• social media
• political communication
• Instagram
• images
• visual communication

Images are an important instrument in political communication strategies (Lilleker 2019), and in the last few years have gained an increasing relevance both for citizens and political candidates (Schill 2012; Towner 2018). This late importance given to these elements is deeply related to the diffusion of new communication technologies (Hjorth and Hendry 2015). On the one hand, to the diffusion of social media and their features which have made images a pivotal element to gain both visibility and popularity during electoral campaigns (Dayan 2013; Parry 2015). On the other hand, to the production of high-quality visual elements by users which can further be shared on social media facilitated by mobile technologies with integrated photo cameras (Messaris 2019).
In this context, digital platforms – especially image-centered ones – have driven political actors to adopt strategies that are ever more often similar to the ones used in the world of entertainment by celebrities (Muñoz and Towner 2017). Moreover, the increasingly frequent use of these platforms has played a key influence in the performative turning point in online political communication in which personalization processes, which were typical practices in traditional media (Van Aelst, Sheafer, and Stanyer 2011), where now being used by political actors on their political communications (Holtz-Bacha, Langer, and Merkle 2014). A performative turning point in digital online political communication resulting from wider social changes related to – among others – our current era of individual mass communication (Castells 2009). Images, in this respect, are strategically used by political actors for promotional purposes, influencing engagement through the diffusion of visual bites of their private lives and emotions (Ekman and Widholm 2014). These have led to more individualized political campaigns (Enli and Skogerbø 2013; Kruikemeier 2014), particularly on social media platforms (Rega and Bracciale 2018), and specially on image-centered platforms (Schroeder 2016), triggering a process of “hyper-personalization” of the political leader’s image.
The present study represents a first attempt to individualize the current declinations that political personalization assumes through images shared on social media (Van Aelst, Sheafer, and Stanyer 2011; Van Santen and Van Zoonen 2010). In the last years, researches on personalization of politics have considered for the operationalization of the phenomenon variables identified on traditional mass media, readapting them to study social media platforms (Filimonov, Russmann, and Svensson 2016; Jung et al. 2017; Ekman and Widholm 2017a; Russmann, Svensson, and Larsson 2019). It is argued that this re-adaptation has not considered neither the structural characteristics of social media platforms nor possible new communication trends that could have developed on these platforms, especially those more image-centered ones.
Instagram is a perfect example on this regard as it is a platform for which personalization is part of its constitutive affordances (Larsson 2019). This platform, launched in October 2010 – one of the most used ones in Italy – allows users to easily capture their activities and life moments using visual elements and share them with other users (Hu, Manikonda, and Kambhampati 2014). It also enables to effortlessly transform images using filters or adding their own text overlays in a process of reconstruction of the image enabling their recirculation (Miller et al. 2016).
Though the importance of this platform and its adoption by many political actors has been studied in the last few years (Jung et al. 2017; Ekman and Widholm 2017a; Muñoz and Towner 2017; Russmann, Svensson, and Larsson 2019), there are still many unanswered questions concerning aspects explicitly related to image usage vis-a-vis the ever more often personalization of politics.
Personalization of politics is not the only aspect that could be identified on political actors’ communication, especially on a visual perspective. In this vein, with the constant fragmentation of the media landscape and the everyday appearance of new communication channels, political communication strategies have been further complemented with processes – in addition to personalization – that are very much related to the popularization of politics (Mazzoleni and Sfardini 2009). Through these processes political leaders seek to build not only a more likable figure, but also a figure ever more similar to the ordinary citizen using popular culture elements (Street 2015), namely we refer to celebrity politics and lifestyle politics (Ceccobelli 2019).
Therefore, this study concentrates on three main aspects being the first one to shed light on both the numerous forms that personalization assumes on Instagram and the specific elements that characterize it on this social media vis-à-vis the elements of mainstream media. The second aspect is to understands political actors’ visual communication also through non personalized images, namely all pictures where the leader is not present. And the third one is to identify which visual aspects influence Instagram users’ engagement.
The organization of this work will see in its first part a theoretical framework that intends to reconstruct the concepts related to the popularization and personalization of political communication and the transformations of the latter from traditional media to new technology developments, in particular social media. The second part of the theoretical framework reconstructs the role images have on political communication, the differences that characterize these elements and the different researches that have been done on this regard especially on image-centered platforms like Instagram. The research design is presented in the following section which begins by describing the data to be analyzed and a detailed explanation of both the automated image classification through third-party convolutional neural networks used for content analysis, and the statistical methods used for data analysis. The structure of the thesis continues with the discussion of image analyses results beginning with an overview of Instagram uses by Italian political actors and two different cases:, i) the first one which is explicitly focused on images displaying the political leader aimed at the identification of new dimensions of personalization process; ii) the second case is an exploratory one which is intended to analyze leaders’ visual political communication also in those images where they are not present. Finally, in the last section of this thesis are presented the conclusions.
Within this context, the following research questions will guide the present study:
RQ1. Which are the main visual elements related to personalization of politics?
RQ2. Can specific visual personalization styles be identified on social media?
RQ3. Do visual personalization aspects influence engagement?
Furthermore, considering that visual communication of political leaders on social media platforms is composed also by not personalized images, further research questions have been added to the study:
RQ4. Which other visual dimensions are identifiable in addition to the personalization ones?
RQ5. What is the relation of these non-personalized elements with engagement?