Special issue on methods in visual politics and protest Mixed methods, data curation & anti-publics

This special issue draws together five articles in the arena of methods in visual politics and protest. It addresses three core methodological challenges across the research process (data access, collection, analysis, visualisation): the proliferation of visual social media, the emergence of novel visual practices, and the increasing application of digital methods. Their key contributions lie in the development of mixed visual methods approaches, new techniques for constructing and curating visual datasets, and methodological explorations of visual anti-publics


Visual Methods
'Visual methods' or 'visual research methods' refer to a range of methodological approaches and techniques for the collection and analysis of visual data, or research methods using visual materials (see Rose 2014).Visual methods pre-date social media research, for example, through the use of photoelicitation, video/film, photography, drawing, collage, and other art forms in participant research such as ethnography (see, e.g., Gauntlett and Holzwarth 2006;Buckingham 2009;Margolis and Pauwels 2011;Rose 2014).Even so, new methodological challenges have arisen following the proliferation of 'visual social media', platforms driven by visual information and communication such as Instagram, TikTok, YouTube, and Douyin (Leaver et al. 2020;Rogers 2021), the ubiquity and networkedness of smartphone photography since the 2010s (Hand 2012;Van House 2011), and the growing use and circulation of digital visual artefacts or digital images across different digital platforms (Rogers 2021).
Research designs consequently need to be adapted in light of new formats, platform logics, and the specific practices that are linked to the evolving infrastructural and technical properties.Challenges arise across the different phases of the research process: data access, data collection, data analysis/interpretation, research ethics, and data visualisation.In data collection, challenges lie in the discoverability of visual contents where they are often linked to textual information, for example through the use of keyword-or hashtag-based queries (where software is used) or specific social media accounts, groups, or communities.In data analysis, challenges lie, above all, in the often implicit nature of visual contents as they may be interpreted differently across varying audiences, for example based on cultural interpretations of specific symbols or the polysemic nature of memes (Boxman-Shabtai and Shifman, 2014), i.e. a message may be interpreted differently based on the specific political or ideological context in which they have been shared.
In response to these challenges, the Visual Cultures section of the European Communication Research and Education Association (ECREA) organised a pre-conference on "Visual Politics and Protest -Current Methodological Challenges" online in October 2022.This special issue follows the themes presented at the pre-conference and draws together a range of papers advancing research methods in visual politics and protest in two parts.This volume constitutes the first of these.In what follows, we outline the rationale of the special issue focus, lay out a range of challenges in contemporary visual research, showcase how the selected articles tackle these issues, and introduce the articles included in this first part.

Visual in politics and protest research
Visuals have long been prominent in political action based on their multiple functions: they may establish an argument, set an agenda, appeal to the audience's emotions or cultural values, and construct (or destroy) a political figure's reputation (Schill 2012).Through their wider distribution, they can also create a sense of collective political identity, for example through connective witnessing (Mortensen 2015) and an embodied collectivity (Pantti 2013), thereby amplifying political messages or protest events (Papacharissi 2016;Valaskivi and Sumiala 2023).While politicians may invest effort in harnessing visual forms of communication to affect the way the public perceives them, whether by creating a powerful, authoritative image (Lalancette and Raynauld 2019;McFarlane 2016) or a channel for personalised connection with their supporters (Farkas and Bene 2021), private individuals may also employ visuals for creating awareness and grouping people around a specific cause (Rovisco and Veneti 2017).
While visual forms of communication have become firmly embedded in global media ecologies, they have shown particular relevance across the spectrum of political action -whether this relates to political campaigns and elections, social movement action repertoires, or the wider practices subsumed under the umbrella of activism (e.g.see Karatzogianni 2015;Özkula 2021).For the purpose of this special issue, we consequently define scholarship on visual politics and protest as the wide remit of research that considers the use of visual formats in politically motivated digital communication.Given this issue's focus on digital communication, this includes, for instance, photographs, cartoons, memes, short or longform videos, GIFs, and emojis shared on social media platforms.Although visual politics and protest have existed long before the advent of the internet (indeed, a considerable amount of social movements research has been dedicated to protest photography, posters, and placards), this issue foregrounds current visual political practices that are embedded in digital networks.
Since these practices are generated by both institutional and private media users, this special issue draws together methodological perspectives across the wide spectrum of activist and political activities.This includes the use of visuals by established institutional political actors, organisational entities, and social movements, as well as networked protest activities -sometimes drawn together under the syntagm of 'digital activism'.The term has variably been described as digitally facilitated dissent, resistance, and rebellion (Hands 2010), or, in more nuanced terms, by Karatzogianni (2015, p. 1) While definitions of digital activism and protest culture conventionally exclude governmental actors, this particular issue considers the wider repertoire of political actions (whether bottom-down or bottom-up) as the departure point.Extant research has shown various visual manifestations of these practices for embedding, expressing, and spreading political messages.This has, for example, included the circulation of protest photography and campaign posters on social media platforms, the use of image filters and frames for articulating political opinion or digital solidarity, 'protest avatars' -political social media profile pictures for collective identification (Gerbaudo 2015), 'hashflags' -temporary visuals attached to hashtags that endorse specific (often political) causes, movements, and interests (Highfield and Miltner 2023), and various forms of video-based activism (e.g., Askanius 2012;2013;2016;2019;Uldam and Askanius 2013).A range of scholars have also drawn attention to specific visual formats or platform features that have led to distinct video-based practices, vernaculars, or phenomena.Such practices have been conceptualised as meme-based or memetic activism -a form of visual activism characterised by the transformative nature of viral social media images (Pilipets and Winter 2017), 'playful activism' -a mode of political engagement that blends audiovisual elements with competitive challenges on social media platforms (Cervi and Divon 2023), and 'slideshow activism' -a social media practice drawing on image sequences in PowerPoint-style, also known as a 'carousel' (Dumitrica and Hockin-Boyers 2022;Schreiber 2023).
This special issue aggregates a range of papers that explore these visual political practices with a specific focus on methodological concerns toward a better understanding of visual politics and protest.

Methodological Challenges
While visual methods constitute a longstanding tradition in social research methods, digital social research has opened up new challenges and opportunities in this field, above all in response to three key developments: 1) the proliferation of visual social media; 2) newly emerging visual practices; and 3) the growing repertoires and application of digital methods.
Platforms.Although internet cultures have drawn on visual formats since their beginning, the proliferation of visual social media and the related photography practices have arguably led to a 'visual turn' in global media ecologies.Visual social media platforms have developed logics that foreground the visual in communication as they focus on the content production, consumption, and (re)circulation of, for example, long-or short-form videos, memes, and cartoons.In part, these developments have emerged from processes of platformisation, a process described by Anne Helmond (2015) as "the rise of the platform as the dominant infrastructural and economic model of the social web", a consequence of the growing popularity of social media platforms as well as the platform affordances these have given rise to.Platform affordances, "(...) the perceived actual or imagined properties of social media, emerging through the relation of technological, social, and contextual, that enable and constrain specific uses of the platforms" (Ronzhyn et al. 2023, p. 14), affect user practices on a given platform, including if and how visuals are shared and to what extent they are central in the communication of political ideas.
Practices.Changing platform logics have at times led to the establishment of new visual practices.This has already been the case prior to visual social media, as exemplified by new forms of visual-textual communication that have emerged through the introduction of emojis, GIFs, stickers, cartoons, and memes.Even so, the growing popularity of visual platforms has created a breadth of new practices that build on the new capabilities of these spaces.This includes algorithmic suggestion mechanisms on the basis of popular visualities and the related audio-/soundscapes (e.g. on TikTok), as well as the replication and re-interpretation of these visual contents through processes of remixing and memefication.Popular visual content is often recreated or re-imagined, with new interpretations sometimes becoming more popular than the originals (e.g., social media "challenges").In relation to protest and politics, new practices have, for instance, included the visual communication of political ideas through specific Research methods.To an extent, scholarship has already addressed the challenges arising from these developments.A range of new platform-sensitive visual methods have developed over the years, of which some are platform-tailored, such as "Instagrammatics" (Highfield and Leaver 2016) and others are platform-comparative, such as "visual cross-platform analysis" (Pearce et al. 2020).These and other approaches have sought to explore new digital-visual phenomena, contextualise them within and/or across individual platform ecologies and cultures, and provide new means of collecting or analysing digital data (see Highfield and Leaver 2016;Marres 2020;Pearce et al. 2020;Rogers 2017;2021).In recent years, many of these questions and solutions have focused on 'digital methods', methods and data native to digital environments and drawing on digital objects towards understanding web-based sociocultural phenomena (Rogers 2019).On the one hand, these methodological approaches have provided new ways of sampling, collecting, analysing, and visualising visual data, for example, through new software packages made available to researchers.On the other hand, these methods may not always provide contextualised and rich accounts that consider wider media practices or medium-specific influences (e.g., see Venturini et al. 2018).
The articles to this special issue contribute to these debates through new methodological techniques and considerations in a field in which methodological approaches such as single-platform Twitter/X research have been prevalent (see Özkula et al. 2023).The contributions in this volume explore a range of visual formats and practices in digital politics and protest, including image or video posts, memes, short videos, and bots across different platforms (Instagram, TikTok, and Twitter) and country contexts (Germany, Italy, Brazil, Portugal, Ukraine, and Russia).In doing so, they draw attention to the ways in which visual practices are embedded in cultural environments based on the platform and national context they take place in.They draw attention to visual logics within their chosen environments and illustrate how methodological approaches and techniques can be adapted to capture these dynamics.In accumulation, these works provide contributions to the field in three thematic areas: mixed visual methods, dataset-building techniques, and explorations of visual anti-publics.

Mixed Visual Methods
The first thematic field of this issue lies in its calls for and proposed techniques for 'mixed visual methods'.Methodological approaches are strongly connected with the theoretical traditions with which they share common roots and ontological premises.While distinctions between qualitative and quantitative research traditions persist, there is a longstanding push for methodological pragmatism.This approach advocates for integrating and blending methods from diverse, and sometimes contrasting, traditions to facilitate more contextualised research of phenomena (Morgan 2007).In part, this trend also follows new epistemological developments in digital social research as digital methods add to the blurring of boundaries between data collection, analysis, and visualisation, as well as between macro-and micro views, induction and deduction (see, e.g., Venturini and Latour 2009), perhaps best illustrated in the notion of a 'quanti-quali approach' (see, e.g., Venturini and Latour 2009;Rogers 2019).The papers presented in this issue contribute to this discussion by providing examples and arguments for crossing methodological boundaries and using complementary methods to circumvent common obstacles and challenges in visual research.
For example, Caldeira illustrates how digital methods can be incorporated into feminist media studies, a research area historically favouring qualitative and interpretive approaches.She argues that mixing digital methods and interpretive feminist methods produces research that provides rich contextualised meaning.Along similar lines, Omena et al. use "quali-quanti visual methods", an approach tailored to the specificity of the chosen platforms, their subcultures of use, and forms of appropriating digital-visual objects in an exploration of social media bots.In doing so, they generate a cross-platform contextsensitive analysis.Rich mixed methods approaches are also employed by Hohner et al., who use computational methods to combine computer vision techniques for classifying images (video frames) with manual (i.e.interpretive) annotations for assigning labels to broader categories.This contextualised approach allows them to categorise multimodal TikTok content produced by far-right groups through a triangulated effort.Mixed-method triangulation is also used by Giorgi and Rama who combine manual content analysis and automated visual analysis for clustering images by pixel similarity, a methodological choice that enabled them to make sense of memes' narrative flows and map recurring templates and their evolution over time.In doing so, these papers illustrate benefits gained from "mixed visual methods": contextualisation, triangulation, closer and multi-faceted readings of visual data, as well as a more detailed engagement with diffused and polysemic visuals.

Dataset-building through soundscapes, social cues, and macros
The second thematic field of this issue is the introduction of new methodological techniques for sampling, building, and curating visual datasets that move beyond the prevalent practice of collecting visual data through text inputs (e.g., keyword or hashtag queries).First, through soundscapes.Using the example of Ukraine-Russia war propaganda on TikTok, Geboers and Pilipets employ TikTok's sound infrastructure for detecting memetic patterns.This approach recognises that sound serves as a primary searchable template and network feature on the platform.This method allows them to navigate TikTok's soundscape, focusing on what they term "memetic masterplots‫,״‬ plots woven from a mix of visual and textual components, including video stickers, music, and physical gestures.Second, through social cues.Omena et al. develop a query design and dataset curation method that relies on bots' social cues, visual stereotypes, and following networks.As such, their starting point is no longer the output of bot-scoring detectors, but the socio-technical knowledge of factors relevant to bot networks.Third and finally, through ephemeral meme macros.In their visual analysis of the 2019 Italian government crisis, Giorgi and Rama analyse meme macros (i.e.text superimposed on images; here on memes), through automated and interpretive content analysis.They analyse what they subsequently call "contingent macros", shortlived/ephemeral memetic layouts that are not able to establish themselves in the visual vernaculars and eventually become obsolete.As such, they focus on the ephemerality and polysemy of memetic contents.In aggregation, the methodological techniques presented in these papers address the challenges of collecting and curating visual data in dynamic, multimodal and transient visual communication on social media platforms.

Exploring visual anti-publics
The third thematic area of this issue is the analytical focus on cases of 'visual anti-publics'."Anti-public spheres" refer to online spaces of socio-political interaction where "discourse routinely and radically flouts the ethical and rational norms of democratic discourse" (Davis 2021, p. 143).This may include fora and practices with white supremacist, anti-climate science, homophobic, racist, misogynistic, and conspiracy theory discourses (Davis 2021).While some of these spaces have entered mainstream public spheres, they are at times (due to their contentious contents) also hidden in subaltern spaces, restricted to certain social groups, or camouflaged towards avoiding content moderation of radical or emblematic content such as propaganda and conspiracy theories.Anti-publics are more often "de-platformed" from mainstream platforms and move towards alternative and less-moderated spaces.In light of these movements, tactics, and rhetorics, anti-publics are methodologically particularly challenging for identifying, sampling, and reading visual contents.This special issue presents a range of case studies that provide insights into these "visual anti-publics": Geboers and Pilipets focus on Russian propaganda on TikTok, Hohner et al. on German far-right videos on TikTok, and Omena et al. on "bolsobots", pro-and anti-Bolsonaro bots in the Brazilian context.These cases illustrate the specific ways in which visual artefacts are employed to recruit and keep users' attention.For example, Geboers and Pilipets illustrate how the use of sound for researching online propaganda provides a deeper understanding of the use of affect and attention-grabbing in memetic masterplots.Along similar lines, Hohner et al. attribute significance to the use of intimacy and familiarity in far-right vlogs.They identify 'talking heads' in vlog-style videos as the primary visual logic of the most popular far-right group in their dataset.Finally, Omena et al. explore bolsobots on the basis that they are identifiable based on specific social cues that allow them to separate bots from their wider sociotechnical environment.While these projects chose different entry points for identifying and making sense of these spaces, they all developed tailored approaches for gaining insights into anti-democratic visual practices.
In aggregation, the papers presented here outline new methodological protocols, based on changing media landscapes that attribute greater significance to visuals in digital communications, as well as the practices that have emerged from this, and the platforms that build on these principles.In particular, they highlight contributions afforded by mixed-methods such as quali-quanti approaches, and combinations of software-based data with on-the-ground engagement and interpretive readings.In accumulation, they also address and discuss the complexities and challenges of multi-and cross-platform study, and advocate the deployment of platform-tailored methodologies.

Featured in this special issue
In what follows we briefly outline the individual papers included in this first part of the special issue.
Hohner et al. contend that TikTok serves as a mobilisation platform for German far-right factions.Their study introduces a unique methodological approach, integrating manual annotation with image frame classification and machine-learning-driven content analysis, to examine the tactics of these groups.They point out that the popularity and engagement levels vary among different far-right factions on TikTok, with nationalist and conspiracy-themed content drawing significant attention.Their study underscores the dangers posed by the proliferation of extremist ideologies on social media platforms and advocates for a systematic monitoring of far-right content on TikTok.
Giorgi & Rama present the idea that the intricacies of Internet memes transcend the conventional concept of an "image macro".An image macro features a photo or image topped with bold, white text, commonly with a black outline and adhering to specific templates.They introduce the term "Contingent Macro" to situate memes within specific contexts and encapsulate the transient/ephemeral nature of memes as tools for discourse.Their work stresses the importance of contextually nuanced meme definitions and offers a methodological framework for examining meme production related to specific events.They illustrate how memes significantly shaped narratives during the 2019 Italian government crisis, using metaphors to depict power dynamics and political alliances.Their argument challenges traditional meme research methods, advocating for more adaptable definitions and methodologies to thoroughly understand the complex nature of meme content.
Omena et al. argue that to effectively understand and identify bots such as "bolsobots" in the Brazilian context, a holistic approach that merges qualitative and quantitative techniques is needed.They critically examine current bot detection methods and advocate for a novel conceptual and analytical framework in bot research.This framework underscores the importance of acknowledging the dynamic characteristics of bots and the intricacies of automated political conduct.They emphasise the role of image repetition and the connection between bots' profile images and related web entities.This is crucial for understanding bot operations within larger networks and their interaction with political identity cultures online.
Caldeira posits that blending digital methods into feminist social media research (using the case of Instagram) offers both opportunities and challenges.She does so through a mixed-methods visual analysis of four feminist hashtags within the Portuguese context.She contends that digital methodologies can yield profound insights into feminist activities on social media, but also bring up ethical and methodological considerations.Caldeira calls for a thoughtful and engaged approach towards digital tools and techniques, and highlights the importance of employing feminist ethics of care in digital social research.She stresses the necessity for methodological innovation in feminist media studies and underscores the critical need for scrutinising digital tools and methodologies in future academic work.
Geboers & Pilipets explore pro-Russian war propaganda on TikTok, focusing on music, stickers, and comments to understand their affective impact and the connections they forge between different expressive forms.They analyse TikTok's 'soundscape' and conduct word pair analyses of comments to examine the resonance and dissonance in video performances and commentaries.Their study employs montage techniques to dissect the sequential narrative and spatial organisation in videos, particularly with the Katyusha remix, revealing memetic linkages and theatrical effects.They map the frequency of word pairs in comments, visualised through heat map-like colour intensities, and consider the role of emojis in fostering collective expressions of belonging.
, as: artefacts, and formats (e.g., augmented reality, artificial intelligence, filters, overlays, slides, emojis), forms of documentation (e.g., witnessing, commenting, stitching, or dueting on TikTok), and various visual styles and narratives (e.g., those specifically adapted for various types of video activism), practices for which methodological approaches need to be adapted.