Migrants and refugees on Twitter in Spain: a study employing
automated analysis into the presence of hate and sentiment
Migrantes y refugiados en Twitter en España: estudio de la presencia de odio y
del sentimiento a partir de un análisis automatizado
doxa.comunicación |
nº 38, pp. 369-389
| 369
January-June of 2024
ISSN: 1696-019X / e-ISSN: 2386-3978
How to cite this article:
Barradas Gurruchaga, A.; Blanco-Herrero, D.; Arcila-Calderón, C. and Sánchez-Holgado, P. (2024). Migrants and refugees on Twitter in Spain: a
study employing automated analysis into the presence of hate and sentiment.
Doxa Comunicación
, 38, pp. 369-389.
https://doi.org/10.31921/doxacom.n38a1734
David Blanco-Herrero.
Postdoctoral researcher at the University of Amsterdam. He has a PhD from the Universidad
de Salamanca, a Master’s in Audio-visual Communication (Universidad de Salamanca), and degrees in Journalism
(Universidad a Distancia de Madrid) and Business Administration (Universidad de León). He is a member of the
Amsterdam School of Communication Research (ASCoR) and his main lines of research are in journalistic ethics,
disinformation and hate speech. He is an editor of the Electronic Yearbook on Studies in Social Communication
“Disertaciones”. He has also worked as a journalist and collaborated with radio and digital media, particularly on
international aairs and audio-visual culture.
University of Salamanca, Spain / University of Amsterdam, the Netherlands
[email protected]
/
[email protected]
ORCID: 0000-0002-7414-2998
Carlos Arcila-Calderón.
Full Professor in the Department of Sociology and Communication of the Universidad de
Salamanca (Spain). Member of the Audio-visual Content Observatory (OCA) and Professor of the Doctorate in Training
in the Knowledge Society. Editor of the Electronic Yearbook on Studies in Social Communication “Disertaciones”. He
has a European Doctorate in “Communication, Social Change and Development” from the Complutense University
of Madrid. Master’s in Data Science and a Master’s in Journalism, both from the Universidad Rey Juan Carlos (URJC).
University of Salamanca, Spain
[email protected]
ORCID: 0000-0002-2636-2849
Andrés Barradas Gurruchaga.
PhD in Communication from the Universitat Autònoma de Barcelona. Full-time Professor
at the Tecnológico de Monterrey (Mexico). He has conducted research into areas of audio-visual communication and
the creative industries, contributing to scientic papers and books. Both scriptwriter and director of short lms shown
at lm festivals such as Cannes. He has addressed audiences in Latin America, the US and Spain. A member of the
Mexican Association of Communication Researchers and of the Spanish Association of Communication Research. He
spent a sabbatical research period at the Universidad de Salamanca, Spain. He has been the Director of the degree in
Communication and Digital Media, Regional Director of the Department of Creative Industries, Research Professor, and
Post-graduate Professor of the Master’s in Finance at the EGADE Business School.
Tecnológico de Monterrey, Mexico
[email protected]
ORCID: 0000-0001-9020-6659
is content is published under Creative Commons Attribution Non-Commercial License. International License CC BY-NC 4.0
<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 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Received: 02/07/2023 - Accepted: 08/09/2023 - Early access: 23/10/2023 - Published: 01/01/2024
Recibido: 02/07/2023 - Aceptado: 08/09/2023 - En edición: 23/10/2023 - Publicado: 01/01/2024
Abstract:
Hate speech directed towards migrants and refugees poses one of
the greatest challenges in communication on social networks. By
means of an automated analysis of 124,337 messages about migration
collected on Twitter in Spain between 2015 and 2020, the presence of
hate and the underlying sentiment in the discourse are analysed, as
is their development and the possible dierences between Spanish
regions. It has been observed that, although greater attention was paid
to migration in 2015 and 2016, the core years of the Mediterranean
refugee crisis, the greatest volume of hate was detected in 2019 and
2020, following the rise of Vox and its anti-immigration rhetoric. In
general, the sentiment in these messages was negative, although the
dierence is slight between those expressing hate and those which do
not. Finally, dierences have been observed between Spanish regions,
with Asturias having the greatest presence of hate and Cantabria
showing the most negative sentiment; such dierences, however, are
not great and no clear patterns have been detected to explain them.
Keywords:
Immigration; racism & xenophobia; hate speech; social networks;
sentiment analysis; Twitter.
Resumen:
El discurso de odio dirigido hacia personas migrantes y refugiadas plan
-
tea uno de los mayores desafíos en la comunicación en redes sociales.
A través del análisis automatizado de 124.337 mensajes sobre migra
-
ción recogidos en Twitter en España entre 2015 y 2020, se analiza la
presencia de odio y el sentimiento subyacente en el discurso, así como
su evolución y las posibles diferencias entre comunidades autónomas.
Se ha observado que, aunque la atención prestada a la cuestión migra
-
toria fue superior en 2015 y 2016, años centrales de la crisis de refugia
-
dos del Mediterráneo, el mayor volumen de odio se detectó en 2019 y
2020, tras el ascenso de Vox y su retórica anti-inmigración. En general,
el sentimiento de estos mensajes fue negativo, si bien la diferencia entre
aquellos que tienen odio y los que no es reducida. Por último, se han
observado diferencias entre regiones autónomas, siendo Asturias la de
mayor presencia de odio y Cantabria la que muestra un sentimiento
más negativo; estas diferencias, no obstante, no son elevadas y no se han
detectado patrones claros que las expliquen.
Palabras clave:
Inmigración; racismo y xenofobia; discurso de odio; redes sociales; aná
-
lisis de sentimientos; Twitter.
Patricia Sánchez-Holgado.
Researcher at the Universidad de Salamanca and member of the Audio-visual Content
Observatory. Graduate in Advertising and Public Relations (Complutense University of Madrid). She is an Associate
Professor at the Faculty of Languages and Education in Nebrija University in Madrid and in the Faculty of Communication
at the Universidad Ponticia de Salamanca. An expert on Big Data (Universidad Ponticia de Salamanca) with a Master’s
degree in Science, Technology and Innovation Studies (Universidad de Oviedo).
University of Salamanca, Spain
[email protected]
ORCID: 0000-0002-6253-7087
1. Introduction
Since the late XXth century Spain has been considered to be a country that receives foreigners (Lacomba et al., 2020), a
phenomenon that has generated press interest and presence since the 1990s and 2000s (Seoane-Pérez, 2017). Migratory
pressure on the European Union has increased more recently, becoming a central aspect of political and media discourse
(Greussing & Boomgarden, 2017).
Although the causes had come earlier, it was in 2015 that migratory ows towards the European Union reached a volume
and attention rarely seen (Splinder, 2015), especially after the discovery of the lifeless body of the young boy Aylan Kurdi in
September on a Turkish beach, an event with considerable media impact (Mielczarek, 2018), which represented a turning
point in coverage of the crisis (Zhang & Hellmueller, 2017). roughout 2016, migratory ows towards the European Union
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 371
began to shift to the central Mediterranean following an agreement between the EU and Turkey, which reduced the number
of arrivals through Greece, until that point the country with the highest number of immigrants and asylum seekers. Both
the election of Donald Trump in the United States and the victory of Brexit in the United Kingdom, two events with strong
nationalist and anti-immigration components (Joppke, 2020), caused immigration as an issue to remain in the media’s eye.
e Strait of Gibraltar began to reappear as a route in 2017, and 2018 saw Spain become the European country with the most
refugees and migrants arriving by sea, mainly along the coasts of Andalusia, via the eastern Mediterranean route, in part after
the tighter measures imposed by the Italian government and, in particular, by the Minister of the Interior, Matteo Salvini. In
June, the Spanish government decided to accept 630 migrants on board the Aquarius in the port of Valencia after the refusal of
their Italian counterparts (CEAR, 2019). ere was a decrease in the number of arrivals to Spain in 2020 as a result of the health
crisis and, in general, in the international movement of people, though this increased again intensely in the latter months of
the year, especially to the Canary Islands (CEAR, 2021).
ese events, apart from inuencing the social and political situation, had certain protagonism in the media. In fact, there
is widespread agreement in academia about the media’s treatment of migration, which tends to be negative, simplistic,
stereotyped and with a lack of focus on individual stories (Igartua et al., 2007; Muñiz, Igartua & Otero, 2006; Fajardo Fernández
& Soriano Miras, 2016; Eberl et al., 2018), something which is also applicable to coverage of the refugee crisis from 2015 on
(Greussing & Boomgaarden, 2017; Brändle, Eisele &Trenz 2019; Fengler et al., 2020). is type of representations reinforces
the association of immigration and foreigners with out-groups, according to Tajfel and social identity theory (1978), in turn
generating less acceptance of immigrants (Esses et al., 2005).
It is true that attitudes towards immigration in Spain tend to be among the most favourable in the European Union, as
conrmed by several Eurobarometers (European Commission, 2019; 2022) and international surveys (Pew Research Center,
2021). However, the presence of hate speech and rejection of migrants and refugees on social networks is a matter of growing
concern (Arcila Calderón, Blanco-Herrero & Valdez Apolo, 2020), partly as a result of the increase in numbers of hate crimes
against the group (OSCE, sf).
at is why this study seeks to analyse public discussion in Spain on the social network Twitter between 2015 and 2020, paying
particular attention to the presence of hate and the underlying sentiments in said discourse. e study aims to continue
contributing to the still small amount of research that use computational techniques to address the issue, oering an analysis
of a large corpus of data. Furthermore, the study aims to ll the existing gap regarding longitudinal approaches that allow
observation of the phenomenon’s development over time. Such studies, in addition to the scientic knowledge they provide,
are key to proposing possible strategic actions a posteriori to combat hate and rejection, both online and in other settings.
2. Hate speech towards migrants and refugees
As noted in the previous section, the presence of expressions of hate, discrimination and rejection has become a phenomenon
of some concern, fuelled by anti-immigration discourse (Arcila-Calderón, de la Vega & Blanco-Herrero, 2020), due to generally
negative coverage in the media (Schemer, 2012) and by the communication model itself on social networks, an environment
in which these types of expressions have increased exponentially in comparison with physical space (Rollnert Liern, 2020).
372 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Focusing on hate speech, this can be dened as apology, promotion or incitement in any form, denigration, hatred or
vilication on the part of a person or group of people, as well as harassment, insult, negative stereotyping, stigmatisation or
threat regarding one or more people (European Commission against Racism & Intolerance, 2005). us, hate speech attacks
the dignity of the defamed communities, which are deprived of their right to be considered t for society (Díaz Soto, 2015).
Added to this is the fact that hate speech may act as a precursor to violent hate crimes (Müller & Schwarz, 2020), exacerbating
the problem.
at is why numerous public and private institutions are striving to stop or combat hate speech, especially on social networks
(Andres & Slivko, 2021). In the particular case of
Twitter,
its rules establish that “you may not encourage violence against other
people or attack or threaten them directly (Twitter, 2022).”
Hate speech with racist or xenophobic motivation is of particular concern, in addition to its signicant presence on digital
platforms, it is the commonest type of hate crime registered both in Spain and in the majority of neighbouring countries
(OSCE, n.d.). It is precisely discourse of this type, which is largely directed at migrants and refugees
1
, that will be examined in
this study.
Content published on Twitter will be used for this analysis, given that, although it is not the most widely used network in
Spain, it is, due to its characteristics, one of the most inuential in public discourse, as it is the most relevant and widely used
by journalists and media, as well as by politicians and institutions (Rodríguez & Ureña, 2011; Campos-Domínguez, 2017).
Furthermore, Twitter, like other social networks, oers a valuable source of information for the analysis of public opinion and
citizens’ attitudes, even overcoming some of the limitations posed by survey-based studies (Arcila Calderón, Blanco-Herrero
& Valdez Apolo, 2020). Moreover, this network, with its easy virilisation of content, its popularity, simplicity of discourse
and speed of communication displays elements that summarise the attributes of Web 2.0 (Ausserhofer & Maireder, 2013),
along with the capacity for dissemination oered by retweets, which together make this the ideal platform for disseminating
information and opinion (Moragas-Fernández, Grau-Masot & Capdevila-Gómez, 2019).
is also contributes to the tendency on Twitter to create homogeneous communities that become echo chambers (Gruzd &
Roy, 2014), where individuals are only exposed to information that conrms their ideas, contributing to radicalisation (Yardi
& Boyd, 2010). And partly derived from that, the network is considered one of the most problematic for the dissemination of
rejection discourse, both by
“the public repercussion of what is disseminated on this social network, and which increases the social alarm of the discourse
expressed in it, as well as the greater methodological simplicity that observation on this social network entails compared to others
1
Although no distinctions will be made in this text, it should be pointed out that there are dierences between the terms ‘migrants’ and ‘refugees’. Since the 1951
Geneva Convention relating to the Status of Refugees, a refugee is considered to be a “person who, due to well-founded fear of being persecuted for reasons of
race, religion, nationality, membership of a particular social group or political opinions, is outside the country of their nationality and is unable or, because of
such fears, unwilling to avail themselves of the protection of that country; or who, lacking nationality and nding themselves, as a result of such events, outside
the country of their former habitual residence, cannot or, due to such fears, is unwilling to return to it” (International Organization for Migration, 2019) . e IOM
includes in the same glossary the word ‘migrant’ as “an umbrella term not dened under international law reecting the common lay understanding of a person
who moves away from his or her place of usual residence, whether within a country or across an international border, temporarily or permanently, and for a
variety of reasons” (International Organization for Migration, 2019).
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 373
(Facebook, social forums) in which “censorship” of violent discourse is swifter and in which it is more dicult to access a large
and homogeneous sample of discourse […]. Short messages are published on Twitter that, although they may contain complex
information, are unlikely to contain the nuances of communication found in other places where this limitation does not exist. Since
the objective of this study is to dene a basic taxonomy of violent discourse, I consider it appropriate to focus analysis on this type of
more basic messages” (Miró Llinares, 2016, p. 86).
For these reasons, the study of discourse on this platform focused on migrants and refugees is of special relevance, as shown
by previous studies (Kreis, 2017). Before addressing hate speech, it is worth evaluating the attention generated by migratory
phenomena. It has been considered that interest in the Mediterranean refugee crisis in Spain came somewhat late (Seoane-Pérez,
2017), and that the rise of Vox, with its anti-immigration discourse, led to an increase in interest in this matter in political discourse
(Castromil, Rodríguez-Díaz & Garrigós, 2020). At the same time, the prominence of immigration in the media has been considered
a factor which may increase the appeal of anti-immigration parties (Damstra et al., 2019; Dennison & Geddes, 2019; Krzy
ż
anowski,
Triandafyllidou & Wodak, 2018). erefore, and by way of contextualisation, we pose the following research question:
RQ1: How did the presence of content about migrants and refugees on Twitter in Spain change between 2015 and 2020?
In line with the above, it is known that rejection and hatred are not immutable categories but are marked by social and media
events. In fact, news events have been observed to aect both the quantity and characteristics of hate speech directed at migrants
and refugees (Arcila-Calderón et al., 2021). At the same time, one must take into account the fact that citizens’ attitudes are
not unchanging, but uctuate over time, as reected in periodic surveys such as Eurobarometer. And the same is true of media
coverage of an issue, in this case migration (Amores, Arcila-Calderón & Blanco-Herrero, 2020; Zhang & Hellmueller, 2017).
For this reason, it is necessary to consider what eects oine events may have had on the volume of hate speech expressed on
Twitter towards migrants and refugees, which is why the following research question is posed:
RQ2: How did the presence of hate speech towards migrants and refugees on Twitter in Spain change between 2015 and
2020?
e existence of dierences between countries is evident, but also within a country itself, as reected in the study by Arcila-
Calderón et al. (2022) on which this study is based. In order to examine more closely any possible dierences between the
dierent regions of Spain, both regarding the presence of this issue and the proportion of hate speech, the following research
questions are posed:
RQ3.1: In which Spanish regions was there a greater presence of content about migrants and refugees on Twitter between
2015 and 2020?
RQ3.2: In which Spanish regions was there a greater presence of hate speech towards migrants and refugees on Twitter
between 2015 and 2020?
Finally, numerous studies have used sentiment analysis to address the representation of migrants and refugees, both in the
mass media (Backfried & Shalunts, 2016) and on social networks (Heidenreich et al., 2020). e study of feelings associated
with the discourse around migrants and refugees –the connection of which to hate speech has already been identied by
Arcila-Calderón et al. (2021)– allows a greater understanding of the attitudes and forms of expression that surround this
374 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
matter. Such analysis of feelings in language is important due to the role that negativity or the presence of incivility has in
radicalising and polarising discourses (Robles et al., 2022). Hameleers, van der Meer and Vliegenthart (2022) have analysed
the convergence between negativity, incivility, hate speech and misinformation, adding layers to the problem. De León and
Trilling (2021) have found that political content with negative tones is more frequently shared and achieves greater reach.
And Yordakul (2021) even points out that negative emotions have a greater eect on political decision-making. at is why the
following research question is posed:
RQ4.1: Was there a correlation between the presence of hate and sentiment expressed in content about migrants and
refugees on Twitter in Spain between 2015 and 2020?
RQ4.2: How did the sentiment expressed in content about migrants and refugees on Twitter in Spain change between 2015
and 2020?
RQ4.3: Were there dierences between the Spanish regions in relation to the sentiment expressed in content about migrants
and refugees on Twitter between 2015 and 2020?
3. Methodology
In order to achieve these objectives, the database of tweets used in the work of Arcila-Calderón et al. has been utilised. (2022).
Tweets focused on the phenomena of migration and asylum were compiled between 2015 and 2020 using the Academic
Research product track that Twitter’s Application Programming Interface (API) makes available to academic researchers. In
order to locate the contents, the keywords ‘migrant’, ‘migrants’, ‘immigrant’, ‘immigrants’, ‘refugee’, ‘refugees’, ‘asylum seeker’
and ‘asylum seekers’ were employed. e next step was to exclude retweets and responses and only select geolocated content,
this was so that local analyses could be carried out with greater reliability; in fact, the use of this type of messages is one
of the main strengths of the study, since, although it signicantly reduces the amount of content available, it allows us to
guarantee the quality of the analysis, knowing its location with certainty and being able to make regional comparisons. Finally,
the downloaded contents were analysed utilising a previously validated automatic hate detection tool (Vrysis et al., 2021) and
again reset for the work of Arcila-Calderón et al. (2022).
Only Spanish data has been utilised for this study, a total of 124,337 tweets, which were passed through the detector to determine
the presence of hate. e presence or absence of hate was measured using a scale between 0 and 1, with 0 representing absence
and 1 the presence of hate.
Similarly, in order to respond to RQ4, sentiment in the messages analysed was evaluated between -1, negative, and 1, positive.
An automatic classication tool based on the SentiStrength algorithm, developed by elwall et al. (2010), was also used for
the same purpose. Although a study of sentiment was also carried out utilising a tool developed from a lexicon composed
ad hoc, the tool based on SentiStrength was chosen due to its generalised validity and for allowing easier replicability and
comparison. e option of employing other scales of sentiment that could lead to greater variety and detail in the analysis was
also evaluated –for example, through the basic emotions developed by Fernández-Abascal, Palmero and Breva (2002) and
Chóliz and Gómez ( 2002)–, however, a binary classication between positivity and negativity was nally chosen, since it is
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 375
precisely negativity, and not a specic type of feeling, that has been associated with phenomena such as polarisation (Buder et
al., 2021) or with a greater aective reaction (Soroka & McAdams, 2015). It is precisely these connections that our work seeks
to evaluate, since they are what may lead to an accentuation or radicalisation of hate speech. is type of study also promotes
the use of the SentiStrength tool, which is widely used internationally and has a greater capacity for replication.
Once the presence of hate and the underlying sentiment in the 124,337 items in the study had been considered, metadata such
as the date of publication or the place of publication (NUTS2, which corresponds to the regions of Spain) was also identied,
which allowed us to examine the temporal and spatial distribution of both hate and sentiment in greater depth. With this,
various statistical tests were carried out to answer the research questions posed, mainly descriptive and frequency analyses
and median comparison tests.
4. Results
4.1. Presence of content relative to migrants and refugees
As indicated above, the 124,337 geolocated messages were obtained over a period of six years, between 2015 and 2020. e
total number of messages collected each year was: in 2015, 24,001; in 2016, 28,614; in 2017, 17,246; in 2018, 19,963; in 2019,
19,337; and in 2020, 15,176. Figure 1 shows graphically that there are dierences, and that the number of messages in 2016 is
close to double that of 2020, which may indicate greater interest in the issue.
Figure 1. Annual number of messages about migrants and refugees on Twitter in Spain between 2015 and 2020
Figura 1. Evolución anual de
los
mensajes sobre migrantes y refugiados en Twitter en España
entre 2015 y 2020
Fuente: elaboración propia
La Figura 2, por su parte, permite observar la cantidad de mensajes sobre migrantes y
refugiados recogidos cada día, entre el 1 de enero de 2015 y el 31 de diciembre de 2020. Se
observan una serie de picos, el mayor de ellos en septiembre de 2015, momento en el que las
llegadas de solicitantes de asilo a la Unión Europeo se convirtieron en un foco de interés
mediático
, en gran medida tras la publicación de la foto de Aylan Kurdi ahogado en una playa
turca
. La cantidad de mensajes se mantuvo alto en 2016, con un importante pico en marzo de
ese mes, coincidiendo con la firma de un acuerdo entre la Unión Europea y Turquía para
evitar las llegadas masivas a Grecia. En junio de 2018 se alcanza otro pico importante a raíz
de lo sucedido con el barco Aquarius, que acabó desembarcando en España. En 2020 se
observan valores generalmente bajos, sobre todo a causa de la pandemia, que además de
frenar durante muchos meses casi todos los movimientos internacionales, concentró
notablemente la atención en los temas de salud, en detrimento de cualquier otra materia.
Ϭ
ϱϬϬϬ
ϭϬϬϬϬ
ϭϱϬϬϬ
ϮϬϬϬϬ
ϮϱϬϬϬ
ϯϬϬϬϬ
ϯϱϬϬϬ
ϮϬϭϱϮϬϭϲϮϬϭϳϮϬϭϴϮϬϭϵϮϬϮϬ
Source: created by the authors
376 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Figure 2 allows us to observe the number of messages about migrants and refugees collected each day, between January 1,
2015, and December 31, 2020. A series of peaks can be seen, the largest of them in September 2015, at which point in time the
arrivals of asylum seekers to the European Union became a focus of media interest, largely after the publication of the photo of
Aylan Kurdi, dead by drowning on a Turkish beach. e number of messages remained high in 2016, with a signicant peak in
March of that year, coinciding with the signing of an agreement between the European Union and Turkey to prevent massive
arrivals in Greece. Another major peak was reached in June 2018 as a result of what happened with the Aquarius, a ship
which ended up disembarking in Spain. Generally low values are observed in 2020, principally due to the pandemic, which in
addition to stopping almost all international movement for many months, notably focused attention on health issues, to the
detriment of all other matters.
Figure 2. Daily number of messages about migrants and refugees on Twitter in Spain between 2015 and 2020
Figura 2. Evolución diaria de
los
mensajes sobre migrantes y refugiados en Twitter en España
entre 2015 y 2020
Fuente: elaboración propia
4.2. Presencia de discurso de odio contra migra
n
tes y refugiados
Centrándonos ahora en la presencia de odio, debemos empezar señalando que de los
124.
337
tweets recogidos sobre el tema, un total de 27
.468
tweets incluyeron muestras de odio y
rechazo superiores al 0,5, lo que podría ya considerarse una forma de discurso de odio
consistente.
Esto supone un 22,
1
%, es decir, más de una quinta parte de la conversación
alrededor de cuestiones migratorias en Twitter en España. Al mismo tiempo, la media de
discurso de odio presente en el conjunto de la muestra es de 0,2
87
(DT=0,264).
S
e puede observar que esta proporción tampoco es estable, y varía con el tiempo, como se
muestra en la Figura 3. De hecho, s
i
realizamos la prueba Anova de un factor para muestras
independientes, comprobamos que las diferencias son estadísticamente significativas; así, el
estadístico F de Welch es significativo [F(5,
54378,503)=90,863
, p<0,001], mientras las
pruebas post
-
hoc realizadas (T3 de Dunnett) indican que la media de odio en el año 2017
(M=0,256; DT=0,251) fue significativamente inferior al resto de
años
, mientras que 2020
(M=0,303; DT=0,76), 2019 (M=0,301; DT=0,277) y 2015 (M=0,300; DT=0,262) tuvieron
una media de odio significativamente superior al resto de años, con 2018 (M=0,276;
DT=0,262) y 2016 (M=0,283; DT=0,255) ofreciendo valores intermedios.
Se puede apreciar que la presencia de odio es mayor en los años iniciales y, sobre todo,
finales, con menor relevancia en los años centrales de la muestra. Los valores iniciales
pueden explicarse por la prominencia que el tema migratorio tuvo en los años 2015 y 2016,
en los que los medios informaban de grandes volúmenes de personas, a menudo retratadas
Ϭ
ϭϬϬ
ϮϬϬ
ϯϬϬ
ϰϬϬ
ϱϬϬ
ϲϬϬ
ϳϬϬ
ϴϬϬ
ϬϭͲĞŶĞͲϭϱϬϭͲĞŶĞͲϭϲϬϭͲĞŶĞͲϭϳϬϭͲĞŶĞͲϭϴϬϭͲĞŶĞͲϭϵϬϭͲĞŶĞͲϮϬ
Source: created by the authors
4.2. Presence of hate speech against migrants and refugees
Turning now to the presence of hate, it is important to begin by pointing out that of the 124,337 tweets collected on the issue,
a total of 27,468 tweets included samples of hate and rejection greater than 0.5, which could already be considered a form of
consistent hate speech. is represents 22.1%, that is, over a fth of the conversation around migration issues on Twitter in
Spain. At the same time, the average value for hate speech present in the whole sample is 0.287 (SD=0.264).
It can be seen that this proportion is not stable, and varies over time, as shown in Figure 3. In fact, carrying out the one-
way Anova test for independent samples veries that the dierences are statistically signicant; us, Welch’s F statistic is
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 377
signicant [F(5, 54378.503)=90.863, p<0.001], while the post-hoc tests carried out (Dunnett’s T3) indicate that the average hate
rate in 2017 (M =0.256; SD=0.251) was signicantly lower than the other years considered, while 2020 (M=0.303; SD=0.76),
2019 (M=0.301; SD=0.277) and 2015 (M=0.300; SD=0.262) had a signicantly higher average hate rate than the other years, with
2018 (M=0.276; SD=0.262) and 2016 (M=0.283; SD=0.255) oering intermediate values.
It can be seen that the presence of hate is greater in the early and, above all, later years, with less weight in the central years
of the sample. e initial values can be explained by the prominence that migration had as an issue in 2015 and 2016, when
the media reported large volumes of people, often portrayed as masses (Amores, Arcila-Calderón & Blanco-Herrero, 2020),
arriving in the European Union. Recent years, although general interest has declined, and contrary to what might be expected,
the values have seen an increase, which in 2019 and 2020 coincided with the rise of Vox, a party with a nationalist and anti-
immigration discourse, causing this type of discourse, which until then had been of little importance, to gain weight in Spain
(Ferreira, 2019; Turnbull-Dugarte, 2019).
Figure 3. Annual changes in the presence of hate speech towards migrants and refugees on Twitter in Spain between 2015 and 2020
como masas (Amores, Arcila
-
Calderón & Blanco
-
Herrero, 2020), llegando a la Unión
Europea. Los últimos años, si bien el interés general ha decaído, y en contra de lo que cabría
esperar, se produce un aumento, que en los años 2019 y 2020 coincide con el ascenso de
Vox, un partido con un discurso nacionalista y anti
-
inmigración, haciendo que en España
ganasen peso este tipo de discursos, que hasta entonces habían resultado poco relevantes
(Ferreira, 2019;
Turnbull-
Dugarte, 2019)
.
Figura 3. Evolución anual de la presencia de discurso de odio hacia migrantes y refugiados en
Twitter en España entre 2015 y 2020
Fuente: elaboración propia
4.3. Diferencias regionales
Hay que señalar, en primer lugar, que la mayor atención al tema se da en las regiones más
pobladas, de manera que Madrid, Cataluña y Andalucía acumulan una mayor cantidad de
mensajes sobre migrantes y refugiados. La Figura 4 resume la cantidad de contenidos sobre
el tema recogidos en cada comunidad autónoma. Sin embargo, esta distribución no se
corresponde con la proporción de odio detectado, donde las regiones con mayor odio son
Asturias
y
Castilla-
La Mancha. Las pruebas de contraste de medias
muestran
que
existen
diferencias significativas entre las regiones españolas [F(17, 13157,451)=27,802, p<0,001],
y las pruebas post hoc confirman estas diferencias. Así, la media de odio detectada en
Asturias (M=0,323, DT=0,285) y en Castilla
-
La Mancha (M=0,318, DT=0,
280)
es
significativamente superior a la detectada en Comunidad Valenciana (M=0,295, DT=0,268),
Islas Canarias (M=0,292, DT=0,264), Madrid (M=0,282, DT=0,263), Melilla (M=0,276,
DT=0,235) Extremadura (M=0,272, DT=0,258), Cataluña (M=0,267, DT=0,255), País
Vasco
(M=0,261, DT=0,248)
y
Navarra
(M=0,257, DT=0,250)
.
También en el caso de las
Ϭ͕Ϯϯ
Ϭ͕Ϯϰ
Ϭ͕Ϯϱ
Ϭ͕Ϯϲ
Ϭ͕Ϯϳ
Ϭ͕Ϯϴ
Ϭ͕Ϯϵ
Ϭ͕ϯ
Ϭ͕ϯϭ
ϮϬϭϱϮϬϭϲϮϬϭϳϮϬϭϴϮϬϭϵϮϬϮϬ
Source: created by the authors
378 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
4.3. Regional dierences
Firstly, it should be noted that the greatest attention to the issue comes in the most populous regions, thus Madrid, Catalonia
and Andalusia accumulate a greater volume of messages about migrants and refugees. Figure 4 summarises the amount of
content on the question compiled in each region. However, the distribution does not correspond to the proportion of hatred
detected, where the regions with the most hatred detected are Asturias and Castilla-La Mancha. Contrasting of averages
shows that there are signicant dierences between the Spanish regions [F(17, 13157.451)=27.802, p<0.001], and post hoc
tests conrm these dierences. us, the average value for hate detected in Asturias (M=0.323, SD=0.285) and in Castilla-La
Mancha (M=0.318, SD=0.280) is signicantly higher than that detected in the Valencian Community (M=0.295, SD=0.268),
Canary Islands (M=0.292, SD=0.264), Madrid (M=0.282, SD=0.263), Melilla (M=0.276, SD=0.235), Extremadura (M=0.272,
SD=0.258), Catalonia (M= 0.267, SD=0.255), the Basque Country (M=0.261, SD=0.248), and Navarra (M=0.257, SD=0.250). In
the case of the Balearic Islands (M=0.311, SD=0.278) the dierences were signicant with respect to the Community of Madrid,
Extremadura, Catalonia, the Basque Country and Navarra.
Figure 4. Number of messages about migrants and refugees by Spanish region on Twitter between 2015 and 2020
Islas Baleares (M=0,311, DT=0,278) las diferencias fueron significativas con respecto a
Comunidad de Madrid, Extremadura, Cataluña, País Vasco y Navarra.
Figura 4. Cantidad de mensajes sobre migrantes y refugiados por comunidad autónoma en
Twitter en España entre 2015 y 2020
Fuente: elaboración propia
La Figura 5 resume los datos anteriores. No se observan, en general, saltos pronunciados
entre unas comunidades y otras, y tampoco se detectan patrones regionales (norte
-
sur, centro
-
periferia), ni ideológicos (la tendencia ideológica de los gobiernos regionales no sigue una
tendencia clara), y tampoco lo hace la pr
o
porción de población
inmigran
te en cada región,
como apuntaban Arcil
a-
Calderón et al. (2022). El único elemento que parece intuirse es la
presencia de las regiones con un mayor PIB per cápita (Madrid, País Vasco, Navarra y
Cataluña) entre aquellas con menor presencia de odio. En cualquier caso, serán necesarios
estudios más detallados que puedan arrojar luz sobre estas diferencias.
Ϭ
ϱϬϬϬ
ϭϬϬϬϬ
ϭϱϬϬϬ
ϮϬϬϬϬ
ϮϱϬϬϬ
ϯϬϬϬϬ
ϯϱϬϬϬ
"ŽŵƵŶŝĚĂĚ ĚĞ DĂĚƌŝĚ
"ĂƚĂůƵŹĂ
$ŶĚĂůƵĐşĂ
"ŽŵƵŶŝĚĂĚ sĂůĞŶĐŝĂŶĂ
WĂşƐ sĂƐĐŽ
"ĂƐƚŝůůĂ LJ >ĞſŶ
/ƐůĂƐ "ĂŶĂƌŝĂƐ
'ĂůŝĐŝĂ
ZĞŐŝſŶ ĚĞ DƵƌĐŝĂ
$ƌĂŐſŶ
"ĂƐƚŝůůĂ Ͳ >Ă DĂŶĐŚĂ
WƌŝŶĐŝƉĂĚŽ ĚĞ $ƐƚƵƌŝĂƐ
%džƚƌĞŵĂĚƵƌĂ
/ƐůĂƐ !ĂůĞĂƌĞƐ
"ŽŵƵŶŝĚĂĚ &ŽƌĂů ĚĞ EĂǀĂƌƌĂ
"ĂŶƚĂďƌŝĂ
DĞůŝůůĂ
>Ă ZŝŽũĂ
Source: created by the authors
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 379
Figure 5 summarises the above data. In general, clear dierences between the regions cannot be observed, be they pronounced
jumps between one region and another, regional patterns (north-south, centre-periphery), ideological (the ideological
tendency of regional governments does not follow a clear trend), nor even the proportion of the immigrant population in each
region, as Arcila-Calderón et al. observed (2022). e only element that seems to be intuited is the presence of the regions with
a higher GDP per capita (Madrid, the Basque Country, Navarra and Catalonia) among those with a lower presence of hate. In
any case, more detailed studies will be necessary that can shed light on these dierences.
Figure 5. Presence of hate speech towards migrants and refugees by region on Twitter in Spain between 2015 and 2020
Figura 5. Presencia de discurso de odio hacia migrantes y refugiados por comunidad
autónoma en Twitter en España entre 2015 y 2020
Fuente: elaboración propi
a
4.4. Estudio de sentimientos
Por último, el estudio del sentimiento muestra un valor medio de
-
0,125 (DT=0,229). Esto
quiere decir que el sentimiento general de los tweets sobre migrantes y refugiados en España
es ligeramente negativo. Existe una correlación significativamente negativa entre el
sentimiento y la presencia de odio
[R(124.031)=-
0,068, p<0,001], es decir, un sentimiento
más negativo está correlacionado con una mayor presencia de odio, aunque el efecto es de
tamaño muy reducido. De hecho, si evaluamos exclusivamente los mensajes con una
presencia de odio consistente (aquellos en los que el valor de la presencia de odio es superior
al 0,5) el valor medio del sentimiento es de
-
0,151 (DT=0,226), es decir, muy ligeramente
más negativo.
Lo anterior es relevante porque implica una aproximación negativa al discurso centrado en
migrantes y refugiados, algo particularmente habitual en aquellos mensajes con odio, pero
no exclusivo de ellos. Es comprensible, puesto que incluso aquellos mensajes que no incluyen
rechazo tienden a centrarse en los aspectos negativos de la migración o al drama que plantea,
algo que concuerda con la cobertura que se suele realizar de la migración en los medios de
Ϭ
Ϭ͕Ϭϱ
Ϭ͕ϭ
Ϭ͕ϭϱ
Ϭ͕Ϯ
Ϭ͕Ϯϱ
Ϭ͕ϯ
Ϭ͕ϯϱ
WƌŝŶĐŝƉĂĚŽ ĚĞ $ƐƚƵƌŝĂƐ
"ĂƐƚŝůůĂ Ͳ >Ă DĂŶĐŚĂ
/ƐůĂƐ !ĂůĞĂƌĞƐ
'ĂůŝĐŝĂ
$ŶĚĂůƵĐşĂ
"ĂƐƚŝůůĂ LJ >ĞſŶ
$ƌĂŐſŶ
ZĞŐŝſŶ ĚĞ DƵƌĐŝĂ
"ŽŵƵŶŝĚĂĚ sĂůĞŶĐŝĂŶĂ
/ƐůĂƐ "ĂŶĂƌŝĂƐ
"ĂŶƚĂďƌŝĂ
%ƐƉĂŹĂ
>Ă ZŝŽũĂ
"ŽŵƵŶŝĚĂĚ ĚĞ DĂĚƌŝĚ
DĞůŝůůĂ
%džƚƌĞŵĂĚƵƌĂ
"ĂƚĂůƵŹĂ
WĂşƐ sĂƐĐŽ
"ŽŵƵŶŝĚĂĚ &ŽƌĂů ĚĞ EĂǀĂƌƌĂ
Source: created by the authors
380 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
4.4. Study of sentiments
Finally, sentiment analysis shows an average value of -0.125 (SD=0.229). is means that the general sentiment in tweets
concerning migrants and refugees in Spain is slightly negative. ere is a signicant negative correlation between sentiment
and the presence of hate [R(124,031)=-0.068, p<0.001], that is to say, a more negative feeling is correlated with a greater
presence of hate, although the eect is very slight. In fact, if we only evaluate messages with a consistent presence of hate
(those in which the value of the presence of hate is greater than 0.5) the average value of sentiment is -0.151 (SD=0.226), in
other words, very slightly more negative.
e above is relevant because it implies a negative approach to discourse focused on migrants and refugees, something
particularly common in hate messages, but not exclusive to them. is is understandable, since even messages that do not
include rejection tend to focus on the negative aspects of migration or the drama it poses, something that is consistent with
media coverage of migration (Fengler et al., 2020; Igartua et al., 2007). In fact, Spanish news stories about issues related to
migration commonly focus on the Mediterranean, Ceuta and Melilla or the Canary Islands (Fajardo Fernández & Soriano Miras,
2016), that is, points of conict, where tragic events sometimes take place; in such cases, feelings of sadness or indignation over
the death of a migrant would not be hatred but do represent negative sentiment.
Looking more deeply at developments over time, there are signicant dierences between the dierent years [F(5,
100,459)=100.459, p<0.001]. e post hoc tests conrm that sentiment shows greater negativity in the years 2016 (M=-0.143,
SD=0.216) and 2015 (M=-0.140, SD=0.215) than in the years 2020 (M=-0.121, SD=0.242), 2019 (M=-0.116, SD=0.248), 2017
(M=-0.110, SD=0.223) and, particularly, 2018 (M=-0.109, SD=0.239). Figure 6 shows these dierences more visually. One can
see that it follows a pattern similar to the distribution of hatred (although inverted because it deals with negative values),
with lower values in the middle years, although in the case of hatred the later years show the greatest presence, while the
earlier years had the most negative feeling. is could be interpreted by the type of feelings associated with the information on
migration generated in those years (López del Ramo & Humanes, 2016; Brändle, Eisele & Trenz, 2019; Amores, Arcila Calderón
& Stanek, 2019), with cases of shipwrecks, detentions, overcrowding and stories with considerable impact in the media, such
as the death of Aylan Kurdi or the discovery of a truck with dozens of suocated immigrants on the border between Austria
and Hungary (Fleming, 2015).
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 381
Figure 6. Annual changes in the underlying sentiment in messages about migrants and refugees on Twitter in Spain between 2015 and 2020
comunicación (Fengler et al., 2020; Igartua et al., 2007
)
. De hecho, en el caso español, son
habituales las historias
sobre
cuestiones relaciona
d
as con la migración se centren en el
Mediterráneo, Ceuta y Melilla o las Islas Canarias (Fajardo Fernández & Soriano Miras,
2016
), es decir, puntos de conflicto, en los que en ocasiones se producen tragedias; en este
sentido, un sentimiento de tristeza o indignación por la muerte de una persona migrante no
sería odio, pero sí sentimiento negativo.
Profundizando en la evolución temporal, existen diferencias significativas entre los distintos
años [F(5, 100.459)=100,459, p<0,001]. Las pruebas post hoc confirman que el sentimiento
es más negativo en los años 2016 (M=
-
0,143, DT=0,216) y 2015 (
M=-
0,140, DT=0,215
)
que
en los años 2020 (M=
-
0,121, DT=0,242), 2019 (M=
-
0,116, DT=0,248), 2017 (M=
-
0,110,
DT=0,223) y, sobre todo, 2018 (M=
-
0,109, DT=0,239). La Figura 6 muestra estas diferencias
de manera más visual. Vemos que sigue un esquema semejante a la distribución del odio
(aunque invertido por tratarse de valores negativos), con valores más reducidos en los años
centrales,
aunque
en el caso del odio los años con mayor presencia eran los últimos, mientras
que los años con sentimiento más negativo son los iniciales. Esto podría interpretarse por el
tipo de sentimientos asociados a las informaciones sobre migración generadas esos años
(López del Ramo & Humanes, 2016; Brändle, Eisele & Trenz, 2019; Amores, Arcila
Calderón & Stanek, 2019
)
, con casos de naufragios, detenciones, hacinamiento y casos de
gran impacto mediático, como la muerte de Aylan Kurdi o el descubrimiento de un camión
con decenas de inmigrantes asfixiados en la frontera entre Austria y Hungría (Fleming, 2015).
Figura 6. Evolución anual del sentimiento subyacente en los mensajes sobre migrantes y
refugiados en Twitter en España entre 2015 y 2020
Fuente: elaboración propia
Por último, la comparación entre las comunidades autónomas vuelve a arrojar diferencias
significativas [F(5, 100.459)=100,459, p<0,001]. En este caso Cantabria tiene un sentimiento
significativamente más negativo (M=
-
0,148, DT=0,252) que Madrid (
M=-
0,121,
DT=0,233
),
ͲϬ͕ϭϲ
ͲϬ͕ϭϰ
ͲϬ͕ϭϮ
ͲϬ͕ϭ
ͲϬ͕Ϭϴ
ͲϬ͕Ϭϲ
ͲϬ͕Ϭϰ
ͲϬ͕ϬϮ
Ϭ
ϮϬϭϱϮϬϭϲϮϬϭϳϮϬϭϴϮϬϭϵϮϬϮϬ
Source: created by the authors
Finally, the comparison between the regions once again shows signicant dierences [F(5, 100.459)=100.459, p<0.001].
Cantabria has signicantly greater negative sentiment (M=-0.148, SD=0.252) than Madrid (M=-0.121, SD=0.233), Castilla and
León (M=-0.120, SD=0.221), Murcia (M=-0.110, SD=0.235), Navarra (M=-0.106, SD=0.238) and Melilla (M=-0.076, SD=0.191).
In the case of Asturias (M=-0.138, SD=0.224) the dierences are signicant when compared to Madrid, Murcia, Navarra, and
Melilla. e case of Melilla is particularly striking, with signicantly less negative sentiment than all the other regions, except
Navarra. Figure 7 summarises this data more graphically.
382 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Figure 7. Underlying sentiment in messages about migrants and refugees by region on Twitter in Spain between 2015 and 2020
Castilla y León
(M=-
0,120, DT=0,221), Región de Murcia
(M=-
0,110, DT=0,235), Navarra
(M=-
0,106, DT=0,238) y Melilla
(M=-
0,076, DT=0,191). En el caso de Asturias (M=
-
0,138,
DT=0,224) las diferencias resultan significativas con Madrid, Murcia, Navarra y Melilla.
Resulta particularmente llamativo el caso de Melilla, con un sentimiento significativamente
menos negativo que todas las demás regiones, salvo Navarra. La Figura 7 resume estos datos
de manera más visual.
Figura 7. Sentimiento subyacente en los mensajes sobre migrantes y refugiados por
comunidad autónoma en Twitter en España entre 2015 y 2020
Fuente: elaboración propia
Aunque en estos casos sí observamos saltos más pronunciados que en lo que respecta a la
presencia de odio, sobre todo en el caso de Melilla. En cualquier caso, el rango es reducido,
y no llega a los 0,10 puntos en un rango ente
-
1 y +1. El orden tampoco parece obedecer a
criterios claros, por lo que será necesario profundizar en potenciales causas
q
ue expliquen
estas diferencias.
5.
Conclusiones
Comenzando por la PI1 sobre la evolución de los contenidos sobre migrantes y refugiados
en Twitter en España, se puede afirmar que el interés ha venido marcado por sucesos de gran
impacto mediático, siendo los años 2015 y 2016, los más relevantes de la crisis de refugiados,
los que más atención se prestó a esta materia, como también confirma el interés académico
sobre la relevancia que estos años tuvieron en la representación mediática de la migrción
(Zhang & Hellmueller, 2017; Amores, Arcila
-
Calderón & Blanco
-
Herrero, 2020). Tras esos
ͲϬ͕ϭϲ
ͲϬ͕ϭϰ
ͲϬ͕ϭϮ
ͲϬ͕ϭ
ͲϬ͕Ϭϴ
ͲϬ͕Ϭϲ
ͲϬ͕Ϭϰ
ͲϬ͕ϬϮ
Ϭ
DĞůŝůůĂ
"ŽŵƵŶŝĚĂĚ &ŽƌĂů ĚĞ EĂǀĂƌƌĂ
ZĞŐŝſŶ ĚĞ DƵƌĐŝĂ
/ƐůĂƐ !ĂůĞĂƌĞƐ
"ĂƐƚŝůůĂ LJ >ĞſŶ
"ŽŵƵŶŝĚĂĚ ĚĞ DĂĚƌŝĚ
/ƐůĂƐ "ĂŶĂƌŝĂƐ
%ƐƉĂŹĂ
%džƚƌĞŵĂĚƵƌĂ
"ŽŵƵŶŝĚĂĚ sĂůĞŶĐŝĂŶĂ
$ŶĚĂůƵĐşĂ
'ĂůŝĐŝĂ
WĂşƐ sĂƐĐŽ
"ĂƚĂůƵŹĂ
>Ă ZŝŽũĂ
$ƌĂŐſŶ
"ĂƐƚŝůůĂ Ͳ >Ă DĂŶĐŚĂ
WƌŝŶĐŝƉĂĚŽ ĚĞ $ƐƚƵƌŝĂƐ
"ĂŶƚĂďƌŝĂ
Source: created by the authors
More pronounced jumps can be observed in these cases than in the presence of hate, especially in the case of Melilla. However,
the range is quite limited, not reaching 0.10 points in a range between -1 and +1. Neither does the order seem to obey clear
criteria, so it will be necessary to look more deeply into potential causes that could explain these dierences.
5. Conclusions
Starting with RQ1, concerning changes in the content about migrants and refugees on Twitter in Spain, it can be armed
that interest has been marked by events of great impact in the media, with 2015 and 2016 being the most noteworthy years
of the refugee crisis, the years when most attention was paid to this matter, as also conrmed by the academic interest in the
importance that these years had in the media representation of migration (Zhang & Hellmueller, 2017; Amores, Arcila-Calderón
& Blanco-Herrero, 2020). After those years, interest seems to have declined, reaching minimum levels in 2020, coinciding with
the pandemic, when attention shifted to health-related matters.
Regarding RQ2, focused on the changes in hate speech, the trend is dierent and concerning, as the volume of messages of
this type has increased, 2020 being the year with the greatest rejection on Twitter in Spain towards migrants and refugees. e
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 383
emergence in the Spanish political arena of a party with an anti-immigration and nationalist discourse seems to have played a
key role in this sense, as previous research has already indicated (Ferreira, 2019; Turnbull-Dugarte, 2019).
To respond to RQ3, regarding regional dierences, we can arm that, although the greatest volume of messages about migrants
and refugees was concentrated in the most populous regions (Madrid, Catalonia, and Andalusia), the highest proportions of
hatred were observed in Asturias, Castilla-La Mancha, the Balearic Islands, and Galicia, without clear patterns being detected.
It may be advisable for future studies to look into these issues in greater detail, since there do not seem to be ideological
patterns –the colour of the regional government, for example–, wealth or population of immigrant origin that explain these
dierences.
Finally, the sentiment analysis carried out to respond to RQ4 indicates that there is a very weak correlation, whereby sentiment
is more negative in messages with a greater presence of hate. In general, the sentiment in discourse about migrants and
refugees on Twitter in Spain between 2015 and 2020 was negative, especially in the rst two years of the sample, something that
coincides with the fact that media coverage and discourse concerning migration tends to focus on negative elements (Amores,
Arcila-Calderón & Blanco-Herrero, 2020). Even the discourse in support of migrants can incorporate negativity, by focusing
on their status as victims, something that could also explain the weakness of the correlation. Sentiment was negative in all the
regions of Spain, Melilla being the place with the least negative sentiment, and Cantabria, the most.
It is worth noting that this work has focused on sentiment analysis from a binary perspective, between positive and negative.
is derives from the relevance of examining the association between negativity and hate speech more closely, a connection
that other studies have already observed (Arcila-Calderón et al., 2021) and which has been conrmed herein. is, however,
sets a limitation, since it does not allow us to evaluate what specic feelings or emotions are associated with this type of
rejection discourse, a line of work that future research in the area may explore, research which could go beyond the exploratory
vocation of this study.
e above allows us to gain greater knowledge about online hate speech based on racist and xenophobic criteria, conrming
the clear relationship between such hatred and news and media events (Arcila-Calderón et al., 2021). e application of
computational techniques and automated analyses has allowed us to carry out one of the most extensive longitudinal studies
yet, with the particularity of including geolocated content which has allowed consistent comparisons to be made between
regions of Spain. is is a considerable advance as a large part of the policies pertaining to coexistence and integration in
Spain are the responsibility of the regional governments. However, and despite the interesting observations made, it must be
pointed out that the main dierences regarding attitudes towards migration occur at the national rather than regional level, as
illustrated by the various Eurobarometers (European Commission, 2019; 2022), hence the interpretation of those dierences
at the regional level still requires future study to give clearer understanding.
Beyond this, a limitation that must be taken into account is the fact that with large volumes of data it is easier for statistical tests
to be statistically signicant. us, it is worth qualifying the observations made, especially those related to regional dierences,
whose eects were smaller than those observed between the dierent years.
It should also be noted regarding regional variables that the corresponding region could not be identied automatically in
4,271 cases, since the metadata did not include enough information. Among these messages it is expected that there will be,
384 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
among others, all those corresponding to the autonomous city of Ceuta, which is not included among the regions or cities
in the study. Given Ceuta’s population, and in view of the number of messages identied in Melilla (858), a similar city, it is
estimated that the eects of this limitation will be slight, but it would be advisable for future studies to continue perfecting the
analyses to look into this question in greater depth.
6. Acknowledgements
is study has been supported by the project “Enhanced migration measures from a multidimensional perspective
(HumMingBird)” nanced by the European Union within the framework of the Horizon 2020 Research and Innovation
Programme, reference number 870661.
is article has been translated into English by Brian O´Halloran
7. Specic contributions of each author
Name & surname
Conception and design of the work
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos
Arcila-Calderón
Methodology
David Blanco-Herrero, Carlos Arcila-Calderón
Data collection and analysis
David Blanco-Herrero, Carlos Arcila-Calderón, Patricia Sánchez-
Holgado
Discussion and conclusionsAndrés Barradas Gurruchaga, David Blanco-Herrero
Drafting, formatting, version review and approval
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos
Arcila-Calderón, Patricia Sánchez-Holgado
8. Conicts of interest
e authors declares that there is no conict of interest contained in this article.
9. Bibliographic references
Amores, J., Arcila-Calderón, C., & Blanco-Herrero, D. (2020). Evolution of negative visual frames of immigrants and refugees in
the main media of Southern Europe.
El Profesional de la Información
29(6), e290624.
https://doi.org/10.3145/epi.2020.nov.24
Amores, J.J., Arcila Calderón, C., & Stanek, M. (2019). Visual frames of migrants and refugees in the main Western European
media.
Economics and Sociology
,
12
(3), 147-161.
https://doi.org/10.14254/2071-789X.2019/12-3/10
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 385
Andres, R. & Slivko, O. (2021). Combating Online Hate Speech: e Impact of Legislation on Twitter.
ZEW - Centre for European
Economic Research Discussion Paper No. 21-103
.
https://doi.org/10.2139/ssrn.4013662
Arcila-Calderón, C., Sánchez-Holgado, P., Quintana-Moreno, C., Amores, J. J., & Blanco-Herrero, D. (2022). Discurso de
odio y aceptación social hacia migrantes en Europa: Análisis de tuits con geolocalización.
Comunicar: Revista Cientíca de
Comunicación y Educación,
30
(71).
https://doi.org/10.3916/C71-2022-02
Arcila-Calderón, C., Blanco-Herrero, D., Frías-Vázquez, M., & Seoane-Pérez, F. (2021). Refugees welcome? Online hate
speech and sentiments in Twitter in Spain during the reception of the boat Aquarius.
Sustainability
,
13
(5), 2728.
https://doi.
org/10.3390/su13052728
Arcila-Calderón, C., Blanco-Herrero, D., & Valdez-Apolo, M.B. (2020). Rechazo y discurso de odio en Twitter: Análisis de
contenido de los tuits sobre migrantes y refugiados en español.
Revista Española de Investigaciones Sociológicas
,
172
, 21-40.
https://doi.org/10.5477/cis/reis.172.21
Arcila-Calderón, C., de la Vega, G., & Blanco-Herrero, D. (2020). Topic Modeling and Characterization of Hate Speech against
Immigrants on Twitter around the Emergence of a Far-Right Party in Spain.
Social Sciences
,
9
(11), 188.
https://doi.org/10.3390/
socsci9110188
Ausserhofer, J., & Maireder, A. (2013). National politics on Twitter: Structures and topics of a networked public sphere.
Information, communication & society
,
16
(3), 291-314.
https://doi.org/10.1080/1369118X.2012.756050
Backfried, G., & Shalunts, G. (2016, octubre). Sentiment analysis of media in german on the refugee crisis in europe.
En
International Conference on Information Systems for Crisis Response and Management in Mediterranean Countries
(pp.
234-241). Springer, Cham.
Brändle, V. K., Eisele, O., & Trenz, H.-J. (2019). Contesting European Solidarity During the “Refugee Crisis”: A Comparative
Investigation of Media Claims in Denmark, Germany, Greece and Italy.
Mass Communication and Society
,
22
(6), 708-732.
https://doi.org/10.1080/15205436.2019.1674877
Buder, J., Rabl, L., Feiks, M., Badermann, M., & Zurstiege, G. (2021). Does negatively toned language use on social media lead
to attitude polarization?
Computers in Human Behavior
,
116
, 106663.
https://doi.org/10.1016/j.chb.2020.106663
Campos-Domínguez, E. (2017). Twitter y la comunicación política.
El professional de la información
,
26
(5), 785-793.
https://
doi.org/10.3145/epi.2017.sep.01
Castromil, A.R., Rodríguez-Díaz, R., & Garrigós, P. (2020). La agenda política en las elecciones de abril de 2019 en España:
programas electorales, visibilidad en Twitter y debates electorales.
El profesional de la información
,
29
(2), e290217.
https://
doi.org/10.3145/epi.2020.mar.17
Chóliz, M., & Gómez, C. (2002). Emociones sociales II (enamoramiento, celos, envidia y empatía). En Palmero, F., Fernández-
Abascal, E., Martínez, F. & Chóliz, M., (Coords.),
Psicología de la motivación y las emociones
(pp. 395-418). McGraw Hill.
Comisión Española de Ayuda al Refugiado. (2019).
Informe 2019: Las personas refugiadas en España y Europa
.
https://tinyurl.
com/bddr35vc
<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>
<p>[Enlace de URL / hc (has AS)]</p>
386 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Comisión Española de Ayuda al Refugiado. (2021).
Informe 2021: Las personas refugiadas en España y Europa
.
https://tinyurl.
com/atvjjecj
Damstra, A., Jacobs, L., Boukes, M., & Vliegenhart, R. (2019). e Impact of Immigration News on Anti-Immigrant Party
Support: Unpacking Agenda-Setting and Issue Ownership Eects Over Time.
Journal of Elections, Public Opinion and Parties,
31
(1), 97-118.
https://doi.org/10.1080/17457289.2019.1607863
de León, E., & Trilling, D. (2021). A sadness bias in political news sharing? e role of discrete emotions in the engagement
and dissemination of political news on Facebook.
Social Media + Society
,
7
(4), 20563051211059710.
https://doi.
org/10.1177/20563051211059710
Dennison, J., & Geddes, A. (2019). A rising tide? e salience of immigration and the rise of anti-immigration political parties
in Western Europe.
e political quarterly
,
90
(1), 107-116.
https://doi.org/10.1111/1467-923X.12620
Díaz Soto, J. M. (2015). Una aproximación al concepto de discurso del odio.
Revista Derecho del Estado
,
34
, 77-101.
https://doi.
org/10.18601/01229893.n34.05
Eberl, J.-M., Meltzer, C. E., Heidenreich, T., Herrero, B., eorin, N., Lind, F., Berganza, R., Boomgaarden, H. G., Schemer,
C., & Strömbäck, J. (2018). e European Media Discourse on Immigration and its Eects: A Literature Review.
Annals of the
International Communication Association
,
42
(3), 207-223.
https://doi.org/10.1080/23808985.2018.1497452
Esses, V. M., Dovidio, J. F., Semenya, A. H., & Jackson, L. M. (2005). Attitudes Towards Immigrants and Immigration: e Role of
National and International Identity. En D. Abrams, M. A. Hogg, & J. M. Marques (eds.),
e Social Psychology of Inclusion and
Exclusion
(pp. 317-337). Psychology Press.
European Commission. (2019).
Special Eurobarometer 493. Discrimination in the European Union
.
https://tinyurl.
com/4xx4spza
European Commission (2022).
Special Eurobarometer 519. Integration of Immigrants in the European Union
.
https://tinyurl.
com/yxyd83n3
European Commission against Racism and Intolerance. (2005).
General Policy Recommendation No.15, on combatting hate
speech
.
https://tinyurl.com/mthh6mf8
Fajardo Fernández, R., & Soriano Miras, R. M. (2016). La construcción mediática de la migración en el Mediterráneo: ¿no-
ciudadanía en la prensa española?
Revista Internacional de Estudios Migratorios
,
6
(1), 141-169.
https://tinyurl.com/3835r2dv
Fengler, S., Bastian, M., Brinkmann, J., Zappe, A. C., Tatah, V., Andindilile, M., Assefa, E. et al. (2020). Covering Migration—in
Africa and Europe: Results from a Comparative Analysis of 11 Countries. Journalism Practice, 16(1), 140-160.
https://doi.org/
10.1080/17512786.2020.1792333
Fernández-Abascal, E., Palmero, F., & Breva, A. (2002). Emociones básicas I (miedo, alegría y sorpresa). En Palmero, F.,
Fernández-Abascal, E., Martínez, F. & Chóliz, M., (Coords.),
Psicología de la motivación y las emociones
(pp. 333-351). McGraw
Hill.
<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>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 387
Ferreira, C. (2019). Vox como representante de la derecha radical en España: un estudio sobre su ideología.
Revista Española
de Ciencia Política
,
51
, 73-98.
https://doi.org/10.21308/recp.51.03
Fleming, M. (2015, 28 de agosto).
Cuerpos sin vida encontrados en un camión cerca de la frontera, mientras los solicitantes de
asilo continúan llegando a Hungría
. ACNUR.
https://tinyurl.com/2j33amdn
Greussing, E., & Boomgaarden, H. G. (2017). Shifting the refugee narrative? An automated frame analysis of Europe’s 2015
refugee crisis.
Journal of Ethnic and Migration Studies
,
43
(11), 1749-1774.
https://doi.org/10.1080/1369183X.2017.1282813
Gruzd, A., y Roy, J. (2014). Investigating political polarization on Twitter: A Canadian perspective.
Policy & Internet
,
6
(1), 28-45.
https://doi.org/10.1002/1944-2866.POI354
Hameleers, M., van der Meer, T., & Vliegenthart, R. (2022). Civilized truths, hateful lies? Incivility and hate speech in false
information–evidence from fact-checked statements in the US.
Information, Communication & Society
,
25
(11), 1596-1613.
https://doi.org/10.1080/1369118X.2021.1874038
Heidenreich, T., Eberl, J. M., Lind, F., & Boomgaarden, H. (2020). Political migration discourses on social media: a comparative
perspective on visibility and sentiment across political Facebook accounts in Europe.
Journal of Ethnic and Migration
Studies
,
46
(7), 1261-1280.
https://doi.org/10.1080/1369183X.2019.1665990
Igartua, J. J., Muñiz, C., Otero-Parra, J. A., & De-la-Fuente-Juan, M. (2007). El Tratamiento Informativo de la Inmigración en
los Medios de Comunicación Españoles. Un Análisis de Contenido Desde la Teoría del Framing.
Estudios Sobre el Mensaje
Periodístico, 13,
91-110.
Joppke, C. (2020). Immigration in the populist crucible: comparing Brexit and Trump.
Comparative migration studies
,
8
(1),
1-18.
https://doi.org/10.1186/s40878-020-00208-y
Kreis, R. (2017). #refugeesnotwelcome: Anti-refugee discourse on Twitter.
Discourse & Communication
,
11
(5), 498-514.
https://
doi.org/10.1177/1750481317714121
Krzy
ż
anowski, M., Triandafyllidou, A., & Wodak, R. (2018). e Mediatization and the Politicization of the “Refugee Crisis” in
Europe.
Journal of Immigrant & Refugee
Studies
,
16
(1–2), 1-14.
https://doi.org/10.1080/15562948.2017.1353189
Lacomba Vázquez, J., Benlloch Doménech, C., Cloquell Lozano, A., & Veira Ramos, A. (2020).
La aportación de la inmigración
a la sociedad española
. Ministerio de Inclusión, Seguridad Social y Migraciones. Observatorio Permanente de la Inmigración.
https://tinyurl.com/mrnerm29
López del Ramo, J. L., & Humanes, M. L. (2016). Análisis del framing visual y sus componentes en el tratamiento fotográco
de la crisis de los refugiados sirios en medios de prensa internacional.
Scire: representación y organización del conocimiento
,
22
(2), 87-97.
Mielczarek, N. (2018). e dead Syrian refugee boy goes viral: funerary Aylan Kurdi memes as tools of mourning and visual
reparation in remix culture.
Visual
Communication
,
19
(4), 506-530.
https://doi.org/10.1177/1470357218797366
Miró Llinares, F. (2016). Taxonomía de la comunicación violenta y el discurso del odio en Internet.
IDP. Revista de Internet,
Derecho y Política
,
22
, 82-107.
<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>
388 |
nº 38, pp. 369-389
|
January-June of 2024
Migrants and refugees on Twitter in Spain: a study employing automated analysis into the presence of hate and sentiment
ISSN: 1696-019X / e-ISSN: 2386-3978
doxa.comunicación
Moragas-Fernández, C. M., Grau-Masot, J. M., & Capdevila-Gómez, A. (2019). Articulación de la inuencia en Twitter ante
el anuncio de la Ley del referéndum en Cataluña.
Profesional
de la Información
,
28
(3), e280320.
https://doi.org/10.3145/
epi.2019.may.20
Müller, K., & Schwarz, C. (2020). Fanning the ames of hate: Social media and hate crime.
Journal of the European Economic
Association, 19
(4), 2131-2167
.
https://doi.org/10.1093/jeea/jvaa045
Muñiz, C., Igartua, J. J., & Otero, J. A. (2006). Imágenes de la Inmigración a Través de la Fotografía de Prensa. Un Análisis de
Contenido.
Communication & Society
,
19
(1) 103-128.
https://doi.org/10.15581/003.19.1.103-128
Organización Internacional de las Migraciones (2019).
Glosario de la OIM sobre migración
.
https://tinyurl.com/mww4cn4h
OSCE (s.f.).
Hate Crime Reporting
,
https://hatecrime.osce.org
Pew Research Center (2021).
Diversity and Division in Advanced Economies. Spring 2021 Global Attitudes Survey
.
https://
tinyurl.com/yeyr8be4
Robles, J., Guevara, J., Casas-Mas, B., & Gómez, D. (2022). When negativity is the fuel. Bots and Political Polarization in the
COVID-19 debate.
Comunicar
,
30
(71), 63-75.
https://doi.org/10.3916/C71-2022-05
Rodríguez Andrés, R., & Ureña Uceda, D. (2011). Diez razones para el uso de Twitter como herramienta en la comunicación
política y electoral.
Comunicación y Pluralismo
,
10
, 89-116.
Rollnert Liern, G. (2020). Redes sociales y discurso del odio: perspectiva internacional.
IDP. Internet, Derecho y Política
,
31
,
1-14.
https://doi.org/10.7238/idp.v0i31.3233
Schemer, C. (2012). e Inuence of news media on stereotypic attitudes toward immigrants in a political campaign.
Journal
of communication
,
62
(5), 739-757.
https://doi.org/10.1111/j.1460-2466.2012.01672.x
Seoane-Pérez, F. (2017). Framing of the Syrian Refugee Crisis in the Spanish Press. En Barlai, M., Fähnrich, B., Griessler, C. &
Rhomberg, M. (Eds.),
e Migrant Crisis: European Perspectives and National Discourses
(pp. 267-282). LIT Verlag.
Soroka, S., & McAdams, S. (2015). News, politics, and negativity.
Political communication
,
32
(1), 1-22.
https://doi.org/10.108
0/10584609.2014.881942
Splinder, W. (2015, 08 de diciembre)
2015: El año de la crisis de refugiados en Europa
. ACNUR.
https://tinyurl.com/34ucruu5
Tajfel, H. (1978).
Dierentiation Between Social Groups: Studies in the Social Psychology of Intergroup Relations
. Academic
Press
elwall, M., Buckley, K., Paltoglou, G., Cai, D., & Kappas, A. (2010). Sentiment strength detection in short informal text.
Journal of the American society for information science and technology
,
61
, 2544-2558.
https://doi.org/10.1002/asi.21416
Turnbull-Dugarte, S.J. (2019). Explaining the end of Spanish exceptionalism and electoral support for Vox.
Research & Politics,
6
(2), 1-8.
https://doi.org/10.1177/2053168019851680
Twitter (2022).
Hateful conduct policy.
https://tinyurl.com/462n2xwj
<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>
<p>[Enlace de URL / hc (has AS)]</p>
<p>[Enlace de URL / hc (has AS)]</p>
doxa.comunicación |
nº 38, pp. 369-389
January-June of 2024
Andrés Barradas Gurruchaga, David Blanco-Herrero, Carlos Arcila-Calderón and Patricia Sánchez-Holgado
ISSN: 1696-019X / e-ISSN: 2386-3978
| 389
Vrysis, L., Vryzas, N., Kotsakis, R., Saridou, T., Matsiola, M., Veglis, A., Arcila-Calderón, C., & Dimoulas, C. (2021). A Web
interface for analyzing hate speech.
Future Internet
,
13
(3), 80.
https://doi.org/10.3390/13030080
Yardi, S., & Boyd, D. (2010). Dynamic debates: An analysis of group polarization over time on twitter.
Bulletin of science,
technology & society
,
30
(5), 316-327.
https://doi.org/10.1177/0270467610380011
Yurdakul, K, H. (2021). How Do the Main Negative Emotions Aect People’s Political Decision Process? Fear, Anxiety and
Anger.
Journal of Academic Inquiries
,
16
(1), 247-261.
https://doi.org/10.17550/akademikincelemeler.708916
Zhang, X., & Hellmueller, L. (2017). Visual framing of the European refugee crisis in Der Spiegel and CNN International:
Global journalism in news photographs.
International Communication Gazette
,
79
(5), 483-510.
https://doi.
org/10.1177/1748048516688134
<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>