The so-called ICTs, Information and Communication Technologies, are part of the evolutionary framework of mass media and introduce qualitative changes in the social forms of interaction between people. This context entails a set of social, economic, political, and, above all, communicative manifestations that significantly impact how people interact with each other. Since the creation of the internet, and with the spread of the world wide web, interpersonal relationships have been digitised to the point that any citizen can express their opinion and actively organise around common interests thanks to the digital society (Porcelli, 2020) and the knowledge society (Silva et al., 2012). The society of the twenty-first century is a network society, that is, a society built around personal and corporate networks operated by digital networks that communicate through the internet. Moreover, since networks are global and know no boundaries, the network society is a society of global networks (Castells, 2012). The expansion of social networks worldwide is evident, as shown in different studies (Vepsäläinen et al., 2017; IAB Spain, 2021). Despite other growing networks, Facebook is still the social network with the most users in the world (Fernández, 2021).

To this, we must add that, in the last decade, social networks have been gaining importance as a communication tool used by companies (Marín Dueñas and Simancas, 2019) or by political parties (Pérez-Curiel et al., 2021). All this establishes a base where citizens can interact (Cabezuelo and Martínez, 2010), not only among themselves, but also through other media (González-Oñate et al., 2020) and where citizens and voters can be persuaded through political campaigns with propagandistic overtones (Gunn and Skogerbø, 2013; Bellido-Pérez and Donstrup, 2020).

Political communication has been transformed thanks to this multiplication and dissemination of messages through social networks, with efficiency and economic profitability rates that were unthinkable until now (Elías Zambrano et al., 2019). In this sense, a space for information, communication and propaganda opens up thanks to which political parties gain access to an online audience by invading their personal spaces (Bernhardt et al., 2012). Parties can now engage in dialogue with users (and voters) and address issues of interest since they can express their opinions (Levonian, 2016).

Currently, there are several studies on the use of Facebook in the political context, a large number of them for predictive purposes, such as Vepsäläinen et al. (2017), who examine the extent to which Facebook likes can be used to predict election results; and the research of David et al. (2016), who argue that it is desirable to identify the targeting of the potential electorate, as personalised messages are much more effective. Engagement is another issue that has been analysed on numerous occasions (Garz et al., 2020; Sally and Wickramasinghe, 2020). Another topic that has been of particular concern to researchers in this context has been the persuasion technics, the emotional tone and content of Facebook conversations, with a negative emotional connotation being the general tone (Bronstein, 2013; Abejón and Mayoral, 2017; Falck et al., 2020; Park et al., 2021; Widmann, 2021). Other authors approach this topic from the perspective of social network analysis, such as the research by Bíró et al. (2016). They explore whether the use of likes (on pages) and posting activity are relevant factors in network growth. However, the results show that parties form closed communities in the networks and have their users.

In Spain, some studies have investigated how users of the leading Spanish parties interact on Facebook. The results point to the existence of an interaction between users characterised by disagreement and more negative and offensive language, especially on the Facebook pages of the most ideologically polarised parties (Fenoll and Cano-Orón, 2017). Also, there are studies about misinformation (Cano-Orón et al., 2021), citizens' comments on Facebook about political leaders' messages (Zurutuza-Muñoz and Lilleker, 2018), or parties' engagement with citizens on Facebook (Ballesteros and Garrido, 2018).

In short, this proliferation in social media has led to a transformation in political communication in which political parties have seen these tools as essential resources that allow them to access an online audience. In this way, they open up space where users can express themselves on issues of topical interest (Levonian, 2016). Current issues include education, employment, welfare, equality, happiness, Etc., which have been gaining prominence in recent decades, especially since 2015, when the international authorities, which are part of the United Nations, met at the Sustainable Development Summit and drew up the 2030 Agenda with the Sustainable Development Goals (United Nations, 2021). Despite the importance of these social issues, insufficient evidence was found in the sources consulted on how political parties deal with the sustainable development goals in their Facebook communications, which will be addressed in the following section.

Digital social networks (DSN), including Facebook, use the internet as a communication channel for socialising and sharing ideas. In these writings, the textual content of the DSN can be enriched with multimedia elements. Together with the interaction, immediacy, and public nature of the conversations, this characteristic makes these tools an excellent potential for the digital public sphere (Milliken and O'Donell, 2008). Moreover, chats generate relevant content for users (Bruns, 2006; Leong and Ho, 2020). As mentioned above, Kaplan and Haenlein (2010) define DSRs as.

If we consider the paragraphs above, it should be noted that social media has become a popular channel for political issues and opinions to exposure (Koiranen et al., 2020; Román-San-Miguel et al., 2021). According to Bond et al. (2012), social media can reflect citizens' political behaviour, which significantly influences their political decision-making. For this reason, politicians are increasingly using social media in their campaigns, as it allows political parties and organisations to connect with citizens in the age of the digital society (Strandberg, 2013).

Citizens' behaviour is a variable that can be influenced by the application of social marketing techniques in the political context. In this sense, Andreasen (1995) stated that social marketing aims to modify behaviours in search of a higher good (individual and social), being this change of attitude or behaviour, voluntarily, which has characterised this discipline in recent years. However, the paradigm of social marketing has evolved since the appearance and use of social networks as communication tools. In this sense, Hestres (2014) already suggested that, in a digital and social media context, efforts should be focused on getting the most significant number of users to talk about social issues, which would make it possible to offer engaging content to the public and, therefore, greater dissemination and visibility of the information. It would help leaders and political parties transform society (Bernhardt et al., 2012; Dooley et al., 2012).

However, despite the importance of issues affecting society such as poverty, health, equal opportunities, Etc. that are part of the SDGs, not much literature has been found that addresses these issues. Sustainable development meets the needs of the present without compromising the ability of future generations to meet their own needs (El Serafi, 1994). In this sense, at the United Nations Rio de Janeiro Conference on Sustainable Development (RIO+20) in 2012, 17 Sustainable Development Goals were developed to end poverty, protect the planet, and ensure peace and prosperity for people (United Nations, 2021). Some examples of research on sustainability focus on methodologies in different contexts such as education, such as the study by Lozano-Díaz and Fernández-Prados (2020). They proposed a research project on global and social citizenship under cyberactivism in university students of Education. This research showed a low level of active, critical and political engagement on the Internet about the environment, human rights or social justice in the online environment that would involve students' personal involvement. It was also observed that the learning methodology used, based on a workshop to train students in digital engagement issues in the field of the SDGs, resulted in a significant improvement in students' digital citizenship competencies. Another proposal focusing on using different methodologies for SDG analysis has been made by Ospina-Forero et al. (2020), who conducted a comparative study on the most suitable estimation methods for SDG networks, building a database with 87 development indicators in four countries over 20 years. The relationships between these development indicators (covering 16 of the 17 SDGs) during 1995–2004 for each country independently were analysed in this research. This information can be used for policy decisions.

Sustainability is also a goal that policymakers little discuss. An example is found in the issue related to climate change, where, while politicians may emphasise the urgency of addressing this issue, they often propose policies that weakly address the fundamental socio-economic roots of the problem (Chaffin et al., 2016). In this sense, numerous examples of policies are motivated by sustainability objectives, such as open space preservation strategies or people transit-oriented development (Krueger and Buckingham, 2012). These examples show that sustainability can remain a volatile notion, often ambiguously defined according to the political tactics of the parties involved in the decision-making process. It is reflected in the complete lack of public debate on sustainability-oriented forms towards objectives other than growth and progress (Latouche, 2009).

As a consequence, political debate remains limited to technicalities rather than political motivation. In other words, decision-making is increasingly seen as a matter of expertise rather than a political position. Political parties deliberately delegate issues to technicians to guide the process. It appears to be an intentional postponement of accountability, occurring as a defensive partisan strategy to cope with the risk of losing electoral power on issues of social relevance to the city (Bossuyt and Savini, 2018).

In short, according to most studies that have explored the relationship between the content of parties' and candidates' posts and the level of engagement they have generated on Facebook, they found that posts that talked about acknowledging their supporters or colleagues as well as those criticising their opponents perform better in terms of likes and comments than other types of content (i.e., information, mobilisation or individual posts) (Larsson and Kalsnes, 2014). Furthermore, regardless of ideology, political parties have made increasing rhetorical use of sustainable development over time, which, on the other hand, does not serve as a dimension for competition between political parties (Fleig and Tosun, 2017).

In this context, this study aims to empirically examine the publications of the leading Spanish political parties on Facebook from a social marketing perspective. In this way, it will be possible to verify how issues related to the Sustainable Development Goals (SDGs) are addressed. On the one hand, it is required for achieving this goal the elaboration of the communication profiles of the main political parties presented in the Spanish General Elections during the period between 2015 and 2019. On the other hand, to analyse whether there are issues that better discriminate or distinguish one political party from another. For this purpose, we have used a methodology based on data mining, content analysis from a quantitative and qualitative approach, and simple correspondence analysis. The findings of this work may have important practical implications for the future implementation of the happiness management philosophy by political parties. A culture that contributes to building the loyalty of political party voters through the intangible vector of happiness, together with assertive communication and supported by a political programme based on the fulfilment of the SDGs.

The specific characteristics of the data coming from social media (large volume of unstructured information) make it necessary to use a specific analysis methodology. This methodology must focus on extracting data and preparing and analysing to obtain valuable and relevant information that responds to the objectives. This procedure is known as Knowledge Discovery in Databases (KDD), defined by Maimon and Rokach (2010) as an organised process of valid, novel, helpful identification that generates understandable patterns from large and complex datasets. This process can be summarised in the phases specified in Figure 1.

Concerning the phases of the KDD process, it should be noted that data preparation (selection, exploration, cleaning and transformation) consists of eliminating unwanted messages, such as duplicate messages, and performing text cleaning by removing extraneous symbols, links, numbers, and numbers punctuation marks. Data mining allows us to discover new and potentially useful information from large amounts of data. Two types of techniques can be used; some are verification techniques, in which the system is limited to testing hypotheses provided by the user (for example, checking whether the use of videos influences the impact of social networks). Others are discovery techniques, in which potentially interesting patterns have to be found automatically, including all prediction techniques, for example, anticipating what kind of messages on social networks will have a more significant impact.

This study will be carried out using a descriptive approach to describe the particularities of the relevant groups and calculate the percentage of units showing a particular characteristic or determine the degree of association between variables (Malhotra, 2004). To this end, a content analysis was carried out (from both a qualitative and quantitative perspective) of the messages published by the main political parties in Spain: Partido Socialista Obrero Español (PSOE), Partido Popular (PP), Ciudadanos (Cs), Unidas Podemos (UP) and Vox. These political formations were chosen because they obtained the highest number of votes in the 2019 general elections, as specified in Table 1.

Posts published on Facebook by the selected parties during the selected periods were captured in an automated way using the Fanpage Karma tool. In total, 9,115 posts were collected. Of these posts, those mentioning social issues were analysed in detail, as explained in the section on content analysis, so that 2,965 posts were examined (Table 2), which were manually coded.

In addition, verification techniques have been used to test the following hypothesis tested:

To confirm the existence of a relationship between social issues and political parties, a simple correspondence analysis (SCA) was carried out.

Content analysis is a technique that seeks to discover the significance of a message, whether it is a speech, a life story, a magazine article, Etc. More specifically, it is a method that consists of classifying and codifying the various elements of a message into categories in order to find patterns that give a specific meaning to the communication that takes place in them (Gómez, 2012). Krippendorff (1990) defined it as “A research technique designed to formulate, from certain data, reproducible and valid inferences that can be applied to their context” (p. 28).

The content analysis aims to achieve a quantitative information analysis and can study phenomena and data from a qualitative point of view. The first aims to measure the data and establish the frequency of occurrence of the elements considered units of information. The second (qualitative) allows us to verify themes, words or concepts in content. This research will take these two perspectives of content analysis to evaluate the effectiveness of the conversations published on the Facebook pages of the selected political parties.

For the procedure of this technique, the approach of Bardin (1983) will be used, which presents the following three stages: Pre-analysis, exploration of the material and processing of the results. Following these steps, we have identified as units of analysis the messages published on the Facebook pages of the political parties, PSOE, PP, Cs, UP and VOX, during the general elections of 2019, 2016 and 2015. The objective to be achieved with this technique is to elaborate the communication profiles of the political parties according to the social issues related to SDGs addressed by the publications. The following variables and categories have been identified to enable the coding of the messages:

In order to determine the validity of these categories and decide whether they are the ones that will ultimately be used in the research, two experts in the field were consulted and provided with a sample of 100 publications chosen at random to be classified into the proposed categories. In this way, the concordance between the classifications of the two experts can be checked. Cohen's kappa coefficient was used to measure this, which yields a significance level of < 0.05. It indicates that a relationship is established between the two specialists. It indicates that the relationship between the two rankings is established, and the null hypothesis that there is no concordance between the results of the two experts is rejected. Also, the value of this parameter is 0.709, which indicates that the agreement between the two coders is satisfactory according to the scale of the researcher mentioned above (Warrens, 2014).

Once the coding process has been carried out, the next step is to deepen the knowledge of the messages. This process aims to extract the crucial questions on each parameter considered that support decision-making concerning social communication actions. For this purpose, the SPSS 25 plus programme has been used to carry out two types of analysis: a descriptive analysis and a correlational analysis. The latter aims to measure the degree of relationship between the parameters specified in the hypothesis to test.

In this study, SCA will confirm whether there is a relationship between political parties and the social issues addressed in Facebook posts. SCA is an interdependence technique that describes a data set and provides a perceptual map based on the association between objects and a set of descriptive characteristics or attributes (Hair, 1995). This multivariate technique first tests the association between variables using a chi-square test and then allows the latent relationship between the two variables to be identified graphically. The method starts with the construction of a contingency table of the variables analysed. Based on the information in this table, a two-dimensional perceptual map is constructed in which dots represent the categories of both variables. The closeness between the points associated with categories of different variables indicates the variables' level of dependence or association. In this sense, the various categories of the variables are represented closer or further away from the dimensions depending on their degree of similarity or differences (Zapillado, 2017). In other words, categories close to the origin of the two-dimensional space will be the categories of the variables that least discriminate each of the dimensions of the solution, and in the categories further away from the origin, the discrimination is more significant. Therefore, it is an ideal method for finding the underlying structure in sets of qualitative variables, with which we can work simultaneously and offer us the underlying behavioural patterns (Camacho, 2010).

Figure 2 below presents the general results of the descriptive analysis corresponding to the communication profiles of the political parties understudy during the 2015, 2016 and 2019 general elections. These results correspond to the analysis of the 2,965 messages selected for addressing social issues related to the SDGs.

Figure 2 shows that the most active parties in the period under consideration were PSOE, PP and VOX. Two interesting facts can be observed in this regard. The first is that Cs was the party with the second-highest average number of likes per message, despite being one of the parties that posted the fewest messages. Secondly, VOX was the party with the highest average number of likes and comments per message.

If we consider the activity corresponding to the different elections, Table 5 shows a disparity between the activity of the parties in the different elections and the public's reaction. More messages have been published in the April 2019 elections (E2), and the most active parties were Ciudadanos, PSOE and UP. However, the campaigns with the highest average number of likes per message were November 2019 (E1) and December 2015 (E4). As for the average number of comments per message, each party obtained the highest value for this indicator in a different campaign.

Table 6 below shows the messages of each political party that have obtained the most significant impact (measured by the sum of likes and comments):

Table 6 shows that most messages talk about terrorism and national politics (Catalan separatism), related to occurrences close to the General Elections. In November 2015, Paris suffered several terrorist attacks (claimed by the jihadist organisation Islamic State) on terraces, bars, restaurants near the Stade de France, and the Bataclan concert hall. The attacks killed 130 people and injured 325 others (El Mundo, 2015). VOX also alludes to terrorism, referring to the nationalist terrorist organisation ETA (Euskadi Ta Askatasuna in Basque, Basque Country and Freedom), which perpetrated more than 800 murders in Spain (Ministerio del Interior (Gobierno de España), 2019). On the other hand, Catalan independence is another of the social issues with the most significant impact on the data analysed. However, based on the results obtained in Figure 1 and Tables 5, 6, it is worth asking what may have been the reasons for this disparity in the results. In order to answer this question, we will examine the messages from a qualitative and quantitative perspective, checking which social issues have been the most published social by each political party and which have encouraged a more significant reaction from the public.

Figure 3 shows the most published social topics and the average number of likes and comments per message according to each topic.

Figure 3 shows that the most recurrent social topic was national politics, which received the most likes per message. Examining the graph shows that the topic that has obtained the highest number of comments per message is gender. For a more detailed analysis, Tables 7, 8 show the communication profiles of each political party.

Table 7 highlights several aspects: firstly, the least published social topic was the environment, and secondly, the most frequently published topic was national politics. Concerning national politics, it should be noted that the most frequent terms have been those related to the separatist issue or the independence of Catalonia and corruption. The separatist issue refers to the Catalan sovereignty process, known as the “Catalan process.” It refers to a set of social and political events developed since 2012 in Catalonia's autonomous community to achieve its independence from Spain. The 1st of October 2017 marked the most decisive week of the whole process. An unconstitutional referendum took place on that day, in a day with incidents with the police and in which the proper guarantees for a vote were not observed. On the 3rd of October, the King of Spain intervened publicly in defence of the constitutional order and the nation's unity. The European Union and member state governments unanimously took a clear stance against it, and thousands of companies moved their tax domiciles out of Catalonia. The week culminated in a massive demonstration in Barcelona against independence. Despite all this, the Parliament decided to go ahead with its plan: it approved the declaration of independence, which was instantly suspended, and gained no international recognition (Andreu, 2020). According to the Corruption Perceptions Index (Corruption Perceptions Index (CPI), 2019), corruption is another of the most worrying issues in Spain; it ranks 30th out of 180 countries surveyed in 2019.

A general trend was also observed concerning the most minor published topic, which was the environment. On the other hand, there is a disparity between the rest of the social issues. For example, in Cs, the second most published topic was work, PP economy, PSOE gender, UP gender and peace, and VOX peace.

Concerning public reaction, it is interesting to note two facts. On the one hand, there is a disparity between the social issues discussed by the different parties in terms of public impact (measured by the sum of likes and comments). That is to say, the Cs messages that have had the most significant impact have been those that talk about health, precisely one that talks about the Single Health Card to defend equality and to allow people to move freely and be attended to regardless of their place of residence. About the PP, messages about peace have had a more significant impact, with a message supporting Rajoy in response to aggression he suffered before delivering a speech at a meeting in A Coruña standing out. In the PSOE, messages on national politics are the most popular, such as advocating a government of change, which takes charge of the needs of citizens and undertakes public policy reforms that will make Spain a reference point for economic progress and social cohesion in the twenty-first century Europe. In UP, the reaction to a message that talks about Julio Anguita stands out, referring to his facet as a professor, so it would not be a message about education in the strict sense of the word. Finally, in VOX, the messages that stand out for their greater impact talk about the environment, especially one in which Iván Espinosa (deputy secretary of International Relations of the VOX political party) talks about the rural world. Once again, politics is once again taking the limelight away from social issues.

On the other hand, there are also differences between the issues most published by each party and those with the most significant impact. That is to say, the topics most posted by politicians are not the ones that have had the most remarkable public reaction. Table 7 shows that the most published topic in Facebook messages by political parties is national politics, marked by the separatist issue in Catalonia and corruption. However, the only political party that has had significantly impacted the messages that talk about this topic has been PSOE. In the others, the reaction has varied, and the issues that have provoked a more significant reaction from the public have been different in each party.

The mention above leads us to think that there may be an association between political parties and the different social issues they publish. In order to verify this, a simple correspondence analysis has been carried out, which will also allow us to create a perceptual map to determine the positioning of the political parties concerning the social issues they publish about and show the similarities or differences between the parties themselves.

Since correspondence analysis is only appropriate if the variables involved in the model show some degree of dependence, we first test this possible relationship between the publication of the various social issues and the political parties. The Chi-square test tests the null hypothesis that the variables considered are independent of each other against the alternative hypothesis that they are dependent. There is sufficient evidence to reject the null hypothesis for very low significance levels (<0.05) and therefore admit that the variables have some degree of relationship with each other. In Table 9, the significance level is null, so it is possible to admit a relationship between the various political parties and the social issues they post on Facebook. Table 9 also shows a complete solution with four dimensions or components, of which two together explain a proportion of inertia of 72.5%. These results make the application of correspondence analysis appropriate and establish a flat representation of the relationships between the variables that preserves 72.5% of the information.

The perceptual map in the SCA determines a point on a Cartesian coordinate axis for each of the variable categories. The closeness of the points associated with different variable categories gives an idea of the degree of association between them. Based on the Figure 4, two exciting results can be concluded. On the one hand, there is a high association between the PSOE and UP parties, and between VOX and Cs (although the latter is not as strong), and on the other hand, concerning the possible associations between social issues and political parties, the following can be deduced:

There are issues that all political parties equally address, such as education and culture.

Other issues are more associated with some parties than others. In particular. There is a relationship between the themes of national politics and peace with the parties VOX and Cs, an association between the themes of environment, gender and health with PSOE, and a relationship between employment with PP. Therefore, we can conclude that social issues discriminate or better distinguish the different political parties, accepting the proposed hypothesis.

It should also be remembered that the SCA uses the Chi-square metric to measure the similarity between the different profiles. In this metric, modalities with low frequencies are weighted more favourably. It leads to the categories associated with low frequencies being located in extreme positions on the map. In this sense, it is noteworthy that the environment is positioned at one of the extremes, as is the Cs party, indicating the low impact of these categories in the study. In conclusion, topics such as the environment, employment and gender, although they still have some correspondence with certain political parties, are losing their discriminatory power in the context of conversations on Facebook on topics related to the SDGs.