Media representation of minors who migrate on their own: The "MENA" in the Spanish press
Abstract
This article analyses Spanish media treatment of a certain type of immigrant: the unaccompanied foreign minor (“MENA” in Spanish). The media play an important role in creating and disseminating ideas and images amongst the general public, thereby promoting the articulation of sets of meanings called discourses. The main goal of this research is to identify the discursive approaches that have been constructed around the term “MENA” in the main Spanish daily newspapers. To this end, we gathered and analysed all the news reports published between January 2017 and October 2019 by the digital editions of the four most widely-read newspapers in Spain (La Vanguardia, El País, El Mundo and ABC). This analysis was performed using text mining techniques (an important field in data science) such as term frequency, inverse document frequency, and correlation networks between words. Our results show that the term “MENA” evokes a criminalising, moralistic, welfare-dependent discourse that is articulated from an adult-centric, nationalist perspective. The study concluded that the conservative press uses the acronym more frequently than the left-wing media. However, no significant discursive differences were observed between conservative and progressive press in terms of the language used, which often had negative connotations that stigmatised the young people concerned.
Keywords
Teens, immigrants, press, content analysis, text mining, MENA
Palabras clave
Adolescentes, inmigrantes, prensa, análisis de contenido, minería de datos, MENA
Resumen
Este artículo tiene como objeto abordar el tratamiento mediático realizado por la prensa española sobre una tipología de inmigrante: el menor extranjero no acompañado «MENA». Los medios de comunicación tienen un papel relevante en la creación y difusión de conceptos e imágenes entre el público, de tal modo, promueven la articulación de conjuntos de significados llamados discursos. La investigación se propone identificar los enfoques discursivos construidos en torno a la sigla «MENA» en los principales diarios de la prensa española. Para llevar a cabo esta tarea se han recopilado y analizado todas las piezas informativas publicadas sobre menores migrantes en las ediciones digitales de los cuatro diarios más leídos en España (La Vanguardia, El País, El Mundo y ABC) entre el 1 de enero de 2017 y 31 de octubre de 2019. Estas piezas han sido analizadas mediante técnicas de minería de datos (un área relevante dentro de la ciencia de datos) tales como la observación de «term frequency» y de «inverse document frequency». Estas técnicas, junto a la construcción de redes de correlaciones entre palabras, han permitido observar que el término «MENA» evoca un discurso asistencialista, criminalizador y moralista desde un enfoque adultocéntrico y nacionalista. Asimismo, se concluye que la prensa conservadora usa más la sigla que la prensa progresista, pero sin divergencias significativas en el lenguaje utilizado.
Keywords
Teens, immigrants, press, content analysis, text mining, MENA
Palabras clave
Adolescentes, inmigrantes, prensa, análisis de contenido, minería de datos, MENA
Introduction and state of the art
Child and youth migrations from the Global South to the Global North are part of the current migration scenario (Bhabha et al., 2018). Since the end of the last century, and under the legal designation of “unaccompanied foreign minors”, international agreements have regulated the institutional reception of minors who migrate alone, through their protection by child care systems (Menjívar & Perreira, 2017). As their parents are not present, local administrations assume this task until these adolescents’ reach adulthood. From that point on, the states begin to apply ordinary immigration law to them (Sigona et al., 2017).
This migration, with thousands of children and teenagers crossing borders between economically and geopolitically linked regions, has strained the legal and social interpretation of Human Rights, giving rise to friction between protection of the most vulnerable and control of access to the territory of each state. (Knezevic, 2017). The media, reflecting the current social and political debates on the recognition of “ius migrandi” (Velasco, 2016), have included the figure of the migrant minor as a new actor that is receiving increasing attention (Rosen & Crafter, 2018).
The aim of this article is to describe and analyse the informative treatment recently given by the Spanish press when reporting on migrant teenagers, often under the legal acronym, “MENAS”. Specifically, the study aims to understand the discursive configurations employed by the Spanish press towards this population. Journalistic information is a source that feeds these configurations (Benson, 2013) and constructs an interpretative framing of reality, a determining factor in the assignment of one sense or another in the observation of reality (Koziner, 2013). There are other expressions of public opinion, such as opinion polls, barometers, etc. (Miller, 2006). In this case, the strengths of the approach chosen are based on the breadth and heterogeneity of sources and information that form part of public opinion, notwithstanding other weaknesses of this methodological option, such as the external and general description of the information.
In the case of the Spanish media, the interest in this youth mobility parallels its two most recent migration cycles. During the first decade of the century, several thousand minors arrived in Spain yearly, at a steadily increasing rate, which reached 8,080 new arrivals in 2008 (Quiroga, 2010), the highest number in that initial stage. As of 2009, this figure decreased to 3,261 in 2012. And towards the end of the second decade, since 2017 (6,414 arrivals) there has been a constant increase in detected arrivals, up to the 12,300 minors registered in 2019 (Save the Children, 2018; State Attorney General's Office, 2020). The sector is highly masculinised (only 6% were girls). Of these, and according to the regions, between 75% and 80% come from North Africa (Morocco and Algeria), and the rest from the west of this continent; with very low presence of Asian minors.
As previously occurred in Italy (Giovannetti, 2016), the local authorities responsible for child protection have seen their capacity to implement resources for foster care overwhelmed. This new “migratory crisis” has contributed to a technical and political debate on the care protocols for these minors, with two key characteristics. It is presented as an issue disconnected from its global dimension and, as in other social policies in southern Europe, it is addressed without taking into account the previous experience in institutional protection, accumulated during the first migration cycle (Gimeno, 2018). Due to these two deficiencies, two poles of discourse arise. One implicit and the other explicit. Explicitly, state and regional legislation establishes the obligation to protect any minor in a legal situation of homelessness, regardless of their nationality. Meanwhile, implicitly, the local authorities, depending on the government’s ideological bent, are responsible for avoiding the “pull effect” that a reception with guarantees might entail (Peris, 2015). This tension between child protection and immigration control would be behind the filtering or selection that underlies the protection measures. The media modulate discourses that contribute to the integration or exclusion of this population (Rigoni & Saitta, 2012).
Frame of contemporary migration policies and the media
As G. Simmel stated at the time, the “foreigner” is a model of interaction that is constantly being re-signified in time and space (Santamaría, 2002). The media play a relevant role in this interplay of meanings (Vickers & Rutter, 2016), imbuing the concept of “immigrant” with content (Bleich et al., 2015) and creating symbolic borders between what is one’s own and what is “the other” (Dines et al., 2015). In this way, Spain has entered the debate of whether the MENAS are part of society, as a moral community of belonging. In fact, a large part of the regulatory framework and social intervention practices do not consider these adolescents “our children” (Gimeno, 2013; Suárez, 2004). Santamaría (2002) observed that most of the metaphors used in the informative treatment of immigration described a disturbing and worrying phenomenon. Several studies in the field of youth migration have confirmed this observation (Rosen & Crafter, 2018; Doná & Veale, 2011).
The metaphors of exile and migration use phytomorphic semantic fields (root, roots …) as well as zoomorphic (herds, flocks …) imagery suggesting collective behaviour. Van-Dijk (2005) talks about watery metaphors (flows, surges, avalanches …) that amplify demographic magnitudes and challenge public powers to channel or contain these currents. Of all provenances, Arab-Muslims are the archetype through which immigration is thought of as the antipode to cultural difference (Santamaría, 2002; Mateo, 2017). In this rhetorical context, the term “MENA” has emerged as a designation for certain migrant minors. This term implies a positive definition (they are Moroccan and Muslim) and a negative one that exposes two shortcomings: they are not yet of legal age and are not accompanied (Gimeno, 2013). The second involves not knowing their previous and present relationships, as well as doubting their ability to establish them in the future. The founding work of the IOÉ Collective (1995) on Spanish discourse regarding foreigners identified the appearance and extension of what is irregular or “illegal” beyond the entry procedure, generating a moral panic of administrative justification (Santamaría, 2002). Thus, the procedure plays a role of procedural illegitimacy of mobility (Doménech, 2013; Perelló & Lacomba, 2020).
The media play a determining role in the social construct of threat and security (Horsti, 2003; Balzacq, 2005; Perelló & Lacomba, 2020). Media outlets have the ability to influence public opinion, both by presenting a certain image of a population, and by amplifying or silencing specific voices (Horsti, 2003; Hoekstra, 2015). Van-Dijk (2005) warns that there is a direct relationship between language and the power exercised with it, as the discourses reproduce the cultural hegemony of elites over the subaltern social groups.
Research works in Spain (Checa-Olmos & Arjona-Garrido, 2011; Llorent-Bedmar, 2012), Italy (Terwal, 1996), Finland (Horsti, 2003; 2012) or Belgium (Van-Gorp, 2005) reflect this relationship. In addition to the value that migrant receiving societies place on administrative procedures, contemporary migration policies are also based on various economic, cultural and social paradigms. (Thorbjørnsrud & Ustad Figenschou, 2014; Eberl et.al. 2018) among which this research highlights two that are apparently contradictory: securitisation and humanitarianism (Perelló & Lacomba, 2020). For Bourbeau (2011), securitisation is the process of discursive and institutional integration of a problem within the frameworks of security, police action, control and defence. Security is a matter of survival. In this way, security is the way for states and societies to rid themselves of threats and maintain their independence and functional integration in the face of forces of change acknowledged as hostile. (Buzan, 2007).
On the other hand, humanitarianism differs from humanitarian discourse by the validity of the approach of immediacy and constant emergency that is triggered in the form of exceptionality. Justice and rights are thus replaced by charity or benevolence, which is tantamount to acting from a paternalistic standpoint (Andersson, 2014; Ticktin, 2015). Urgency in the implementation of some reception operations entails the decontextualisation of migratory phenomena, as they are treated as unique and unforeseen crises, when in reality they have a long historical tradition (Ticktin, 2015). Another feature of humanitarian discourse is the outsourcing by the State to private agents of the provision of medical and social assistance to migrants in controlled spaces (Perelló & Lacomba, 2020). The humanitarian construct produces victims and the security construct produces threats (Van-Gorp, 2005; Horsti, 2012), in such a way that the news stratifies the legitimisation by the media of migrant types. It can be said that the discursive construct of the MENAS falls into the category of “threatening social figure” that is moulded as a prototype of the “European anti-subject” (Santamaría, 2002).
Material and methods
To carry out the research, a database was created, containing all the journalistic articles published between January 1, 2017 and October 1, 2019 by the newspapers ABC, El País, El Mundo and La Vanguardia in which reference is made to “foreign minors”. These four periodicals were chosen because they are the generalist newspapers with the largest number of readers according to the General Media Study of the year 2019 and because they allow a minimal representation of the progressive and conservative ideological spectrum (Martínez-Nicolás et al., 2014). This research conceives the progressive press as following an editorial line related to secular, cosmopolitan and egalitarian values, while the conservative one is closer to religious values, centred on the nation-state and individual freedom. The timescale considered the increase in the number of unaccompanied migrant minors in 2017 and ends on October 1, 2019 as the period prior to the campaign –and pre-campaign– of the general elections on November 10, thus avoiding skewing of the results –and the consequent impossibility of comparison with the previous period– that the introduction of the issue studied into the electoral agenda could lead to.
The journalistic articles were compiled –except in the case of El País, which has its own tag for this topic–by searching in the digital newspaper archives of the newspapers for the term “foreign minors”. This search yielded 651 pieces that were processed using “web scraping” techniques, which allow, through the writing of code in R language, the reading, downloading and processing of HTML files by data analysis software (in this case, R-Studio). Thus, a database was created that houses the text of the articles (news, social commentaries and reports), their title, date of publication, the newspaper they belong to and whether or not they use the category “MENA” 1 . This latter variable is of vital importance as, given the object of study, the interest of this research focuses on the 344 journalistic articles (Table 1) in which the term is used. However, this question does not mean that the remaining items are irrelevant. On the contrary, their analysis will play a key role in contrasting what was observed in the pieces in which the acronym is mentioned, that is, the characteristics of the articles in which it is mentioned, comparing them with the features of those items in which it is not mentioned.
Through data scraping or text mining, different content analyses have been carried out, among which the observation of word usage frequencies and the construction of correlation networks between them are noteworthy. Beyond the ability to process large volumes of information 2 , this technique has the main advantage of allowing the use of algorithms that automatically process the corpus of documents, which significantly reduces the likelihood of researchers introducing their own biases in the process. Furthermore, as the processing and analysis operations were carried out using the R programming language, all of them have been recorded in “scripts” 3 that can be reviewed and executed by other researchers in order to verify the outcomes obtained.
Resorting to data mining techniques to carry out a content analysis does not mean giving up the explanatory level of the analysis introduced by the researcher's interpretation. The techniques mentioned allow us to delve into the “informational use of discourse” (Gutiérrez-Brito, 2010: 254), but this does not imply that the analysis should be restricted to what is merely descriptive. The techniques used generate rigorous outputs, which must be interpreted by the researchers, allowing the analysis to move towards the explanatory level of the discourses studied.
Word frequency: tf and tf-idf
To observe whether different discourses can be seen within the corpus of documents studied, as well as to determine which words would make up these discourses, we first resorted to analysing the “term frequency” (tf), that is, the frequency with which the different words appear in the corpus according to different variables. Using tf allows us to observe which words are the most used by each newspaper, or determine whether different words are used in the articles in which the “MENA” category is mentioned. The tf is complemented by “inverse document frequency” (idf), formulated as follows:
The idf allows us to reduce the weight that is given to the words most used –present in a large part of the groups created– and increase the weight that is given to those that seem more specific to each group (Silge & Robinson, 2019). Thus, tf and idf are combined (multiplying both quantities) in the tf-idf of a term, which expresses the frequency of an adjusted category based on how common or exceptional its use is within the “corpus” (ibíd.). Consequently, tf-idf helps find the most specific terms of a discourse and, therefore, those that distinguish it from other discourses.
In the current research, tf and tf-idf was used to try to find differences between the discourses formed by the news from one newspaper and others, as well as to observe whether in the articles in which the acronym «MENA» is used, words different from those in the articles in which it does not appear are used.
Networks of correlations between words
The construction of correlation networks allows the visualisation of relationships between words based on their tendency to appear together within the same article (Silge & Robinson, 2019). Taking the words that make up the corpus of documents as a unit of analysis, their degree of correlation with other words within the same article was studied. To measure this correlation, the correlation coefficient phi (ϕ) between all the words in the corpus was used, calculated for any pair of words X and Y by the following formula:
All these correlations can be represented graphically by Markov chains, as networks of correlations between words. This way, an immense network can be built, consisting of all the words in the corpus in which the intensity of the links will vary depending on the phi between the words connected.
Filtering this network by the most frequent words 4 and keeping only those moderate or higher correlations (ϕ≥0.3) will allow us to discover networks of words that are relevant in the corpus and that are interconnected. So, by analysing the correlations between words, clusters of terms present in the corpus are discovered. (Silge & Robinson, 2019). In this way, different networks can be discovered and, therefore, different discourses present within the corpus of documents. The main advantage of this application of content analysis is that cluster discovery replaces the category construction present in traditional content analyses (Piñuel-Raigada, 2002), which removes the main source of arbitrariness and possible biases from the analytical process.
Analysis and results
Temporal evolution of publications
If we observe the evolution of the articles in which the term «MENA» is used over time, it can be seen that, regardless of the medium, it was barely mentioned by a handful of news items during 2017 (Table 2).
However, this amount grew substantially in 2018 and continued to grow a little more for ABC and El Mundo during 2019, while it decreased slightly in El País and La Vanguardia. However, it should be borne in mind that only the first nine months of 2019 are included, so that year stands out above the previous ones for which the full 12 months were compiled. Based on the number of articles published per month (Figure 1), the growing trend described can be confirmed, as well as highlighting that, since August 2018, there has not been a single month in which at least one of the four media has not published an article in which the acronym is mentioned. The peak of information from La Vanguardia between August and November 2018 was due to the difficulty of the care system for children and teenagers in Catalonia in responding to the arrival of more than 2000 minors from Morocco, which caused a collapse in shelters during the weeks after that summer.
Term frequency
In the four media outlets, the highest frequency words reference the description of the MENA acronym without expressly mentioning the term (“minors” is the word most frequently used in the media and the terms “foreigners” and “(un)accompanied” occupy positions between the second and sixth most frequently used in all cases) except for the newspaper ABC, which uses the MENA acronym more than the descriptors. However, the most relevant finding in the tf studied is the similarity present among the media. Notably, for example, if we observe the twenty-one words most used by each media outlet (those with a tf greater than 0.002) there are only one to three that are only among the most used by one medium and not by the rest 5 . Also taking into account the categories with the highest tf, slight differences are observed between the articles in which the MENA category is mentioned compared to those in which it is not cited. Therefore, a certain discursive convergence and affinity can be seen in the language used by the media, although the inclusion or omission of the acronym marks the path of the informative frame.
Although the tf-idf analysis should highlight words that distinguish different groups of documents, the tf-idf obtained were extremely low, and those terms with a slightly higher value than the others have a very low tf, so they cannot be considered relevant within the news corpus. For example, the word with the highest tf-idf is «Lanzarote», which appears only in the articles in which the MENA category is not mentioned, but its tf-idf is 0.0001288 and its tf is 0.0001858. Both the study of the tf and of the tf-idf seem to indicate a degree of uniformity in the discourse present throughout the corpus studied. This would not allow us, a priori, to talk about different discourses depending on the appearance of the term “MENA” in the article. But Table 1 does clearly reflect that ABC and El Mundo more frequently include the MENA acronym every time they discuss foreign minors. However, it should be borne in mind that the words used by a discourse are only one of its constituent elements. The way in which these categories are related is another of the relevant dimensions of the discourse and, as will be seen below, exploring their relationships will allow us to appreciate different discursive axes present within the corpus.
Correlation network
Figure 2 shows the correlations between the 150 terms most frequently found in the articles in which the MENA category is used. Different clusters of different sizes can be seen in the figure, including the largest, which occupies a central position in the representation. This cluster shows the existence of an important number of terms that tend to appear related in the different articles in the form of two discursive lines. On the one hand, a geographical axis that traces the migratory route. This line links Morocco with the autonomous cities of Ceuta and Melilla, the community of Andalusia (entry points for migration) and ends in the Community of Madrid, with a specific protection centre (Hortaleza). The second axis articulates the different public institutions responsible, including the Andalusian Government, the Ministry of the Interior, Spain’s president, Pedro Sánchez, and the resources necessary for care (“millions of euros”, “reception centres”, “workers” and “places”).
The small clusters are grouped into three sets: in the first set there are three nodes on public policies, the child protection system and the lack of resources for care. The second set groups four nodes that characterise the population in question: boys “17 and 18 years old”, in “the street”, who are “single migrants” and “unaccompanied foreigners”. The last set consists of only one very significant node, as it cites “crimes”, “mossos” (Catalonian police force), “Generalitat”, “Barcelona” and “Catalonia”.
The appearance of the word “crimes” in news items about migrant minors is, by itself, a relevant fact that suggests the presence of a securitising and criminalising discourse. It is even more significant to note that the word “crimes” does not appear within the word network in which the MENA acronym does not appear (Figure 3).
The first cluster observed (Figure 2) contrasts with the one that does not use the acronym. As Figure 3 shows, the cluster that includes the cited terms is also the main cluster, but this one is larger and more compact. In other words, it integrates more categories linked by a greater number of relationships. This difference in size between one cluster and another is due, as can be seen, to the fact that the one in Figure 3 includes elements that are also present in Figure 2 but are not related to the main discourse, such as reference to the ages, social policies or the status of unaccompanied minors of the subjects.
It can be seen, therefore, that in the articles in which the MENA acronym is mentioned, there is a main discourse accompanied by secondary discourses, whereas when it does not appear, the principal discourse is more compact and integrates elements that were scattered in the previous case.
In the smaller scattered clusters, words that refer to the situational status of minors (“alone”, unaccompanied, without care resources, minors, “street”), their foreignness (“migrants”, “countries”, “foreigners”) and, in the case of Barcelona, terms that criminalise them (“mossos” and “crimes”) stand out. By relating the small clusters, it can be inferred, therefore, that these subjects are foreign minors and migrants, alone, without care or company, who come from other countries, who are on the street and who are monitored by the police anticipating the perpetration of crimes. In this way, there is a discursive spectrum of the Spanish newspapers analysed that point to several informative frames regarding this population: state-nationalist, because they emphasise their status as foreigners; protector-assistance, because they highlight their loneliness and neglect; moral-adult-centric, because they highlight their minority of age and their location on the street; securitising-criminalising, because they invoke the security forces and crimes.
Discussion and conclusions
The study has verified the progressive increase in the number of articles in the Spanish press concerning adolescents who migrate alone to Spain between 2017 and 2019. Likewise, there are no substantial differences in the informational framing of the four media outlets, although the word frequencies suggest incipient trends whereby the media emphasise certain social, partisan, police or welfare concepts. The conservative press more frequently includes the MENA acronym every time it discusses foreign minors. Similarly, the study by Boeva (2016) found that conservative newspapers tend to resort to the legal category more frequently than liberal ones and portray refugees in a more negative light.
The way information on foreign teenagers is framed is determined by the inclusion or omission of the acronym MENA. Articles citing foreign minors –without making reference to the acronym– delimit a semantic field characterised by a greater geographical contextualisation, by a human, personal and, sometimes, humanitarian and assistance-oriented outlook, than those articles that include the acronym. The second block of articles is characterised by an approach linked to threat, security, moral disapproval and political intervention. The word MENA describes a homogeneous collective with gregarious behaviour patterns and few contextual and subjective nuances.
Both views emphasise the adult, national and moral condition of the frame. In addition, there is also an emphasis on the solitude of minors, as omissions are appreciated in family networks and the embryonic support and reception networks of which they are part. The articles omit and silence the complexity of migratory processes and transnational social relations.
The work of the IOÉ Collective (1995) on Spanish discourses regarding foreigners identified nationalist, cosmopolitan, racist, ethnocentric, universalist, egalitarian and socially-committed standpoints. Our research identified informative frames close to the discursive positions of protectionist nationalism, ethnocentrism and solidarity, but did not identify the other discursive spectra.
Regarding the different approaches of the conservative and progressive press, the study by Boeva (2016) found that most articles in the Daily Mail, The Guardian, FOX News and ABC News on refugees were related to their nationality or country of origin. Nevertheless, as in our study, Boeva’s found no noticeable difference between conservative and liberal newspapers. It concluded that refugees are represented as “needy” rather than as “threats”, although in our work we observed that the two representations are associated with the inclusion or omission of the MENA acronym.
Our research coincides with other studies (Horsti, 2003; Checa-Olmos & Arjona-Garrido, 2011) in which the pejorative image of immigrants and the dichotomous visions that separate the “good” from the “bad” hold sway. However, we found specific elements of the Spanish context, at the end of the second decade of the 21st century, that reveal a growing ideological polarisation that cuts across political, social and media organisations. This trend accentuates the simplification of perceptions and interpretations of the studied population.
Although the method employed has strengths based on the quantity, breadth and heterogeneity of journalistic information, the empirical material for an in-depth study of the structure of latent psychosocial images and representations in the Spanish press was insufficient. The analysis enables the description of the discursive morphology of the press, but the relationships of meanings are still incipient. It would be convenient to complement this analysis with other studies that analyse and interpret opinion polls on minors who migrate alone. There is a need to continue with the exploration of findings related to the adult-centric and nationalistic news frames of the other print media, as well as the audiovisual media with the highest audience penetration. (1)