The representation of phenotypic diversity in audiovisual advertising in Spain: professional perspectives and implications for the cultural dynamics of inclusion

Visual and audiovisual advertising is an influential social language that transforms representations of reality into cultural references, with direct implications for the perception of diversity. However, despite its impact on the fabric of society, few studies have tackled the representation of phenotypic variability in advertising by combining academic and professional perspectives on the phenomenon in Spain. Our study explores this matter through systematic data collection from 834 TV adverts, coding data on 359 on-screen characters, and in-depth interviews with 38 advertising professionals. The conclusions of the study show, among other findings: (a) the low level of phenotypic diversity in television advertising; (b) the under-representation of certain groups; and (c) the complexity of decision-making between clients and agencies, who have to reconcile risk aversion with the effectiveness required for any advertising campaign. This research can be of use for companies in the advertising sector (giving them improved knowledge of specific targets) and policymakers (by helping to promote better intercultural dynamics and the integration of minority groups).


Introduction
In the early twentieth century, R.H. Tawney claimed that no historian could write as if Marx had not existed. To adapt this idea to a different context, we could argue that no social scientist can write about contemporary culture and social organisation as if advertising did not exist. This is far from an exagger-ation, as today's societies are consumer entities which shape the culture to the extent that "the marketplace (and its main ideological tool, advertising) is the major structuring institution of contemporary consumer society" (Jhally, 1990: 506).
Hence, advertising has become an important social language (Caro Almela, 2007) that turns representations of reality into first-order cultural references whose capacity to penetrate the social imaginary enables them to influence popular culture on a global scale. Television is undoubtedly the main platform for shaping the collective imagination through popular culture (Anderson and Gabszewicz, 2006). Much of its potential lies in its ability to build, reproduce and transmit social models of behaviour and references which make an impact on audiences. They do so by generating desirable prototypes for the physical and attitudinal make-up of broad social sectors, according to their age, gender, ethnicity or class affiliation (Johnson, 2012).
However, there are few studies that interrogate the sociocultural characteristics of advertising, including the degree to which it can mimic or transform the real-world phenomena upon which it is built. While other areas, such as TV fiction, have been more widely studied in this sense (Marcos et al, 2014), television advertising has yet to be explored in terms of delineating the interrelationships between representation and reality.
According to Bermejo and Martínez (2013: 2), there is little research available within Europe on the impact made by the presence of immigrants in advertising or on the native population's perception of advertising featuring different ethnic groups. In Spain, it has not been studied at all. This is particularly relevant because, since 1998, Spain has become home to one of the largest immigrant populations in Europe. It has also experienced one of the largest increases in the percentage of non-nationals in the total population, which reached a peak of 11.4% in 2008 (Steingress, 2012). Such a major change is likely to encourage a corresponding transformation in sociocultural discourses and representations of difference, as well as in the perception of immigration and socially-articulated values and attitudes relating to it (Lorite, 2016). On this point, Verena Stolcke has noted that: This perception came first-hand by an intimate connection between the perception of physical traits and the ultimate cultural character they supposedly entail (the social dimension of phenotype). […] Phenotype tends now to be employed as a marker of immigrant origin. (Stolcke, 1995: 4-5;8) This paper therefore addresses the issue of phenotypic variation in the advertising field, and examines the extent and forms of such diversity as seen in Spanish television advertising. It should be noted that the expression 'phenotype' or 'phenotypic appearance' is often used in Spanish sociological and anthropological research to encompass cultural and physical traits (Cea and Valles, 2011), given that social organisation is based on the articulation between perceived morphology and ascribed social positioning (Sollors, 2008).
Subsequently, we will investigate whether the representations found in television advertising in Spain correspond to the country's socio-demographic reality, and we will reflect on the influence that this concurrence (or lack thereof) may have on specific intercultural dynamics, especially regarding inclusive social processes. In doing so, we also consider the views of advertising professionals on this matter.
Our study therefore helps to fill the gap identified by Bermejo and Martínez (2013) in research on the perception of otherness in advertising in Spain, and offers an approach to the issue of coding and analysing physical appearance and attributed ethnic cues in advertisements aired on real-time TV broadcasts in Spain.
To accomplish this, we carried out exhaustive data collection from the total range of available audiovisual data (a subsample of 834 adverts, drawn from a total of 4627 spots), based on 41 different variables. We then executed a thematic analysis of the subsample's selected adverts and the 359 characters represented in them. In parallel, we conducted in-depth interviews with 38 professionals from the worlds of advertising and cultural representation, which allowed us to address the issue from the perspective of advertising production routines as well.

The social relevance of TV advertising
It is difficult to exaggerate the importance of television advertising on an audience's collective imaginary of difference and cultural diversity. We will therefore focus on the well-documented phenomenon by which advertising campaigns identify specific targets, by configuring receivers and potential consumers according to specific criteria established by each advertiser for gender, age, geographical origin, ethnic affiliation or the cultural preferences of relatively well-defined social groups (Burton, McAlister and Hoyer, 2015).
Our goal here is to examine representations of diversity in advertising at the phenotypic level, which does not necessarily correspond to a diverse range of geographic or cultural origins but can sometimes be read in terms of ethnic or racial diversity (Shankar, 2015). Over the last 20 years, the increasing immigrant population in Spain has contributed to numerous urban and rural populations encountering a heterogeneous range of phenotypic traits on a daily basis. In this context, exploring the extent to which this new situation has transformed media representations of such diversity is especially relevant.
Consequently, this research follows the path of previous international studies dealing with the confluence of phenotype, identity and advertising. Beccaria et al. (2018) and Henry (2020) have shown how identities based on supposed phenotypical differences are highly malleable in terms of their performativity. After all, a phenotype results from cultural ideations about ethnicity and race embedded in oral, graphic or audiovisual narratives and dis-courses (Malta, Santos and Santos, 2021). Several studies have addressed the connections between phenotype, race, ethnicity and social consumption through the lens of advertising (see Banks, 2020, for a comprehensive literature review on this matter). International studies also point to a strong intersectional link between phenotypes and the representation of gender, social class and social or ethnic minorities at different levels by: (a) highlighting the underrepresentation of women and certain ethnic groups in advertising (Taylor et al., 2019); (b) promoting stereotyped images of sexualized female bodies, thus rendering visible the performative character of gendered identities (Beccaria et al., 2018); (c) underlining the agency of minoritized groups that purposely use their intersectional ethnic and gendered identities to develop specific messages concerning political or sociocultural issues (Santia and Bauer, 2022); and (d) evaluating those groups' reactions to specific advertising campaigns (El Hazzouri and Hamilton, 2019). Allan and Coltrane (1996: 185) argue that the most ubiquitous and stereotypical models reproduced in the collective imagination come from television advertisements, which are "preeminent among current purveyors of popular cultural imaginary". Advertisers have developed specific tools to measure and evaluate the effectiveness of an advertisement in relation to the interests of the brand that sponsors it (Burton, McAlister and Hoyer, 2015). These interests guide a campaign's development, but their effects go far beyond the strictly economic dimension of commercial marketing, aimed at influencing people's decisions when choosing that product. In this sense, the use and handling of stereotypes and cultural tropes in television commercials is far from incidental, and this may suggest the existence of certain cultural references pertaining to certain social groups, generational cohorts or genders.

From stereotypes to inclusive practices: advertising as a cultural operator
Consequently, following Napoli (1999), we should question whether certain social or demographic groups are accurately depicted. In this line of research, several studies stress the lack of correlation between reality and what is represented. Boulton (2016), for example, focuses on "race inequality inside advertising" within the framework of representations of African-Americans in advertising. In the United Kingdom, a Lloyds Banking Group report in 2016 indicated that only 19% of the people portrayed in advertising came from minority groups, and just 47% of people considered themselves adequately represented in advertising; additionally, 65% of consumers claimed they would feel more comfortable with a brand whose ads displayed greater social diversity (Kehoe, 2016). In India, Patil (2014) also notes the underrepresentation of people from "culturally diverse backgrounds". The situation is similar when looking at factors such as gender or age (Hetsroni, 2012).
In Spain, the situation is very similar when it comes to age, gender and minority groups. Even in the more specific field of how immigrants are portrayed in televised fiction, Marcos et al. (2014) conclude that migrant charac-ters are underrepresented compared to their actual position in Spanish society. Furthermore, our research shows very similar results in the field of phenotypic diversity on screen (Figure 1). However, some research has had an impact on the advertising industry's interest in demographic diversity, both regarding its purpose, its object of study and its staff, in addition to its multidisciplinary and multifunctional composition (Lynch and West, 2017). Further, several studies suggest that improvement in the media and advertising representation of marginal or minority groups can make the positive integration of these social segments more effective (Avraham and First, 2010).
Additionally, some empirical research has explored the impact that the ethnicity of advertising subjects has on consumers (Appiah, 2007), and how the degree of an advert's cultural embeddedness influences consumer response (Appiah and Liu, 2009). Advertisers have also been the subject of significant research, whether as employers of personnel with different ethnic ascriptions and cultural cues, or as executives of an industry with a substantial social impact due to the combination of product marketing, brand promotion and consumer appeal (Beard, 2017;Crockett, 2008).
It is for all these reasons that we have moved beyond a thematic analysis of the available audiovisual material, to incorporate the perspectives of the various professional agents involved in managing the representation of diversity in advertising.

The issue of self-regulation
Although it is not our purpose to analyse regulatory practices within the industry, the issues discussed above point to the need for further research on the self-regulation system in advertising (ASR or advertising self-regulation). While this matter has mainly been addressed in English-speaking countries (Dickinson-Delaporte et al., 2020), several previous studies in Spain have highlighted some significant weaknesses in this area (Perelló-Oliver and Muela-Molina, 2019; Muela-Molina, Romero-Rodríguez and García-Arranz, 2020). The consequences of such regulatory shortcomings are noticeable among certain age groups (Jiménez-Morales, Montaña and Vàzquez, 2019) and broad consumer targets (Perelló-Oliver and Muela-Molina, 2016), concerning not only images and visual representations but also radio messages broadcast by major national stations (Perelló-Oliver, Muela-Molina and Romero-Rodríguez, 2020).

Methods
Our research involves three successive levels of analysis: (1) First, a representative sample of TV adverts broadcast in a prime-time slot on five Spanish television channels was selected, four channels from the private sector (Antena 3, Cuatro, Telecinco and La Sexta) and the fifth a public channel broadcast in the autonomous community of Catalonia (TV3). We systematically recorded every advertising block broadcast on each of these channels over one week in March 2014, a period that did not coincide with any specific seasonal campaign. We ended up with 140 hours of recording, from which 4627 ads were extracted. We eliminated variants of the same ad of different lengths, as well as self-promotional spaces for the selected channels. The sample was refined to a subsample of 834 ads, which were classified according to the following main categories: group, brand, product, language, time, number of spots in the campaign, main language of the advertisement, presence of characters with/without indicators of diversity and a typology of diversity. To allow for potential subjective variation, we developed a codebook to be the basic resource for the coding team (four members). To do this, we used various ads from the sample in a control-group experiment with 12 volunteers of different ages, nationalities and academic training from within the university environment. We collected their labelling of skin, hair and eye colour, physique and relative height. A final discussion in focus groups served to reduce the list of labels to over-arching categories.
(2) The characters (359) who appeared in the selected ads were then codified and a second coding was subsequently carried out for those in whom distinguishable diversity indicators could be identified, such as phenotypical appearance and voice (128). These samples were coded according to 41 variables, including sex, physique, age group, the character's role, skin colour, eye colour, clothing, type of sounds and tone, using textual, graphic, visual and audiovisual languages as reference points.
(3) Finally, in-depth interviews were carried out with 38 professionals from the academic and audiovisual worlds, including creative artists, advertising entrepreneurs, casting agency directors, university professors and representatives of NGOs actively involved in combating racism and xenophobia in Spain.

Results
A strict comparison between our coding of television adverts and official statistical data is impossible, because the former is based on audiovisual representations and the latter on the officially recorded birthplace or nationality of registered residents. Nevertheless, we can observe that in 2014, near 10% of the total population were of non-Spanish nationality (4,177,294 versus 46,109,565). Table 1 shows the population's distribution by age and nationality, while Table 2 shows the age distribution in the country's real population and that observed in the sample of 834 adverts analysed.  The most frequent commercial categories in advertising broadcasts (Table 3) correspond to products that seem to be aimed precisely at the most well-represented targets, which appear to be the population segments with the highest consumer potential. As Table 4 shows, the young adult category is the most well-represented in both segments, and babies/toddlers the least. The few cases in which changes are observed in representation (e.g. 4% of seniors in ads without indicators of diversity versus 0% in ads with such features) can be explained by variance in the total number of on-screen characters (231 versus 128). Thus, we can conclude that a significantly higher number of characters in the sample does not imply a significant change in the correlations between age or role categories. Our analysis of the phenotypic traits of on-screen characters with notable features of diversity (128) will centre primarily on skin colour (Table 5), given its relevance as a social marker and the popular assumption that this is the focal point of difference, in 'racial' terms, despite the problems surrounding the use of this term, even in the field of biological sciences (Yudell et al., 2016). Representation also varies according to gender. Table 5 shows that male characters clearly outnumber female (83 versus 45), while the distribution of men's roles is slightly more balanced (nearly 41% have starring roles, while 31% have supporting roles and nearly 28% can be considered incidental characters). In contrast, women's roles are distributed as follows: roughly 44% starring, 42% supporting and 13% incidental. As for the combination of gender, skin colour and acting role, the highest percentages of women in starring roles are light brown 1 and most female supporting characters have reddish-brown skin. In contrast, most male characters have black skin in all three categories.
Regarding the physique of on-screen characters (Table 6), we followed Sheldon's conventional somatotype classification (1954), which includes three basic types: the ectomorph (thin with light musculature), the mesomorph (slender, with more-developed musculature) and the endomorph (tending to accumulate fat). Out of the 359 characters analysed, the predominant type is mesomorphic (63%), followed at a considerable distance by ectomorphic (19%) and endomorphic (6%). We were only able to specify a body type for 12% of the characters (generally because the composition of the scene did not allow the character's entire body to be seen).
1. In this article, 'Light brown' refers to the Spanish mulato: the term used by control-group participants for brown-skinned characters perceived as having a mixed-race background. We avoid 'mulatto' due to its connotations in the English language.

Discussion
As these results show, there is a clear disparity between the population segments represented in advertising and the age distribution in Spanish society. The most evident disparity is the considerable underrepresentation of the two ends of the scale (babies/toddlers and seniors) and, in contrast, an overrepresentation of children and young adults. If we examine the correlation of ages in the advertising sample more closely, we can see little difference between those adverts that show evident indicators of diversity and those that do not, particularly if we consider the role played in the advert by the character in question.
Only 10.1% of the coded characters have white skin, the diversity of these characters stemming from other aspects, such as spoken language (other than Spanish) or accent (e.g. Latin American). Dark skin predominates, accounting for 39% of on-screen characters, and almost a quarter have a reddish-brown skin tone. If we compare this data with the composition of the population of non-Spanish residents in mid-2014, we find that nationalities from other EU countries are predominant (37%), followed by 32% from South America, and Africa ranking third at around 17%, which suggests that there is an overrepresentation on television of individuals with darker skin among those determined to be non-Spanish. Additionally, in the first half of 2014, 37% of immigrants came from other EU countries, followed by 17% from South America and 15% from Africa. Therefore, it is reasonable to make an empirical generalisation that a significant proportion of the non-Spanish population represented in the statistical data during this period would appear to be white.
It could be argued that there is a certain correlation between the predominant skin colour in the emigrant sectors of the countries of origin and the representation of dark to reddish-brown tones in television advertising. However, this equivalence falls away if we set aside the subsample of adverts with features related to diversity, and base the comparison on the entire sample of adverts analysed, in which only 7.8% (65) of the 834 advertisements in the initial subsample include characters with visible indicators of diversity. This limited plurality does not correspond closely to what Spaniards experience in their daily sociocultural environment. No significant differences were observed in terms of gender, a slender physique being predominant in both women and men (69% versus 60%), and the endomorphic physique the least represented (5% versus 4%). If we take skin colour as a reference in the subsample of ads in which diversity-related features were present, we can see how these sub-sections of physique include, in general, little diversity (Table 6).

The professional viewpoint
According to several experts interviewed, the public representation of diversity by the audiovisual sector -including advertising -should be pluralistic but, above all, must be suitable for the particular context in which the brand and the advertised product are being presented.
With reference to the representation of diversity in television advertising, Larry, a professional who has had one of the longest and most consolidated advertising careers in Spain, argues that: "[…] brands in general are conservative. They aim to please as many people as possible and avoid upsetting everyone. And they therefore tend to focus on the most common denominator." An interviewee from a casting agency, however, is somewhat more nuanced. While she admits that comparatively few of the models they use are recognizably non-Spanish, she adds that: There has always been a limit on colour… when I make a casting call for ethnic groups, obviously, firstly, I don't put down [the term] race anymore, I use 'ethnic groups', and I don't use colour tone because it's something that gets passed about, and that email could end up in anyone's home. But I know […] and you don't need to talk to the directors, that from a very jetblack tone onwards, very dark […] it's completely clear that they're not going to select the black person! (Sylvia, director of a casting agency) There is some consensus among professionals in the advertising field that the underrepresentation of phenotypic diversity does not depend on substantive ideological issues (racism), but rather on eminently practical ones: aligning with the target audience's preferences and adjusting brand representations to them. However, the insistence on business conservatism, or the fear that certain on-screen characters may cause an undesired effect among particular target-audience segments, points to the relevance of cultural assumptions in this area. Additionally, the significance of ideology can be clearly seen in the recurrent commentary on stereotyped representations of diversity. Despite an awareness that "the cultural stereotype in general is like a false truth which is said so many times that in the end you believe that it's true" (Leonard, director of an advertising agency), interviewees acknowledged that the handling of diversity can easily end up as caricature.
Thus, the humorous or satirical hyperbole of certain stereotypes is generally considered a double-edged sword. Take, for example, the case of the character selected from a Kelisto advert (Figure 2). Some professionals stressed that the nature of the advertised product can make a significant difference (in this case, a comparison website): "It is clear that this is a much more direct kind of creativity […]" (Brian, casting director). Beyond this point of relative agreement, there is no consensus on whether this kind of treatment is degrading or if it is merely a strategy for gaining visibility, essentially seeking to attract the audience's attention.
Nor is media visibility blind to social class. In fact, Jeffrey (an expert in deontological codes) insisted that, despite the underrepresentation of certain social groups, it is entirely possible that this imbalance was not perceived as a discriminatory factor in the professional field. On this point, several advertisers insisted on the social perception of normality: that it takes some time for the diversity of appearances to be processed and become part of a habitual environment. Interviewees repeatedly insisted on the fact that advertising is, after all, a sales tool and that decisions made within the context of developing a sales strategy do not have to deliver a faithful representation of reality.
However, it is interesting to note that many of our respondents admitted that they could not support their explanations for the representation of different phenotypes in their productions with solid numbers. Similarly to Beattie, Khan and Philo's findings on the proportional representation of nonwhite ethnic minorities on screen (2013), there is seemingly no formal monitoring by the advertising industry of the number of ethnic minorities in TV adverts, despite the recommendations of regulatory bodies in Spain. Further research should address this topic in the future, including the role of advertising self-regulation (ASR) in Spain concerning the representation of phenotypic diversity in audiovisual advertising.

Conclusions and implications for practice
Phenotypic variability is only a superficial indicator of diversity, involving no more than the visibility of different physical features. However, from a broader social perspective, his variability is almost inexorably transformed into grouping mechanisms that entail social classification based on physical appearance, which ultimately has a great influence on interpersonal relationships in any given society, as well as on the collective imaginary concerning difference and inequality.
Given the social significance of perceived differences, we consider it crucial to learn more about the specific forms that this variety is given in television advertising, and the effects this can have on wider social relationships. As Patil (2014) has suggested, the advertising medium generates powerful audiovisual narratives which, in the specific field of phenotypic diversity, can have a considerable impact on the social perception of ethnic or 'racial' inclusion.
Previous research has demonstrated the need to explore the representation of individuals with different physical appearances in advertising in greater depth. Our research takes a step further in interrogating the perception of otherness in this field through an approach to the problem that involves the coding and analysis of phenotypes from a representative sample of TV advertisements broadcast in a prime-time slot. Taking phenotype as a cultural marker (Stolcke, 1995) helps us understand the connections between physical appearance, social organisation and ethnic cues. Our study helps to fill the gap concerning the presence of immigrants in advertising (Bermejo and Martínez, 2013) by providing an analysis of advertisements that incorporates the advertisers' points of view.
Broadly speaking, audiovisual advertising ignores the real phenotypic diversity of current Spanish society, which, from a professional perspective, raises two points for consideration. On the one hand, it demonstrates the need to expand the visible range of people with physical features that differ from those of the majority currently represented in advertising; on the other hand, it underlines the risk posed to brands by choosing more realistic representations of diversity. In this sense, a feedback mechanism of risk aversion between clients and agencies may be generated: the former prefers a sure-fire sales strategy and minimising the uncertainty that alternative representations bring, while the latter may see their creative freedom constrained by the margin of risk the brand is willing to take on. However, we should make it clear that clients are not inherently resistant to agencies' creative proposals. Rather, they are forced to consider other factors in the equation, such as the effect the chosen advertising strategy may have on the specific target audience towards which the advert is directed.
This said, we are not dealing with discriminatory situations, as Jeffrey stated, since neither brands nor agencies are likely to perceive underrepresentation in that way. Instead, this is a case of advertising reflecting the way in which diversity is socially perceived and associated, in an a priori way, with political and integrational issues related to migration and social groups' ethnic (or other forms of) plurality.
From the point of view of advertising creatives, time is a key factor in the effort to avoid underrepresentation and asymmetries: for balanced representations of diversity to be achieved, they must first be perceived as normal in everyday social contexts. And that sense of normality is not achieved overnight; nor is it likely to be attained within a few years.
Therefore, this research could be useful for companies in the advertising sector and for policymakers. For the former, it is an opportunity to make use of improved knowledge of specific targets, so as to differentiate themselves in the market and feel trusted by their potential clients (Kehoe, 2016). For the latter, it could contribute to better intercultural dynamics and the improved integration of minority groups, to the extent that: "[…] given the reach and cultural influence of advertising, advertising which better reflects the diversity of the population may help to support the integration of minorities into the multicultural matrix" (Ambwani, Heslop and Dyke, 2011: 345).