1369118X.2018.1449882
Hameleers, M., Brosius, A. & de Vreese, C. H. (2022a). Whom to trust? Media exposure patterns of citizens with perceptions of
misinformation and disinformation related to the news media.
European Journal of Communication
,
37
(3), 237–268.
https://
doi.org/10.1177/02673231211072667
Hudders, L., De Jans, S. & De Veirman, M. (2020). e commercialization of social media stars: A literature review and
conceptual framework on the strategic use of social media inuencers.
International Journal of Advertising
,
40
(3), 327–375.
https://doi.org/10.1080/02650487.2020.1836925
Katz, E. (1957). e two‐step ow of communication: An up‐to‐date report on an hypothesis.
Public Opinion Quarterly
,
21
(1),
61–78.
https://doi.org/10.1086/266687
Katz, E., Haas, H. & Gurevitch, M. (1973). On the use of the mass media for important things.
American Sociological Review
,
38
(2), 164–181.
https://doi.org/10.2307/2094393
Katz, E., Blumler, J. G. & Gurevitch, M. (1974). Utilization of mass communication by the individual. En J. G. Blumler & E.
Katz (Eds.),
e uses of mass communications: Current perspectives on gratications research
(pp. 19–31). Beverly Hills: Sage.
https://archive.org/details/usesofmasscommun0000unse/page/8/mode/2up
Lazarsfeld, P. F., Berelson, B. & Gaudet, H. (1944).
e people’s choice: How the voter makes up his mind in a presidential
campaign
. Duell, Sloan and Pearce.
https://archive.org/details/peopleschoicehow0000laza
Lowenstein-Barkai, H. & Lev-on, A. (2022) News videos consumption in an age of new media: a comparison between
adolescents and adults,
Journal of Children and Media
, 16:1, 78-94.
https://doi.org/10.1080/17482798.2021.1915831
McQuail, D., Blumler, J. G. & Brown, R. (1972). e television audience: A revised perspective. In D. McQuail (Ed.),
Sociology of
mass communication
. Middlesex: Penguin.
https://cir.nii.ac.jp/crid/1570854176052168448
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
370 |
nº 43, pp. 347-371
|
July-December of 2026
News consumption proles facing disinformation: exposure and responses to suspect information
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Milligan, G. W. & Cooper, M. C. (1985). An examination of procedures for determining the number of clusters in a data set.
Psychometrika
, 50(2), 159-179.
https://doi.org/10.1007/BF02294245
Montiel Torres, M. F.; Teruel Rodríguez, L.; García-Faroldi, L. & Martín-Martín, F. M. (2025). Colectivos vulnerables y
desinformación. Análisis de la realidad andaluza.
Doxa Comunicación
, 41, pp. 369-393.
https://doi.org/10.31921/doxacom.
n41a2902
Pariser, E. (2011).
e lter bubble: How the new personalized web is changing what we read and how we think
. London: Penguin.
Park, S., Fisher, C., Flew, T. & Dulleck, U. (2020). Global mistrust in news: e impact of social media on trust.
e International
Journal on Media Management
,
22
(2), 83–96.
https://doi.org/10.1080/14241277.2020.1799794
Pastora Estebanez, P. & García-Faroldi, L. (2025). ¿Nueva herramienta informativa o generadora de bulos? Percepción
ciudadana de la Inteligencia Articial Generativa en la era de la desinformación.
Cuadernos del Audiovisual | CAA
, (14).
https://doi.org/10.62269/cavcaa.59
Peter, C. & Muth, L. (2023). Social Media Inuencers’ Role in Shaping Political Opinions and Actions of Young Audiences.
Media and Communication, 11 (3), Pages 164–174.
https://doi.org/10.17645/mac.v11i3.6750
Punj, G. & Stewart, D. W. (1983). Cluster analysis in marketing research: Review and suggestions for application.
Journal of
Marketing Research
,
20
(2), 134-148.
https://doi.org/10.1177/002224378302000204
Rasul, M. E., Calabrese, C., Oh, Y. J., Cho, H. J., Jeon, M. & Boukes, M. (2025). “It’s All Fake News!”: How Perceptions of
Misinformation and Disinformation Inuence News Consumption Across Traditional Media, Social Media, and AI.
Journalism
& Mass Communication Quarterly
, 102(4), 993-1019.
https://doi.org/10.1177/10776990251373085
Sedano, J., Blanco, S. & Palomo, B. (2025): “Imagen y desinformación en la era de la inteligencia articial”. En: L. Teruel & L.
García Faroldi (Eds).
Los medios de comunicación ante la desinformación: inteligencia articial, discursos de odio, teorías de la
conspiración y vericación
, Tirant lo Blanch, pp. 337-361.
https://riuma.uma.es/xmlui/handle/10630/39110
Shao, C., Ciampaglia, G. L., Varol, O., Flammini, A. & Menczer, F. (2017). e spread of low-credibility by social bots. arXiv
preprint arXiv:1707.07592, 96, 104. ArXiv e-prints.
https://doi.org/10.48550/arXiv.1707.07592
Shu, K., Sliva, A., Wang, S., Tang, J. & Liu, H. (2017). Fake news detection on social media: A data mining perspective
. ACM
SIGKDD Explorations Newsletter
, 19(1), 22–36.
https://doi.org/10.48550/arXiv.1708.01967
Shu, K., Bhattacharjee, A., Alatawi, F., H. Nazer, T.H., Ding, K., Karami, M. & Liu, H. (2020). Combating disinformation in a
social media age.
WIREs Data Mining Knowl Discov
., 10:e1385.
https://doi.org/10.1002/widm.1385
Stehr, P., Rössler, P., Leißner, L. & Schönhardt, F. (2015). Parasocial opinion leadership media personalities’ inuence within
parasocial relations: eoretical conceptualization and preliminary results.
International Journal of Communication
,
9
, 982–
1001.
https://ijoc.org/index.php/ijoc/article/view/2717
Terren, L. & Borge, R. (2021). Echo Chambers on Social Media: A Systematic Review of the Literature.
Review of Communication
Research
,
9
, 1–39.
https://doi.org/10.12840/ISSN.2255-4165.028
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 43, pp. 347-371
July-December of 2026
Pablo Pastora Estebanez and Livia García-Faroldi
ISSN: 1696-019X / e-ISSN: 2386-3978
| 371
Teruel, L., Blanco, S., Cea Esteruelas, N. & Congosto, M. (2026). Polarización en Twitter/X: análisis del debate generado por los
medios de comunicación durante las elecciones locales y autonómicas del 28-M-2023 en España.
REDES, Revisa Hispana para
el Análisis de Redes Sociales
, 37 (2).
https://doi.org/10.5565/rev/redes.1133
Tucker, J. A., Guess, A., Barberá, P., Vaccari, C., Siegel, A., Sanovich, S., Stukal, D. & Nyhan, B. (2018).
Social Media, Political
Polarization, and Political Disinformation: A Review of the Scientic Literature
.
https://doi.org/10.2139/ssrn.3144139
Turel, O. (2024). Polarized sharing of fake news on social media: the complex roles of partisan identication and gender.
Behaviour & Information Technology
,
43
(11), 2424–2441.
https://doi.org/10.1080/0144929X.2023.2248282
Vosoughi, S., Roy, D. & Aral, S. (2018). e spread of true and false news online.
Science
, 359(6380), 1146–1151.
https://doi.
org/10.1126/science.aap9559
Ward, J. H., Jr. (1963). Hierarchical grouping to optimize an objective function.
Journal of the American Statistical Association
,
58
(301), 236-244.
https://doi.org/10.1080/01621459.1963.10500845
Wojcieszak, M., Menchen-Trevino, E., Clemm von Hohenberg, B., de Leeuw, S., Gonçalves, J., Davidson, S. & Gonçalves, A.
(2024). Non-News Websites Expose People to More Political Content an News Websites: Evidence from Browsing Data in
ree Countries.
Political Communication
,
41
(1), 129–151.
https://doi.org/10.1080/10584609.2023.2238641
Zafra Arroyo, A., Sánchez González, M. & Sánchez Gonzales, H. M. (2025). Innovaciones con IA generativa para alfabetización
y vericación en la Unión Europea.
Doxa Comunicación. Revista Interdisciplinar De Estudios De Comunicación Y Ciencias
Sociales
, 41, 489-509.
https://doi.org/10.31921/doxacom.n41a2874
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>