Introduction

The increased visibility of hate speech and other forms of offensive communication online is proving to be a challenge for connected societies. The fact that the internet is being used to spread messages of hate and violence has become a social concern due to the capacity attributed to it to disperse, generate, and enhance the effects of these messages (Ardèvol, 2016; Gualda, Borrero, & Carpio, 2015; Miró, 2016). The concern has increased after the appearance of studies that have confirmed the relationship between what happens in the online and offline worlds (Olmos, Rubio, Lastres, & Martín, 2020). In fact, some research proposes to predict social behavior based on the messages published on social networks (Congosto & Aragón, 2012; Congosto, 2014), such as hate messages.

According to the United Nations, hate speech is “any form of verbal, written, or behavioral communication that attacks or uses pejorative or discriminatory language concerning a person or a group based on who they are or, in other words, because of their religion, ethnic origin, nationality, race, color, descent, gender, or another identity factor” (UN, 2019). The events that occurred throughout the summer of 2021 in Spain, such as verbal and physical attacks (even resulting in death) against members of the LGTBIQ+ community, have put homophobia at the center of hate speech.

In 2012, the European Parliament defined the term "homophobia" as "the irrational fear and aversion to male and female homosexuality, which is based on prejudice and is comparable to racism, xenophobia, anti-Semitism, and sexism, and which is manifested in the public and private spheres in different ways, such as incitement to hatred and discrimination, mockery, and verbal, psychological, and physical violence, as well as persecution and murder, discrimination in violation of the principle of equality, and unjustified and unreasonable limitations of rights, often hidden behind justifications of public order, religious freedom, and the right to conscientious objection” (Resolution (2012/2657(RSP)) of the European Parliament, of May 24, 2012). These demonstrations are carried out today through social networks such as Instagram.

In its terms and conditions, the social network ensures that it removes “any content that includes credible threats or language that incites hatred, content directed at individuals to humiliate or embarrass them, personal information used to blackmail or harass someone, and repeated unwanted messages” (Instagram, 2021). For the platform, "it is unacceptable to encourage the use of violence or attack someone for reasons of race, ethnicity, national origin, sex, sexual identity, sexual orientation, religious beliefs, disability, or illness" (Instagram, 2021). Despite the limitations in the conditions of use of this social network, the mechanism is not well explained to users (Heins, 2014), or its implementation is carried out in a somewhat dysfunctional way (Cabo & García, 2016).

This fact makes the collaboration of the rest of those involved in the communicative process necessary. One of them is the media. They are not only in charge of giving visibility to the reality that the LGTBIQ+ community sometimes suffers but also of monitoring their contents and what is said in them both on their web pages and their social network profiles. Although they cannot be responsible for the opinions expressed by readers, they have committed to trying to prevent hate speech from spreading. Therefore, offensive, insulting, hurtful, derogatory, or obscene language directed at a person or group on social networks has fueled the debate on content moderation. In conflict situations, such as the dissemination of hate speech, the role of the media is key and will influence how it ends (Aguiló, 2020).

As in most spheres of society, social networks have had a strong impact on journalism (Viner, 2016). AsPérez-Soler (2017) points out, these platforms place the user at the center of the business strategy, encouraging participation. This supposes a change of mentality above a technological change. This change supposes, among other aspects, that the events most likely to be commented on and shared on social networks are those that journalists will choose to turn into news (Pérez-Soler, 2017) because the disclosure of the content is one of the strategic factors of the media today (González, 2020).

Newspapers, televisions, and radios began to create profiles on social networks to expand and improve the distribution of their content. Beyond reaching the audience, they also want the audience to interact with the content by sharing it and/or through comments (Arréguez & Merlo, 2020). One of the most used social networks for this purpose is Instagram. The media are taking advantage of its ability to create content and interact with the public. However, although there is a great deal of research on Facebook or Twitter, studies on the use of Instagram by the media are scarce (Lirola et al., 2015).

Objectives

The objective of this work is to describe and quantify the presence of hate speech directed at the LGTBIQ+ community in the comments on Instagram posts of the main Spanish newspapers with national circulation.

It is based on the hypothesis that, although the publications about the group arouse negative comments that can be recognized as hate speech in its different manifestations, the visual nature of Instagram, together with the filters of both the social network itself and the specific to each media outlet, prevent hate speech against the community from being the majority discourse.

Methodology

Descriptive research allows a critical review and in-depth analysis of the research line to review the conditions that caused its appearance, as well as the most relevant results achieved by the various studies on the subject, in addition to pointing out the existing conceptual problems and methodological limitations (Tinto, 2013). This is the case of this work, which, based on empirical content analysis, aims to extract a set of variables with which to establish relationships and interpretations of the reality under study.

The present research worked on a case study analyzing all the comments made in the publications of the Instagram accounts belonging to the five main Spanish newspapers with national circulation (El País, El Mundo, ABC, La Vanguardia, and 20Minutos) related to what is known as "Pride Week". The study period comprised from June 21st to July 4th, 2021, two weeks between which the celebrations are framed. The final sample comprised the analysis of the comments generated (N=6,013) from a total of N=20 publications, which became the analysis units of the study (Table 1). A quantitative content analysis was chosen, considered as a technique that allows objective and systematic knowledge of the content of the analyzed messages (Igartua, 2006; Piñuel, 2002) from the collection of information after the automatic download of the comments using the Export Comments tool. Downloaded comments were supplemented manually because the software does not download the contents of private accounts (10%-30%).

Table 1: Newspapers, Instagram accounts, and the number of posts analyzed

Source: own elaboration

The coding protocol of each unit of analysis included the registration, in each post published by each newspaper, of the number and type of comments generated. For the classification of comments, a classificatory variable proposed byLosada-Díaz, Zamora-Medina, and Martinez-Martínez (2021) aimed to measure "the numerical presence, among all the comments generated in each publication, of six parameters that include all the options on which hate speech is based: criticism, insults, contempt, threats, mockery, and others" (p.200) was used. To these, the category “in favor” has been added. Table 2 explains how each of these parameters was interpreted in the coding process:

Table 2: Analysis units classification parameters

In this way, the study combines quantitative methods, focusing on counting the values established for each variable, and qualitative methods aimed at a deeper understanding of the comments for a critical and constructive perspective. To minimize subjective bias, the classification was made twice by the author and a third time by an external party, with the three reviews coinciding in 94.8% of the cases. Subsequently, the SPSS version 20 computer program was used to proceed with the statistical treatment of the data. The descriptive analysis included carrying out the frequencies of the variables included in the study, as well as the elaboration of contingency tables.

Results

The total number of comments analyzed from the 20 posts studied (see Annexes) amounts to 6,013 comments. El País and La Vanguardia are the newspapers that register the greatest interaction by their users if we compare the posts in which the different newspapers coincide (Table 3). P2 of El País is the same as P1 of ABC and P2 of La Vanguardia and P1 of El Mundo. However, the number of comments is much higher in the case of the Prisa group newspaper. In the case of the Catalan newspaper, its P3 is the same as P3 of ABC and P1 of El País, which it exceeds in comments. Although ABC is one of the newspapers that published the most content related to Pride Week, most of its publications do not reach 100 comments. In the case of 20Minutos, it is the one that published the least content related to the studied subject.

Table 3: number of comments analyzed per newspaper

Of the total number of comments, 41.8% were registered as “Others”. This high percentage is because, on many occasions, users start discussions that end up being unrelated to the topic of the publication. Furthermore, as can be seen in the examples in Table 2, the use of emoticons is also frequent; Sometimes, it is impossible to elucidate the meaning that the user wanted to give, for example, to a face with closed eyes and sticking out the tongue. Finally, regarding this category, it is worth mentioning that those publications with the highest number of comments classified as “Others” are those that generated discussions among some users. As already mentioned, these discussions usually end up addressing issues totally unrelated to the subject being studied and the arguments are fundamentally based on personal disqualification.

Leaving aside this category, the majority of the comments are in favor of the LGTBIQ+ community (32.6%). "Criticism" represents 16%, leaving far behind " Mockery" (4.9%), "Contempt" (3.1%), "Insults" (1.5%), and "Threats” (0.1%). The small percentage of comments that can directly be considered hate speech (mockery, contempt, insults, and threats) reveals that, contrary to what might be expected, the presence of hate speech on Instagram is not very significant, unlike what can happen in other social networks like Twitter.

The publication that received the highest number of comments was P3 of El Mundo (Image 1). It is a video in which members of the Civil Guard appear acknowledging that they are homosexual or transsexual. It was the only newspaper of those analyzed that echoed the video broadcast by the meritorious on its profile on the TikTok social network. Of the 1,357 comments, 470 (34.6%) were classified as “Others”, 453 (33.4%) were “In Favor”, and 297 (21.9%) were “Criticism”. As for hate messages, it received 67 (4.9%) "Mockery", 59 (4.3%) "Contempt" messages, eight (0.6%) "Insults", and it is the publication in which a greater number of threats, three (0.2%), were detected.

Source: @elmundo_es on Instagram

In contrast, the publication that received the fewest comments was ABC's P3. Of the 37 comments, 11 (29.7%) are "Criticism", seven (18.9%) are "In favor", seven (18.9%) also deal with "Others", five (13.5%) are “Insults”, the same number of “Mockery”, and two (5.4%) are “Contempt” comments. The content of this publication, the jump of a spontaneous fan with the flag of the community to the soccer field in a match of the European Championship that faced Germany vs. Hungary, is the same as P1 of El País and P3 of La Vanguardia (Image 2). In the case of the publication of El País, and contrary to what happens with that of ABC, it is the content that received the most comments. Of the 628 comments, 335 (53.3%) were “In favor”, 86 (13.7%) were “Criticism”, 22 (3.5%) “Mocking”, 19 (3%) “Contempt” comments, seven (1.1%) “Insults”, and one (0.1%) “Threat”. In the publication of La Vanguardia, also the one that received the most comments, the majority of the comments are about “Other” (52.5%) topics, followed by the comments “In favor” (22.8%), “Criticism” (12%), “Mockery” (8.7%), “Contempt” (2.5%), “Insults” (1%), and it is the only publication of this newspaper in which “Threats” were detected (0.3%).

Source: @lavanguardia on Instagram

The other content shared by several newspapers is related to the lighting of the Allianz Arena in Munich with the colors of the LGTBIQ+ flag during the same match of the content analyzed above. These are ABC's P1, El Mundo's P1, El País's P2 (Image 3), and La Vanguardia's P2. The media outlet in which it generated the fewest comments was ABC (89 comments), the majority of them (59.5%) being “Criticism” and only 15.7% “In favor”. Where more comments were produced (533) was in El País. 45.3% of them were classified as "Others", 39.8% as "In Favor", 11.8% as "Criticisms", 1.3% as "Contempt" and "Mockery", and 0. 4% as “Insults”. A similar number of comments were received in El Mundo and La Vanguardia (256 and 239 respectively). In the first case, the majority were classified as "Other" (49.6%) and then "In favor" (29.7%), while in the second the majority were comments "In favor" (50.6%).

Source: @el_pais on Instagram

The category “Other” is the majority in the accounts of El Mundo, La Vanguardia, and ABC (Chart 1). In the case of El Mundo, it is the predominant category in all its publications. In general terms, this is because its users hold debates that, as indicated before, move away from the main theme of the publication. The case of La Vanguardia is different since the total of “Other” is due to a single publication (P3, analyzed above); In the other four publications, the category with the highest number of comments is “In favor” (33%). They follow the same line as 20minutos (30%) and El País (42.3%), newspapers whose majority of comments fall into this category. In ABC, only in P5 (Image 4) is “Other” the majority. It is the publication of the newspaper that has the highest number of comments (254), being more “Insults” than “Mockery” or “Contempt”, although practically all of them are addressed to the Minister of Equality and not to the community; It is also the only case in this newspaper in which a threat was found, also directed at Montero. In the rest of the publications, the majority is “Criticism” (28.1%).

Source: @abc_diario on Instagram

Source: own elaboration

When specifying how hate materializes in negative comments, in each case a fairly similar pattern is observed in all the newspapers, since a clear predominance of negative comments centered on the "Criticism" typology is detected, followed by the “Mockery” option. Although, as mentioned before, the presence of comments that can be considered hate speech in the global count of comments is not very evident, the fact that “Insults” and “Contempt” are recorded in all the accounts analyzed, shows that Instagram users also use this social network destructively. However, the most worrying typology, “Threats”, is almost non-existent in all the publications.

Discussion and conclusions

There is no doubt that the internet, and social networks, in particular, have given greater diffusion to the hate speech present in society. The different characteristics of these platforms, such as interaction, instantaneity, and anonymity, allow hate speech to be carried out more impulsively. In fact, since the reform of the Criminal Code carried out in 2015, the penalties for hate crimes are greater when they are carried out through media that allows the message to reach a large number of people. Likewise, and also as of that year, the Ministry of the Interior registers the presence of hate speech in Spain, taking into account any event in which the medium used for its commission is the internet, social media, streaming pages, shared file networks, direct download pages, blogs, emails, etc., and is related to any of the areas described for hate crimes (Carratalá & Herrero-Jiménez, 2019).

Even though there are automatic language detection tools for the study of hate speech, this work has highlighted the difficulty of classifying content, especially if it is done exclusively through software. As mentioned and exemplified, there are a large number of comments that use emoticons that even for a person are difficult to interpret. Added to this use of language are the ironic intention with which some users write, so a comment that is actually in favor could be classified as against and vice versa. With this work, we want to record the need to continue working on methodologies that use both technological advances in linguistic analysis and human knowledge.

The objective of this research was to describe and quantify the presence of hate speech directed at the LGTBIQ+ community in the comments to the Instagram posts of the main Spanish newspapers with national circulation. After analyzing 6,013 comments, it is concluded that hate speech, although present in publications related to the group, appears in its least harmful form on the scale of possible manifestations of hate speech, where criticism is the protagonist and threats have a residual presence in all the studied cases. In fact, the general tone is one of support for the community because, leaving aside the category "Other", the comments "In favor" are the majority. The only exception is the ABC newspaper. Taking into account the conservative editorial line of this newspaper, the result was predictable and coincides with the results of the research byVega (2019) on Facebook.

There may be several reasons why extreme forms of hate speech have not been found in the analyzed sample. The first of these can be found in Spanish society itself, in which members of the LGTBIQ+ collective have recognized rights and are widely accepted. Another reason is found in the limits imposed by Instagram and discussed in the first section of this work. AsSorigó (2020) comments, unlike other social networks such as Twitter, Instagram has a series of filters that prevent the publication of certain types of messages, so homophobia is more present on the microblogging network. It can also happen, as in M Martínez and Mayagoitia (2021), that hate messages are sent privately.

Finally, there is the mediation of the social network managers of the analyzed newspapers, who can moderate the comments by adding filters to those already established by default by the platform or even delete those that they consider offensive. They were contacted on four different occasions (July, August, September, and October) to ask about this management, and in no case did they respond. Even so, and as users have stated in some comments, these managers do filter content, being one more reason why the most aggressive hate speech does not appear published. This confirms the initial hypothesis that proposed that, although the publications about the community arouse negative comments that can be recognized as hate speech in its different manifestations, the visual nature of Instagram, together with the filters of both the network itself like those specific to each media outlet, prevent hate speech against the community from being the majority discourse.

This exploratory work on the presence of hate speech in social networks in which young people have a greater presence could be completed in the future with the use of other qualitative methodologies such as the discussion group. Likewise, a comparison of the comments and reactions to the same publications studied here, but published on the Twitter accounts of the newspapers, could be carried out; even compare them with the comments left by readers on the media website. Furthermore, it would be essential and enriching to have in-depth interviews with the managers of the newspapers' social networks, as well as interviews with members of the LGTBIQ+ community to learn how these messages affect the victims.

Annexes