10.4185/RLCS-2019-1331en | ISSN 1138 - 5820 | RLCS, 74-2019 | |
The representation of immigration in contemporary Spanish primetime TV series
Carla Portillo Delgado [CV] [ORCID] [GS]. Student in the PhD programme in “Education in the Knowledge Society”. Universidad de Salamanca (USAL) / University of Salamanca (Spain). email@example.com
Keywords Immigration; television fiction; prime time; content analysis.
Translation by CA Martínez-Arcos
According to the surveys carried out the Sociological Research Centre and several statistics studies, immigration and emigration are two of the most concerning issues for Spanish people. Spain has recently become an immigration country  (Díez Nicolás, 2004), with immigrants representing an important share of its population . In addition, in recent years, “as a result of the economic crisis”, an important share of the Spanish population has emigrated to other countries (INE, 2017).
If we consider that our perception of the world is, in part, shaped by the media, it is essential to analyse the media’s representation of this recent phenomenon in Spain (Van Dijk, 1997, 2003). In recent years, the analysis of immigrants in the media has been a recurrent subject of study (Van Dijk, 1989; Entman, 1992; Romer Jamieson and De Coteau, 1998; Dixon and Linz, 2000; Igartua, Muñiz and Cheng, 2005; Igartua, Muñiz, Otero and De la Fuente, 2007).
Research on immigration and ethnic minorities has focused on news coverage and treatment and has pointed out that these are one of the factors causing the increase in xenophobia in the country (Igartua, Cheng, Moral, Fernández, Frutos, Gómez-Isla and Otero, 2008; Igartua and Cheng, 2009). The representation of immigration in the media, especially in the written press and television news formats, has been a fairly researched subject in social sciences (Van Dijk, 1989; Entman, 1992; Romer, Jamieson and Coteau, 1998; Dixon and Linz, 2000; Igartua, Muñiz and Cheng, 2005; Gartua, Muñiz, Otero and De la Fuente, 2007; Igartua, Moral and Fernández, 2011; Igartua Muñiz, Otero, De la Fuente, 2013). These studies have concluded that the image of the immigrant and ethnic minorities is associated with socio-economic and cultural threats, aberration, crime and violence in the media (Van Dijk, 1997 and Cea D’Ancona, 2004). This is because media representations of immigrants are based on stereotypes, which generates stereotyped and prejudiced perceptions of immigration in the audience. Social science researchers have been carried out numerous studies on the informative treatment of immigration in the media and its socio-cognitive effects. These studies have indicated that there are more negative news pieces on immigration than positive ones (for example, on the positive contribution of immigration to host countries), and that representations of immigration tend to associate it with delinquency, crime and other social problems (Van Dijk, 1989; Van Gorp, 2005; Igartua, Muñiz and Cheng, 2005; Igartua, Muñiz, Otero and De la Fuente, 2007; Igartua, Moral y Fernández, 2011). In this regard, Van Dijk (1994, 1997), by way of example, argues that the media increasingly associate immigrants, refugees and ethnic minorities with socio-economic and cultural threats, aberration, crime and violence .
On the other hand, research on the socio-cognitive effects of news frames of the subject –the treatment performed by the media– has pointed out that frames themselves influence perception, attitudes and beliefs about immigration in the host country (Domke, McCoy and Torres, 1999; Brader Valentino and Suhay, 2008; Igartua and Cheng, 2009; Igartua, Moral and Fernández, 2011). One of the main consequences of this informative treatment is the formation and/or maintenance of certain stereotypes  and prejudices about immigrants. In this sense, Seiter (1986) has pointed out that the media, especially television, are actively involved in the creation of stereotypes.
The image of immigration in television fiction has not been extensively analysed in Spain despite fiction is a basic component of the primetime television programming (Ruiz-Collantes, Ferrés, Obradors, Pujadas and Pérez, 2006; Galán, 2006; Lacalle, 2008; Marcos et al., 2014). Using content analysis to “observe and study stereotypes and the representation of immigrants”, Galán (2006) conducted a study of two TV series of great trajectory in national fiction -El Comisario (“The Commissioner”) and Hospital Central (“Central Hospital”)- and found out that they offered “discriminatory or biased representation of immigrants” (2006)
Lacalle (2004) has pointed out that, up until approximately 2003, immigrants barely appeared in Spanish fiction and that when they did, they used to be characterised as secondary characters and were associated with negative stereotypes, playing the protagonist’s friend role, and always occupying a circumstantial and passive position. Their professions tended to be related to the service sector, the entertainment industry or domestic work and illegal activities. In a study carried out later, Lacalle (2008) indicated that the image of the immigrant in national fiction television was predominantly associated to an irregular situation and a low education level. There was a high presence of immigrants who acted as criminals or were victims of crimes or violent actions, while the presence of immigrants with a high education level or a central narrative role in the series analysed was uncommon. According to Ruiz-Collantes, Ferrés, Obradors, Pujadas and Pérez (2006), the image offered by national fiction is: non-leading characters, whose representation is mainly negative, associated with problems and victimisation. Moreover, immigrants are shown to be uncapable of achieving their goals and resorting to simulation, manipulation or deception.
The studies of Ruiz-Collantes, Ferrés, Obradors, Pujadas and Pérez (2006) and Lacalle (2008) focused on the analysis of immigrant/foreign characters in national fiction, as in the research presented here. Marcos et al. (2014) analysed 114 TV shows and 2,623 characters, the largest sample examined to date, and all the primetime programming, regardless of their national origin and format -taking into account series and feature films. The authors concluded that immigrants were underrepresented in primetime fiction and that immigrant/foreign characters mostly played background roles, so they are scarcely represented because of their scarce appearance and narrative weight. When these characters appear in fictional series, they are more likely to be illiterate, to perform unskilled jobs and lack a stable occupation than native/national characters. They usually play antagonist roles, so they are involved in plots where there is a greater presence of violent acts. Moreover, the socioeconomic level of immigrants is lower than that of native/national characters. Since they are background characters, they do not intervene in many plots and, consequently, do not participate in many dialogues, so their conversational performance is much lower than that of natives/national characters. In addition, immigrant/foreign characters are defined with more negative personality attributes: they are more aggressive, conflicting, unfair or treacherous and intolerant than native/national characters.
The research presented here examined a relatively wide sample of programming by means of content analysis as a research method, which allowed us to carry out a socio-demographic analysis of the representation of immigrant characters in primetime Spanish television series broadcast by the major mainstream channels during 2016 and the first half of 2017. Characters were not observed in isolation but taking into consideration their relationships with other characters. This allowed us to conclude how immigrant characters were portrayed, which is a strategy that has been used in previous studies on the image of ethnic minorities in television fiction (e.g., Mastro and Greenberg, 2000, and Harwood and Anderson, 2002, Marcos et al., 2014). Research on the image or representation of immigrants and ethnic minorities, especially African-Americans and Latinos, in the television fiction, and previous research on the analysis of news framing of immigration were taken into account. Attention was also paid to characters’ gender to determine whether there was equality in the representation of male, female and non-binary characters, although this article does not present the results on this aspect .
Based on the objective of evaluating the status of the representation of immigration in primetime Spanish television series, we have established the following hypotheses:
(1) Immigrants/foreigners will be underrepresented in the primetime fiction programming broadcast on national TV channels, considering their real demographic weight in Spain.
(2) Immigrant/foreign characters will play secondary or background roles to a greater than native/national characters, and will appear to a lesser extent in leading roles.
(3) Immigrant/foreign characters will have a lower level of educational, a lower socio-economic level and less-skill occupations than native characters.
The selected methodological strategy was content analysis, which “includes special procedures for the processing of scientific data” (Krippendorff, 1990, p. 28) and allows us to quantify data and provide objective conclusions, supported by numbers representing real phenomena. As stated by Juan José Igartua (2006, p. 180),
As a research technique, content analysis is very useful and necessary in the social sciences because it allows us “to formulate, based on certain data, replicable and valid inferences that can be applied to their context” (Krippendorff, 1990, p. 28). It also treats data not as physical events but as symbolic phenomena. Therefore, content analysis has become one of the most widely used techniques in this field.
This method was fundamental because it allowed us to examine characters as a basic unit  and allowed us to work on different aspects in the other studies. In this way, the programme as a whole and the characters were treated as units of registration, or meaning, as Bardin points out, since they are “the segment of content that will be necessary to consider as base with a view to the categorization and frequency count” (Bardin, 1996, p. 79). It is in this step where the variables are measured, where numbers are attributed to the manifestations of the analysis unit. In addition, we must “submit these numbers to certain mathematical techniques” (Igartua, 2006, p. 203) that allow us to draw quantitative conclusions with which to develop theories after the analysis of these data.
2.1.1. Population and sample
The total sample was composed of 26 programmes and 723 characters detected through the analysis of the fiction TV series , excluding co-productions, broadcast throughout 2016 and the first half of 2017 by the six national mainstream channels: La 1, La 2, Antena 3, Cuatro, Telecinco and La Sexta. Together these networks reach a 66.5% share (Barlovento Comunicación, 2017). The sample selection resulted from the analysis of the TV programming of national fiction broadcast in the two-year period of analysis. Later, we coded the most-watched episode according to Kantar Media. A total of 26 programmes and 723 characters were identified (Table 1). La 1 provided 42.3% of the programmes and 42.2% of the characters; Antena 3 provided 38.5% of the programmes and 39% of the characters; Telecinco provided 15.4% of the programmes and 16.2% of the characters, and La Sexta, 3.8% of the programmes and 2.6% of the characters. Neither La 2 nor Cuatro provided programmes nor characters for the analysis of this work (table 2).
Table 1. Description of the analysed sample
Source: Authors’ own creation.
Table 2. Description of characters and programmes by networks
Source: Authors’ own creation.
If we take into account the data of the 26 programmes included in the sample, the programme with the lowest audience was iFamily (La 1, 2017, SO1e01 “And suddenly a stranger”), with a 8.7% share, while the programme with the highest audience, with a 29.2% share, was El Príncipe (Telecinco, 2016, S02e18 “Inghimasi”), so the sample range was 20.5. The average audience of all programmes was 18.08% (DT=5.78). Taking into account the broadcast network, the average audience of the episodes of the series was as follows: Telecinco, 22.72%; Antena 3, 21.2%; La 1, 14.75%; La Sexta, 7.6%. Neither Cuatro nor La 2 provided any primetime fiction series in the period of analysis (Kantar Media).
2.1.2. Data collection instruments
For the analysis of the fictional programmes and their characters, a coding scheme was developed based on the work of Neuendorf et al. (2010), Marcos Ramos et al. (2014), Álvarez-Hernández, et al. (2015). This coding scheme articulates, therefore, the variables to analyse around the typology of characters, the narrative roles and the social sphere of characters object of this study. Thus, we collected data on the name of the variables, sub-variables and categories, as well as their definitions.
For this study, we established nine large blocks of variables that were, in turn, divided into other sub-variables and categories that provided fundamental data to analyse the unit of analysis: the character. The nine large blocks of variables are:
1. Basic identification data. The following aspects were evaluated: character number (analysis unit number), programme number, coder number, broadcast year, and programme’s broadcast television network.
2. Character type (Mastro and Greenberg, 2000). It was evaluated with the following code: 1=leading (its presence is essential for the development of the narrative), 2=secondary (it is involved in the narrative, but it is not essential to it), 3=background (its presence is non-essential, peripheral or very episodic).
3. Character’s socio-demographic aspects. The following variables regarding each character were evaluated: a) sex (1= cisgender male, 2= cisgender female, 3=other, non-binary, including transgender and intersex persons); b) sexual orientation (1=heterosexual, 2=homosexual, 3=bisexual and 4=other (asexual, pansexual, demisexual); c) age group (1=child, 0-12 years; 2=teen, 13-17 years; 3=young adult, 18-30 years; 4=adult, 31-64 years; 5=elderly, 65 and over; d) educational level (1=no studies, 2=compulsory studies, 3=university studies); e) socio-economic level (1=low, working or lower class, cannot satisfactorily meet basic needs with income; 2=middle class, the character works for a living, meets its needs and can afford small luxuries; 3=upper class, characters that do not need to work to maintain their living standards or have a job that allows them to enjoy many luxuries not accessible for most people); and e) occupation, coded according to a list established by Spain’s Sociological Research Center (CIS, for its initials in Spanish), which includes 17 different professions; f) religion (1=religious; 2=non-religious); g) marital status (1=single, 2=married or living as a couple, 3=divorced, 4=widower) and h) change of marital status (1=no change, 2=to married or living as a couple, 3=to divorced or split, 4=to widower). For all these variables, 99 was used for those that could not be coded.
A relevant aspect of the present study was the coding of the character’s nationality. Given that it was often difficult to discern the place of birth of the character, the identification of this criterion was based on a set of traits or attributes that had to be evaluated jointly or separately: a) the character’s place of birth (provided an explicit mention was made on this aspect in the programme); b) birthplace of one of the parents of the character, considering the possibility of the character being a “second generation immigrant” (when at least one of the parents had been born outside the country); c) biological characteristics or phenotypical traits (such as eye shape, skin colour and hairstyle); d) cultural characteristics (such as way of dressing, name, accent, etc.); and e) reason for being in the country (work, studies, holidays). Nationality was assessed by taking into account the country in which most of the action in the narrative time took place. Taking as a reference the abovementioned criteria, the following code was used to classify the nationality of the character: 1=citizen of the country where the main narrative action takes place (native, if living in his or her country of origin); 2=foreigner, i.e., a person born or coming from a country other than the one he or she resides temporarily (for studies, holidays or business); 3=immigrant, i.e., a person who comes to live permanently in a foreign country with a concrete work project; a character could also be coded as “immigrant” (second generation) if at least one of the parents had not been born in the country where the main action takes place and had settled in another country for work reasons. For the purposes of coding, both the foreign and immigrant sub variables were united to better analyse the data.
Another relevant aspect is the geographical origin of the character, which was coded in the following way: 1=Spain; 2=Another European country; 3=the United States; 4=Canada; 5=Latin America; 6=Asia; 7=Africa Y, 8=Oceania. The ethnicity of the character was also coded: 1=Caucasian; 2=African American/African; 3=Asian/East Asian; 4=Arabic/Middle East; 5=Latin American; 6=Gypsy; 7=other. For all these variables, 99 was used for those that could not be coded.
4. Character’s narrative level. It assessed whether the character had defined goals (0=no, 1=yes) and, if any, whether they were related to its personal life (0=no, 1=yes), job (0=no, 1=yes), and whether it pursued them actively (0=no, 1=yes) or passively (0=no , 1=yes). It also analysed the way it pursued these goals: sex (0=no, 1=yes), violence (0=no, 1=yes), ethically (0=no, 1=yes).
Another sub-variable that was measured within this section was the character’s hypersexualisation. To determine it, we observed the character within the whole of the series based on five concepts, four of which were proposed by the Geena Davis Institute of Gender in Media, in its report titled “Gender Bias Without Borders” (2014): use of sexually suggestive clothing, naked (partially or fully if it is not a power fantasy), thinness (minimum amount of body fat or muscle) and attractive (comments are made about their physical appearance). The fifth criterion is the erotic focus of the camera on a fragmented part of the body of the character. If in the episode analysed, at least three of these concepts are at some point applied to the character, it will be considered as hypersexualised (0=no, 1=yes).
4. Social sphere of the character. This variable explores the social interactions of the coded characters, according to gender-related criteria extracted from the Test Bechdel-Wallace (Bechdel, 1985, p. 22). Interaction with other characters was measured exclusively in the following way: 1=interacts mainly with men; 2=interacts mainly with women; 3=interacts mainly with non-binary characters. The analysis also considered conversations with other characters, measuring whether the character spoke with characters of the same gender (0=no, 1=yes). In addition, we coded the topic of the conversations: if the character identifies as a woman, does she talk to other women about something besides a man (0=no, 1=yes). and if it identifies as a man, does he talk to other men about a woman (0=no, 1=yes).
5. Character’s violent behaviour. A dichotomous scale (0=no; 1=yes) was used to code the degree of presence of violent behaviours or forms of violence based on the classification developed by Potter and Warren (1998): a) performing “major physical attacks”; b) performing “minor physical attacks”; c) performing acts that cause “property damage”; d) performing acts of “intimidation”; and e) making “hostile comments”
6. Violent behaviours performed against the character. Based on a dichotomous scale (0=no; 1= yes) and the classification developed by Potter and Warren (1998), we coded whether the character suffered from or was a victim of the following types or modes of violence: a) major physical attacks; b) minor physical attacks; c) acts that cause property damage; d) acts of intimidation; and e) hostile comments.
7. Problematic health behaviour shown by the character. We coded (1=yes, 0= no) whether the character analysed: a) consumes alcoholic beverages; b) smokes tobacco; c) uses prescription drugs; d) uses illegal drugs; and e) has an eating disorder.
8. Conversation topics engaged in by the character. A dichotomous scale (0=no, 1=yes) was used to code whether the character conversed with other characters at some point during the programme about the following topics: love, violence, friendship, sex, money, machismo, work, environment, health, education, family, politics, sports, racism, Immigration and empowerment.
9. Character’s personality traits. Taking as reference the work carried out by Igartua, del Rio, Álvarez et al. (1998), we evaluated on a three-point scale (1=not characteristic of the character; 2=partially or moderately defines the character’s personality; 3=perfectly defines the character’s personality; 99=cannot be coded), to what extent the following traits characterised the personality of the analysed character: friendly, open (extrovert), good (good-hearted), disloyal, unfair, treacherous, aggressive, intelligent, hard-working, distrustful, thankful, conflicting, racist and tolerant.
2.1.3. Coding and reliability
Given that the next step, coding, is one of the most important, “it must be carried out in a systematic way” (Igartua, 2006, p. 212), since one of the objectives of content analysis is that it can be replicable, i.e., that “any analyst who repeats the process must reach the same conclusions” (Igartua, 2006, p. 212). It is essential that all analysts have internalised each of the variables and categories, understand the same and understand the process perfectly. As Krippendorff (1990, p. 104) points out, “observers, coders and judges must be familiar with the nature of the material to be coded, but they must also be able to reliably manage the categories and terms that make up the data language. It is not easy to comply with this double requirement”.
Once the coding of the whole sample was completed, we analysed 10.23% of the characters (n=74) in the total sample to calculate the reliability of the coding process. Intercoder reliability was calculated with the observed agreement coefficient (OA) and Krippendorff’s alpha (αk) (Igartua, 2006). The use of these two instruments is justified by the fact that it has been shown that Krippendorff’s alpha coefficient yields very low coefficients even when levels of simple agreement are high in variables whose data are very skewed  (Lovejoy, J., Watson, B. R., Lacy, S., & Riffe, D., 2016, pp. 4-5). Therefore, reliability was calculated with Krippendorff’s alpha in the appropriate variables while the observed agreement coefficient was performed on the remaining variables. If we take into account the 64 variables considered , the mean observed agreement in the index (OA) was .87 (DT: 0.38), while the mean obtained in Krippendorff’s alpha coefficient was αk =.74 (DT: 2.12). Both are sufficiently high reliability values (Igartua, 2006, P. 221).
Table 3. Reliability data
Source: Authors’ own creation.
Two of the most important and sensitive variables for the study are nationality and geographical origin of the character, as they allow us to determine the number of natives and immigrants/foreigners to analyse in the research, so it was of vital importance for these variables to obtain sufficiently high values in the indices of the intercoder reliability indices. Based on the obtained indices, the sample is considered reliable: the variables “nationality” and “geographical origin” yielded very acceptable values: αk =.82 and αk =.95, respectively. However, it is important to note that the results of the variables not discarded around the value αk =.60, but lower than αk =.70, should be considered tentative and therefore interpreted cautiously (Neuendorf, 2002).
This section presents the findings related to the hypotheses posed about immigrant/foreign and national/native characters with respect to their nationality and geographical origin, character type, educational level and professional status.
In this sense, hypothesis 1 tried to analyse whether there was the percentage of the immigrant/foreign population represented in Spanish TV series is the same as the one in the official records of the Spanish society. Since the TV series that were analysed were broadcast in 2016 and 2017, we examined the 2017 data provided by Spain’s National Statistics Institute (INE).
The analysis of the nationality of the 710 characters in the sample (the total sample was 723 characters, of which 13 had lost values) indicates that 89% are native characters, that is, characters born in the same country in which the fictional narrative takes place. As for immigrant/foreign and emigrant  characters, they represent 7.1% and 3.9% of the sample, respectively.
Table 4. Relationship between character nationality and geographical origin (% column)
Source: Authors’ own creation.
In order to make a comparison, demographic data were reviewed to see to what extent immigrants/foreigners were represented in fiction television. According to Spain’s National Statistics Institute (INE, 2017), as of 1 January 2017, Spain had a population of 46,528,966, of which 9.5% were registered as foreign, representing 4,424,409 people. So, if we look at the sample, the percentage of immigrant/foreign characters is 7.1%. According to the data, it could be said that there is an underrepresentation of immigrants/foreigners in national fiction television, in this case by a difference of 2.4 percentage points with respect to their real demographic weight in Spain.
On the other hand, it is important to note that the total sample not only included TV series set in Spain with the presence of Spanish nationals and immigrants/foreigners from different countries, but also contemplated series mostly set in Germany (Buscando el norte, Antena 3, 2016) and Thailand (La Embajada, Antena 3, 2016). In this sense, the condition of Spanish emigrants is also represented in contemporary Spanish fiction television, reaching 3.9% of the characters. If we compare this percentage with data offered by Spain’s National Statistics Institute (INE, 2017b), which indicates that 1.7% of the people born in Spain live abroad, it can be concluded that the Spanish emigration is overrepresented by a difference of 2.2 percentage points in the national fiction television .
Table 5. Relationship between character’s geographical origin and nationality (% column)
-Value statistically lower than the total percentage (analysis of adjusted standardised residuals).
Another element to be taken into account in this section is the origin of immigrant/foreign characters, as shown in table 5. In this sense the comparison with the data provided by Spain’s Statistics Institute (INE, 2017) with respect to the nationalities with greater immigration to Spain in 2016 is significant. According to Spain’s INE (2017), immigrants come, ordered from highest to lowest, from: Africa (Morocco), Europe (Romania) and Latin America (Colombia). China occupies the eighth place in the representation of immigrants in Spain today and the United States, compared to what happens in the sample analysed, has no statistical presence within the first fifteen nationalities representative of immigration in the country. Therefore, the geographical origin of the characters does not coincide with the demographic data, being the Latin American and African representation especially relevant.
In short, the analysis of the data confirms the initial hypothesis, which proposed that there was an underrepresentation of immigrant characters and an overrepresentation of Spanish emigrants in primetime national fiction television.
The second hypothesis proposed that there would be a relationship between the type of character –leading, secondary or background- and its nationality. In this way, it was expected that, in comparison with native/national characters, immigrant/foreign characters would occupy secondary or background roles to a greater extent and leading roles to a lesser extent.
The distribution of characters by type was as follows: the majority were background characters –472 characters that represent 65.7% of the total sample-, followed by secondary characters -190 characters, 26.5%- and, lastly, leading characters -7.8% (n=718 with 5 cases of missing data). Of all the 627 valid characters  coded as native, 65.6% –411 characters- played background roles, 26.5% –166 characters- performed secondary roles, and 8% –50 characters- played leading roles in the narrative. With regards to immigrant/foreign characters, the distribution was as follows: 64.1% (50 characters) were coded as background; 28.2% (22 characters) as secondary; and 7.7% (6 characters) as protagonists.
Table 6. Relationship between character’s nationality and role type (% column)
(χ2 [2, N = 705] = .107, p < .948)
Pearson’s chi square statistic (χ²) was used to determine the extent to what the two variables were independent or not. We observed a non-statistically significant relation between the two variables (χ2 [2, N = 705] = .107, p <.948), since the percentage of leading characters was similar between immigrant/foreign and native/national characters. In secondary characters it was slightly higher –the difference is 1.7 percentage points- in immigrant/foreign characters. Finally, immigrant/foreign characters played background roles in slightly lower proportion than native/national characters (64.1% versus 65.6%), so this hypothesis seemed to be rejected by the data.
However, the fact that some of the TV series analysed were not set in Spanish territory and that, therefore, the Spanish characters involved in them were coded as immigrants, is determinant in the reading of these data. To check whether there was really a statistically significant relationship between the type of character and its immigrant status, we separated immigration from emigration through the correlation of the variables “character type” and “geographical origin”.
Table 7. Relationship between character’s geographical origin and character type (% column)
-Value statistically lower than the total percentage (analysis of adjusted standardised residuals).
A statistically significant relationship was observed between the geographical origin and type of character with the Pearson’s chi square (χ2 [2, N = 711] = 10.964, p < .004) so that Spanish characters were more likely to carry more narrative weight than characters from other countries, as proven by the following data: of the characters from the rest of the world, 82.9% played background roles, 15.7% played secondary roles and only 1.4% were protagonists. On the other hand, Spanish characters were protagonists in 8.6% of the cases; secondary in 27.6% and background in 63.8%.
Therefore, the hypothesis is partially confirmed. Immigrant/foreign/emigrant characters (Spanish emigrants in fictions set in other countries and immigrants or foreigners in Spain) do not show statistically significant differences by type of character. However, there are differences according to characters’ geographical origin, comparing characters of Spanish origin and characters with the greatest narrative weight.
With regards to hypothesis 3, we analysed the educational level and professional status of the characters, assuming based on previous studies that immigrant/foreign characters, in comparison with native/national characters, would have a lower educational level, a lower socio-economic level and they would perform lower-skilled professions. To work with the data in this hypothesis we recoded the “nationality” variable, separating immigrants/foreigners in Spain from Spanish emigrants in series set in another country.
With respect to the educational level, we detected statistically significant differences (χ2 [4, N = 470] = 9.734, p < .045). There was a high percentage of cases (35%) in which the educational level of the character could not be identified. In the cases in which it was possible to do so, there is an outstanding high percentage of characters with university education (60.2%) (n=470 with 253 lost cases). Likewise, in the comparison by nationalities, the high percentage of Spanish emigrants with university studies (94.4%) stood out, as well as the lack of representation of characters with compulsory studies and without studies (2.1%). For their part, the percentages of university education are slightly higher in immigrants/foreigners than in native characters (62.1% vs. 58.6%), which partially refutes the hypothesis.
Table 8. Relationship between character’s nationality and educational
-Value statistically lower than the total percentage (analysis of adjusted standardised residuals).
On the other hand, a statistically significant association was observed between the nationality and economic level of the character (χ2 [4, N = 666] = 41.633, p < .000). Immigrant/foreign characters (17.8%) and Spanish emigrants (22.2%) appeared more frequently with a low socio-economic level than native/national characters (6.1%); appeared less frequently with a middle socio-economic level; and were portrayed as upper-class to a greater extent than natives/national characters. The polarisation at the low and upper-class levels of immigrants/foreigners and especially of Spanish emigrants is symptomatic of the differences in migratory processes. So, for example, the characters in La Embajada are Spanish emigrants with a high standard of living while emigrants from Buscando el norte are in an economically unfavourable situation, because this series narrates the experiences of a group of Spaniards who had to emigrate due to the economic crisis.
Table 9. Relationship between character’s nationality and socio-economic level (% column)
-Value statistically lower than the total percentage (analysis of adjusted standardised residuals).
Finally, we analysed the relationship between the character’s nationality and occupation. A statistically significant association was observed in terms of occupation with Pearson’s contrast test (χ2 [34, N = 710] = 80.633, p < .000). Thus, there is a statistically significant higher percentage of mid-level technicians and students among native/national characters than among immigrants/foreigners. However, the percentage of criminal, police/military activities is considerably higher among immigrants/foreigners than native/nationals, as shown in Table 10. This last finding, a priori unexpected, partly responds to the inclusion in the sample of an episode of El Caso: crónica de sucesos that narrated a murder on an American military base. Finally, the percentage of unemployment was significantly higher among Spanish emigrants.
By way of summary, the educational level of immigrant/foreign characters was not lower than that of native/national characters, but Spanish emigrants were portrayed as highly qualified. In addition, immigrants, foreigners and migrants presented a polarisation in the socioeconomic level, while native/national characters occupied the middle class. With regards to occupation, there was an overrepresentation of immigrants/foreigners in criminal and police/military activities, as well as of technical or mid-level occupations. The overrepresentation of unemployed migrants is also significant. Thus, it can be concluded that the fourth hypothesis was only partially confirmed by the data.
Table 10. Relationship between character’s nationality and main occupation (% column)
-Value statistically lower than the total percentage (analysis of adjusted standardised residuals).
4. Discussion and conclusions
Analysing the representation of immigration on television is a good way to examine the perceptions of the Spanish population regarding this group because a large part of what of people’s perceptions is conditioned by the opinion of the media. This study should be seen as a continuation of the previous studies carried out in Spain (Marcos Ramos et al., 2014).
One of the first questions to be answered was the extent to what immigrant/foreign characters appeared in the primetime national fiction television. The objective was to measure how similar the proportion of immigrant/foreign characters was to that of Spain’s real demography. Thus, it was found that there was an underrepresentation of immigrant/foreign characters in the national fiction television. Surprisingly, there was an overrepresentation of emigrant Spanish characters in fiction series set in a country other than Spain. In other words, the national fiction series set in Spain included few immigrant/foreign characters while fictional series set abroad included a proportion of Spanish emigrant characters that was higher than that provided by official figures.
A more revealing finding is that immigrant/foreign characters mostly perform background roles while Spanish characters perform leading roles to a greater extent even when the fiction series is set outside of Spain. According to the data obtained in this research, it can be argued that immigrant/foreign characters, in comparison to emigrant characters, exhibit significant differences in terms of educational level, which is higher in emigrants, but there are no differences in the socio-economic level, which is low in both groups. This is in line with the fact that both immigrants/foreigners and migrants have the same goal: the desire to prosper in life outside their countries of origin.
An aspect in which the representation of immigrant/foreign characters has improved in national fiction television versus previous research (Marcos et al., 2014) is the portrayal of their educational level. Studies previous found that immigrants/foreigners were characterised by low educational levels, while in this study their education level is higher, and at the same level as native/national characters. This data has corroborated the results of the National Immigrant Survey report (Reher et al., 2008), which indicated that 59% of immigrants have completed first and second year of secondary education, 17% have higher education studies and only 23% belong to the elementary and non-educated groups. These data are very similar to those shown by the Spanish population.
The results of this study are consistent with the results of studies carried out in these areas and with previous studies developed in the United States in relation to the representation of ethnic minorities in television fiction (Mastro and Greenberg, 2000; Mastro and Behm-Morawitz, 2005; Marcos et al., 2014).
On the one hand, we can talk about the low presence of immigrant/foreign characters in fiction. The lack of diversity in television fiction can condition their social visibility and power since a percentage of the population in Spain is not only not represented but is also invisible for a certain population sector that, for various reasons, may not have contact in their daily lives with immigrant/foreign people and whose only means of contact with this sector is through the media, especially television, which is the medium with the highest level of penetration and accessibility.
This lack of media visibility makes it more difficult for the native population to establish a vicarious parasocial contact with characters from other nations that have a notable presence in Spanish society (Harwood and Anderson, 2002 and Ortiz and Harwood, 2007). If we add the fact that when they appear, the image of immigrants/foreigners tends to be stereotypical and/or negative, it can reinforce or promote prejudiced attitudes towards immigrants.
The media can play the opposite role, that is, not to encourage and maintain prejudice, but to contribute to changing attitudes and beliefs about immigration by reflecting, for example, other immigrant/foreign character models, and to promote vicarious contact which, as shown, can have a positive impact on the reduction of prejudice and, therefore, favour the establishment of more harmonious relations between citizens of different ethnic and national origins (Müller, 2009; Igartua, 2010; Park, 2012). This way, fiction television narratives should depict positive interactions and contacts between native/national characters and immigrant/foreign characters, in a way that contributes to the reduction of the negative perceptions regarding immigrants/foreigners, who are considered by society as a threat to them, and encourages the improvement of relations between both groups.
[1 ]For the purposes of this study, the term “immigrant” refers to all the people who have left their country to settle in another country permanently regardless of their motives. Thus, a political exiled, for example, will be considered an immigrant for the study.
 According to the Spanish National Statistics Institute (INE, 2017), as of 1 January 2017, the total population of Spain was 46,528,966 inhabitants. Of this total, 42,104,557 have Spanish nationality and 4,424,409 are foreigners, representing 10.5% of total registered population. This research will work with 2016 data, to be able to compare the demographics of fiction television and the real world. Immigration increased by 21.9%, while emigration fell by 4.6% over the previous year.
 This article is part of a larger study on the state of immigration and gender in Spanish primetime fiction television series. For this reason and for space limitations, this article only presents the results regarding the first topic.
 The unit of analysis was the individual character. The analysis of characters will focus on those who are human, leaving aside animals, extra-terrestrials, fantasy or science fiction beings and animated characters (cartoons). Among the human characters, the analysis only takes into account those who meet the following requirement: for a character to be part of the analysis he or she must appear throughout the programme and engage in dialogue with other characters (talking individuals) (Koeman, Peeters and D’Haenens, 2007).
 The term “fiction” refers to “the simulation or illusion of reality” (Estébanez Calderón, 2002, p. 411) that is produced in the artistic -and in this case audiovisual- invention through the representation of beings and events that develop in an imaginary world. To determine what is meant by a fictional programme, we took into account the following definition: format intended for entertainment, with a clear narrative structure (presentation, conflict, resolution) and a cast of leading, secondary and background characters that are involved in the action.
 "Another controversy involves Alpha, Pi, and Kappa and the fact that they can produce very low coefficients even when levels of simple agreement are high (Feng, 2015; Gwet, 2008; Zhao et al., 2012), which can occur when data distributions are skewed (e.g., most of the coded units are in one category; see Riffe et al., 2014). Krippendorff (2013b) labeled this “insufficient variation” (p. 319), writing that such data “. . . cannot be correlated with anything either, their analytical meanings are largely void, and they cannot convey sufficient information from the analyzed text to the research question” (p. 320). That conclusion seems to ignore the fact that there have been, and will continue to be, populations with skewed distributions of categories that are nonetheless important to study. For example, Robinson and Anderson (2006) studied portrayal of older characters in animated children’s television. Only 8% of characters were older and of these 107 characters, only1% were African American. The authors reported simple agreement to assess reliability. Monk-Turner, Heiserman, Johnson, Cotton, and Jackson (2010) found only 5% of primetime TV characters were Hispanic and fewer than 2% were Asian American. As with Robinson and Anderson, the article reported only simple agreement. The authors do not report why they did not provide chance-corrected reliability coefficients, but it may be because of the skewed distribution phenomenon", in Lovejoy, J., Watson, B. R., Lacy, S., & Riffe, D. (2016, pp. 4-5).
 A priori, Spanish emigrants were coded as immigrants/foreigners if the series was set in a country other than Spain. However, given the specific characteristics of this group, it was decided to recode the variable by separating immigrants, foreigners and emigrants.
 As of 1 January 2017, 794,209 people born in Spain lived abroad, which represents 33% of the total number of Spaniards living abroad (INE, 2017b, p. 2). Therefore, 1.7% of the total population in Spain (46,528.966) plus emigrants born in Spain (794,209).
 In the coding process, the nationality of 1.8% (13) of the characters could not be coded because it was not sufficiently clear. The type of character was not identified in 0.7% (5) of the cases. These data were treated as lost system data.
C Álvarez-Hernández, B González-de Garay-Domínguez & FJ Frutos-Esteban (2015): “Representación de género. Las películas españolas contemporáneas de adolescentes (2009-2014)”, en Revista Latina de Comunicación Social, 70(8), pp. 934-960. doi: 10.4185/RLCS-2015-1079.
L Bardin (1996): Análisis de contenido. Madrid, España: Ediciones Akal.
Barlovento comunicación (2017): Análisis mensual del comportamiento de la audiencia televisiva (diciembre 2017). Disponible en:
T Brader, NA Valentino & E Suhay (2008): “What triggers public opposition to immigration? Anxiety, group cues, and immigration threat” en American Journal of Political Science, 52(4), pp. 959-978.
MA Cea-D´Ancona (2004): “La activación de la xenofobia en España. ¿Qué miden las encuestas?” en Colección Monografías, 210, Madrid: Centro de Investigaciones Sociológicas/Siglo XXI.
J Díez-Nicolás (2004): El dilema de la supervivencia. Los españoles ante el medio ambiente. Madrid, España: Ed. Obra Social Caja Madrid.
TL Dixon & D Linz (2000): “Overrepresentation and underrepresentation of African Americans and Latinos as lawbreakers on television news” en Journal of Communication, 50(2), pp. 131-154.
D Domke, K McCoy & M Torres (1999): “News media, racial perceptions and political cognition” en Communication Research, 26(5), pp. 570-607.
R Entman (1992): “Blacks in the news: television, modern racism and cultural change” en Journalism Quarterly, 69(2), pp. 341-361.
D Estébanez-Calderón (2002): Diccionario de términos literarios. Madrid, España: Akal.
E Galán (2006): "La representación de los inmigrantes en la ficción televisiva en España. Propuesta para un análisis de contenido. El Comisario y Hospital Central” en Revista Latina de Comunicación Social, 61. Disponible en http://www.ull.es/publicaciones/latina/200608galan.htm
J Harwood & K Anderson (2002): “The presence and portrayal of social groups on prime-time television” en Communication Reports, 15(2), pp. 81-97.
JJ Igartua (2006): Métodos cuantitativos de investigación en comunicación. Barcelona, España: Bosch.
JJ Igartua (2010): “Identification with characters and narrative persuasion through fictional feature films” en Communications. The European Journal of Communication Research, 35(4), pp. 347-373.
JJ Igartua, IM Barrios & F Ortega (2012: “Analysis of the Image of Immigration in Prime Time Television Fiction” en Comunicación y Sociedad, 2, pp. 5-28.
JJ Igartua & L Cheng (2009): “Moderating effect of group cue while processing news on immigration. Is framing effect a heuristic process?” en Journal of Communication, 59(4), pp. 726-749.
JJ Igartua, L Cheng, F Moral, I Fernández, FJ Frutos, J Gómez-Isla & JA Otero (2008): “Encuadrar la inmigración en las noticias y sus efectos socio-cognitivos” en Palabra Clave, 11(1), pp. 87-107.
JJ Igartua, F Moral & I Fernández (2011): “Cognitive, attitudinal and emotional effects of the news frame and group cues on processing news about immigration” en Journal of Media Psychology, 23(4), pp. 174-185.
JJ Igartua, C Muñiz & L Cheng (2005): “La inmigración en la prensa española. Aportaciones empíricas y metodológicas desde la teoría del encuadre noticioso” en Migraciones, 17, pp. 143-181.
JJ Igartua, C Otero, JA Otero & M de la Fuente (2007): “El tratamiento informativo de la inmigración en los medios de comunicación españoles. Un análisis de contenido desde la Teoría del Framing” en Estudios sobre el Mensaje Periodístico, 13, pp. 91-110.
INE (2017a): Cifras de Población a 1 de enero de 2017 – Estadística de Migraciones 2016 a 29 de junio de 2017. Nota de prensa. Instituto Nacional de Estadística, Madrid, 2017. Disponible en: http://www.ine.es/prensa/cp_2017_p.pdf
INE (2017b): Estadística del Padrón de Españoles Residentes en el Extranjero a 1 de enero de 2017. Nota de prensa. Instituto Nacional de Estadística, Madrid, 2017. Disponible en: http://www.ine.es/prensa/pere_2017.pdf
J Koeman, A Peeters & L D´Haenes (2007): “Diversity Monitor 2005. Diversity as a quality aspect of television in the Netherlands” en Communications, 32, pp. 97-121.
K Krippendorff (1990): Metodología de análisis de contenido. Teoría y práctica. Barcelona: Paidós Comunicación.
C Lacalle (2004): “Comunicación y diversidad cultural” en Fórum Barcelona 2004. Disponible en: http://www.forumbcn2004.org/
C Lacalle (2008): El discurso televisivo sobre la inmigración. Ficción y construcción de identidad. Barcelona: Ediciones Omega.
J Lovejoy, BR Watson, S Lacy & D Riffe (2016): “Three decades of reliability in communication content analyses: Reporting of reliability statistics and coefficient levels in three top journals”. En Journalism & Mass Communication Quarterly, 93(4), pp. 1135-1159. doi: 10.1177/1077699016644558
M Marcos Ramos, J Igartua, F Frutos, I Barrios, F Ortega & V Piñeiro (2014): “La representación de los personajes inmigrantes en los programas de ficción” en Vivat Academia, 0(127), pp. 43-71. doi:10.15178/va.2014.127.43-71
D Mastro & E Behm-Morawitz (2005): “Latino representation on primetime television” en Journalism and Mass Communication Quarterly, 82(1), pp. 110-130.
D Mastro & BS Greenberg (2000): “The portrayal of racial minorities on prime time television” en Journal of Broadcasting and Electronic Media, 44(4), pp. 690-703.
F Müller (2009): “Entertainment anti-racism. Multicultural television drama, identification and perceptions of ethnic threat” en Communications. European Journal of Communication Research, 34(3), pp. 239-256.
C Muñiz & JJ Igartua (2004): “Información noticiosa sobre la inmigración en los medios de comunicación. Un análisis de la prensa y televisión españolas”. En J Latorre, A Vara & M Díaz (Eds.), Ecología de la televisión: tecnología, contenidos y desafíos empresariales (pp. 281-290). Pamplona: Eunate.
KA Neuendorf (2002): The content analysis guidebook. Thousand Oaks, CA: Sage.
KA Neuendorf, TD Gore, A Dalessanddro, P Janstova & S Snyder-Suhy (2010): “Shaken and Stirred: A Content Analysis of Women´s Portrayals in James Bond Films” en Sex Roles, 62, 747-761. doi: 10.1007/s11199-009-9644-2
WJ Potter & R Warren (1998): “Humor as camouflage of televised violence” en Journal of Communication, 48(2), pp. 40-57.
DS Reher, L Cortés, F González, M Requena, MI Sánchez, A Sanz & M Stanek (2008): “Informe Encuesta Nacional de Inmigrantes (ENI-2007)”. Documentos de trabajo, 2(08).
D Romer, KH Jamieson & NJ De Coteau (1998): “The treatment of persons of color in local television news. Ethnic blame discourse or realistic group conflict?”, en Communication Research, 25(3), pp. 286-305.
X Ruiz Collantes, J Ferrés, M Obradors, E Pujadas & O Pérez (2006): “La imagen pública de la inmigración en las series de televisión españolas” en Política y cultura, (26), pp. 93-108.
E Seiter (1986): “Stereotypes and the media: a re-evaluation”, en Journal of Communication, 36(4), pp. 14-26.
TA Van Dijk (1989): “Race, riots and the press. An analysis of editorials in the British press about the 1985 disorders” en Gazette, 43(3), pp. 229-253.
TA Van Dijk (1997): Racismo y análisis crítico de los medios. Paidós: Barcelona.
TA Van Dijk (2003): Ideología y discurso. Barcelona: Gedisa.
B Van Gorp (2005): “Where is the frame? Victims and intruders in the Belgian press coverage on the asylum issue”, en European Journal of Communication, 20(4), pp. 484-507.
How to cite this article in bibliographies / References
M Marcos Ramos, B González de Garay, C Portillo Delgado (2019): “The representation of immigration in contemporary Spanish prime time TV series”. Revista Latina de Comunicación Social, 74, pp. 285 to 307.
Article received on 3 December 2018. Accepted on 17 January.