Visibility Under Constraint: Format Publics and the Politics of Visibility on TikTok Protest Videos During the Mahsa Amini Movement in Iran

 

Fatemeh Oudlajani[1]

Allameh Tabataba’i University, Iran

 

 

Short-video platforms have become important sites for protest communication, yet we know little about how visibility is negotiated when appearing on camera entails political risk. This study examines TikTok videos produced during the Mahsa Amini (“Woman, Life, Freedom”) movement in Iran under conditions of state repression and surveillance. Drawing on a mixed-methods content analysis of 145 videos, the study analyzes how narrative style, symbolic imagery, audio cues, and hashtag practices relate to two dimensions of visibility: reach and engagement. Symbolic visuals, such as hair cutting and scarf burning, are associated with higher reach, while original speech, ambient sound, and protest chants predict higher engagement. Hashtag practices show no meaningful relationship with reach. Building on these findings, the article advances the concept of format publics and conceptualizes visibility under constraint as a layered condition shaped by state surveillance and platform curation.

 

Keywords: TikTok, digital activism, algorithmic visibility, affective publics, Iran

 

Fatemeh Oudlajani: [email protected]

Date submitted: 2025-11-24

 

Publics on social media have been theorized as affective (Papacharissi, 2014) and connective (Bennett & Segerberg, 2013). These frameworks explain how emotion and personalized participation mobilize people in open digital environments. They do not, however, account for visibility when expression is risky, and distribution is shaped by platform-format logics. Despite a growing body of research on digital activism, we still lack a clear framework for understanding how publics negotiate visibility on short-video platforms when appearance can trigger political risk.

 

On short-video platforms like TikTok, participants navigate political repression and platform curation that influence which content is promoted or suppressed. This study does not observe TikTok’s algorithm directly; it infers platform influence from observable patterns in content, reach, and engagement.

 

I theorize visibility under constraint: a condition in which expressive participation must be simultaneously politically safe and platform-recognizable. This concept highlights visibility not as a simple outcome, but as a strategic balancing act shaped by both state surveillance and platform formatting norms. I argue that participation under constraint travels through format publics—collectives that cohere around reproducible, low-exposure, platform-aligned formats (e.g., symbolic visuals, shared protest sounds) rather than fixed identities or organizational leadership. In this account, format is not decoration, but mechanism: It mediates who can appear, how far content circulates, and how audiences respond.

 

Empirically, I test this framework with a mixed-methods study of 145 TikTok videos from the Mahsa Amini movement. I distinguish two outcomes—scale (reach) and responsiveness (engagement)—and ask how multimodal tactics (symbolic imagery, authentic/protest audio, and hashtags) shape each outcome under constraint.

 

The analysis assesses whether the creative levers that expand visibility also intensify interaction. This approach advances debates on visibility, affect, and participation by specifying how format-driven platform logics condition protest communication in repressive contexts. It proposes the format-publics model for understanding digital activism, where what travels is a repeatable format that balances platform legibility with reduced personal exposure. This article, therefore, provides a conceptual and empirical framework for understanding how multimodal protest expression circulates in environments characterized by political repression and algorithmic governance.

 

Literature Review

 

Networked, Affective, and Connective Publics

 

Digital media scholarship has long emphasized that publics on social platforms emerge through the interaction of technological infrastructures and user practices (boyd, 2010). Two influential models capture the dynamics of contemporary mobilization. First, Papacharissi’s (2014) notion of affective publics describes collectivities that cohere through shared affective expressions, ambient communication, and distributed storytelling. These publics form not through formal organization, but through emotional circulation and connective threads of personal narrative.

 

Second, Bennett and Segerberg’s (2013) model of connective action explains how large-scale mobilizations operate without strong organizational leadership. Here, personal action frames—flexible, individualized calls to participate—travel easily across platforms because participants adapt them to their own repertoires.

 

While these frameworks illuminate how publics emerge through emotion and personalization, both assume communicative settings in which visibility is relatively safe. They leave less conceptual room for contexts in which public expression is dangerous, and participation must be carefully navigated. In repressive environments, the problem is not only how publics connect but also how they manage when, how, and through which formats they appear.

 

Iran’s digital sphere has long been shaped by cycles of political expression and state repression. Sreberny and Khiabany’s (2010) account of “Blogistan” documents how online participation in Iran has historically been entangled with surveillance, risk, and state control, creating a communicative environment in which visibility is never neutral. This longer history provides crucial context for understanding contemporary protest communication on platforms like TikTok, where participants continue to navigate the dangers of appearing online while engaging in new, multimodal forms of collective expression.

 

This shift in emphasis—from whether to be visible to how safely to be visible—requires reconsidering how affect, expression, and networked organization operate when exposure carries risk. Empirical work on hashtag activism further shows how networked and affective publics connect online discourse to offline struggle.

 

Freelon et al. (2016) study of #Ferguson and #BlackLivesMatter demonstrates that hashtags can serve as infrastructures for information sharing, narrative contestation, and protest coordination, linking social media practices to demands for offline justice. This research underscores how publics use platforms to weave together storytelling, visibility, and mobilization, focusing on text-centric environments like Twitter/X, where participation is anchored in hashtags rather than multimodal short videos.

 

However, existing work offers little guidance on how these dynamics unfold on short-video platforms, where multimodal template use shapes visibility in ways distinct from text-centric environments.

 

From Hashtag Publics to Template-Based Participation

 

As social platforms evolve, so do the infrastructures that shape how publics assemble. On text-based platforms such as Twitter/X, attention often coheres through hashtags, which serve as organizational markers that structure discourse and visibility (Bruns & Burgess, 2015). Short-video environments operate differently. Platforms like TikTok organize participation around audio-visual templates: shared sounds, gestures, filters, editing grammar, and visual tropes. These standardized expressive building blocks form the basis of what Zulli and Zulli (2020) call TikTok’s “technological mimesis,” in which imitation and replication are encouraged by design. As Kaye et al. (2022) also argue, TikTok’s short-video grammar is built around modular, repeatable templates that structure how expression becomes visible.

 

Recent research on TikTok in conflict contexts demonstrates how short-video platforms enable civilian participation through standardized and platform-recognizable multimodal formats (Kalnes & Bjørge, 2025). Examining Ukrainian women’s participation during the war, the authors show how music, visual cues, and narrative templates function as the primary means of political expression, allowing content to circulate without relying on formal organizational structures. Participation in this context involves visibility labor, as creators adapt their performances to TikTok’s media logic in anticipation of circulation and attention.

 

Cervi et al.’s (2021) analysis of Podemos’ TikTok strategy further illustrates how political actors adapt to the platform’s visual, affective, and memetic language to reach younger publics. Their findings show that TikTok’s communicative grammar, informality, performativity, and highly stylized multimodal templates reshape how parties construct political messages. While this work demonstrates how institutional actors mobilize TikTok’s expressive resources in democratic contexts, it also highlights a gap: Far less is known about how noninstitutional protest publics adopt and reconfigure these same platform-native formats under conditions of repression. This study addresses that gap by examining how TikTok’s political “language” functions when visibility carries significant personal and collective risk.

 

Broader scholarship on digital activism also highlights how such platformed expressive logics shape contemporary collective action. Reviewing the evolution of social-media-based activism, Castillo-Esparcia et al. (2023) argue that affective, low-threshold, and highly replicable formats now underpin much activist communication. However, they also note that activists face structural disadvantages in terms of algorithmic visibility compared with influencers or institutional actors. This tension between participatory affordances and unequal circulation provides an important backdrop for understanding how TikTok users engaged in high-risk protest, such as those in Iran, adopt platform-native formats while negotiating limited visibility and political danger.

 

Abidin’s (2021) work on TikTok influencer cultures similarly shows that the platform’s attention economy is organized around visibility labor, in which creators continuously tune their content to platform norms, trends, and templates to secure reach and monetizable celebrity. Her analysis demonstrates that duets, challenges, and other iterative formats function as standardized, low-friction expressive units that travel more easily than singular, authored messages. I take this as a baseline account of how TikTok structures ordinary celebrity visibility and extend it to a high-risk setting by asking how the same format-dependent logics are reconfigured when participants are protesters facing repression rather than commercial influencers seeking brand growth.

 

Lee et al. (2023) argue that TikTok’s short-video culture and participatory affordances are increasingly mobilized by social movement actors and publics, transforming pop-cultural formats into vehicles for activism and civic engagement. This perspective integrates platform form, cultural template logic, and protest mobilization. It provides a valuable foundation for examining how template-based participation on TikTok is repurposed in highly repressive settings where the logics of visibility and format must be managed under risk.

 

Recent work on TikTok activism further underscores the centrality of remixable formats and depersonalized participation. Gitomer et al. (2024) show how youth activists use TikTok’s remix tools to produce what they call “(de)personalized” political content, where participation often occurs through templated sounds, filters, and editing conventions rather than face-forward testimony. In their account, remix enables activists to embed political messages into recognizable trends, allowing users to signal alignment while partially displacing individual visibility onto shared templates (Gitomer et al., 2024). This emphasis on depersonalized remix aligns with the idea that short-video publics assemble around standardized formats and provides an important point of comparison for analyzing how similar strategies are adapted in high-risk, repressive settings.

 

Participation in such environments is enacted less through text or argument and more through the reuse of formats: users signal alignment by repeating a sound, gesture, or symbolic act. What remains underexplored is how template-based participation functions in political communication, particularly in settings where individuals face real consequences for appearing on camera. Under such conditions, standardized formats can offer a protective space: they enable contribution without requiring personal disclosure, allowing publics to assemble through depersonalized expressive templates rather than face-forward performance.

 

Visibility, Power, and Platformed Norms

 

Visibility is not merely an outcome on digital platforms; it is a field structured by power. Brighenti (2007) conceptualizes visibility as a strategic and relational process in which being seen can generate recognition or impose control. Visibility is thus inherently ambivalent: It can empower marginalized actors through public recognition, but also expose them to surveillance and sanction.

 

In platform contexts, this ambivalence intersects with the techno-commercial logics of platform governance. Van Dijck and Poell (2016) describe contemporary public life as embedded in a platform society, where ranking systems, curation mechanisms, and data-driven optimization shape communication. These systems privilege certain media forms—short, emotive, and visually legible formats—while suppressing others.

 

Building on this, Gillespie (2018) conceptualizes platforms as “custodians” of the Internet whose moderation practices quietly organize what can be seen and spoken. Rather than neutral conduits, platforms actively sort, remove, and downrank content through a mix of policies, human labor, and automated systems that are only partially visible to users (Gillespie, 2018). These hidden moderation decisions define the boundaries of acceptable expression and shape which publics can materialize in the first place. In the context of TikTok protest communication, such custodial governance intersects with recommendation logics to determine not only what becomes visible but also which forms of dissent are rendered risky, illegible, or effectively invisible.

 

Building on this governance perspective, Cunningham and Craig (2019) argue that platforms exercise power over creators not only through formal moderation but also through ongoing and opaque systems of creator governance. Visibility is shaped by shifting platform policies, algorithmic adjustments, and incentive structures that encourage anticipatory compliance over direct enforcement. Importantly, they show how platform self-regulation enables states and other institutions to shape visibility indirectly, allowing political power to be exercised at a distance. From this perspective, creators internalize platform rules and adjust their expressive practices in advance, treating visibility as a conditional and negotiated outcome rather than a stable reward.

 

Bucher’s (2018) account of algorithmic power further clarifies how such hierarchies of visibility are operationalized. She shows that platforms govern participation through conditional “if . . . then” logics that shape what becomes amplifiable, searchable, or suppressed. Visibility is therefore not a neutral outcome of user activity, but the effect of programmed thresholds and opaque decision rules. This perspective foregrounds TikTok not merely as a communicative environment, but as a governance infrastructure that structures what kinds of protest expressions can circulate and under what constraints, making algorithmic power integral to understanding visibility under repression.

 

As a result, publics must orient themselves not only to political risk but also to the expressive norms that platforms reward.

 

Bishop’s analysis of “algorithmic gossip” on YouTube shows how creators manage visibility in conditions of algorithmic opacity by developing folk theories about what the recommendation system rewards and adjusting their practices accordingly (Bishop, 2019). This emphasis on user interpretations of platform logics complements accounts of formal governance: Visibility is shaped not only by ranking infrastructures but also by how publics imagine and respond to them. In the context of TikTok protest communication, such speculative knowledge about which formats, sounds, and symbols might travel under opaque curation becomes part of how participants navigate visibility and risk.

 

Similar dynamics are observed by Duffy and Meisner (2023), who show that marginalized creators experience sudden losses of visibility and opaque moderation that reinforce the instability of platform-governed attention.

 

Research on digital activism under repression similarly shows how participants negotiate visibility amid threats of surveillance, harassment, and state retaliation. Kreutz and Makrogianni (2024), examining the transnational #MilkTeaAlliance movement, demonstrate how activists adapt their communicative practices to mitigate risk while sustaining collective presence across borders. Their findings highlight that repression produces distinct visibility dilemmas: Actors must appear enough to maintain momentum yet remain obscured enough to avoid identification. These dynamics parallel the constraints faced by Iranian TikTok protest publics, for whom participation requires balancing political danger with the platform’s demand for legible, repeatable formats.

 

In repressive contexts, these pressures converge. Participants must manage a layered field of visibility: state visibility, which threatens personal safety, and platform visibility, which determines circulation and discoverability.

 

The condition that emerges from this intersection is what I refer to as visibility under constraint: a communicative condition in which appearing publicly requires simultaneously minimizing individual risk and aligning with platform-format norms that determine reach. Rather than choosing between visibility and invisibility, participants negotiate how to be visible safely.

 

Iran, Digital Activism, and Visual Protest

 

Iran’s recent protest movements have unfolded within a highly constrained media environment, where public expression is closely monitored and often criminalized. Long before TikTok, activists relied on a multiplatform ecology—including blogs, Facebook, Twitter, Telegram, and Instagram—to document state violence, sustain anonymity, and build transnational solidarity networks. Rahimi’s (2011) work on the 2009 Green Movement shows how social media functioned simultaneously as a space of dissent and a field of state power, rather than a simple tool of liberation. Building on this, Faris and Rahimi’s (2015) edited volume on social media in Iran and Danesh’s and Athari’s (2024) recent study of cyber activism during the 2022 protests demonstrate how networked communication has become integral to contentious politics under authoritarian rule.

 

Scholars of the “Woman, Life, Freedom” uprising emphasize that visual and symbolic acts—such as hair-cutting, hijab burning, and protest signage—play a central role in transforming individual grief into collective action. Izadi’s social semiotic analysis of protest signage details how images and slogans condense complex political demands into easily shareable motifs (Izadi & Dryden, 2024). Navarro’s (2023) study of the “Hair for Freedom” campaign and Kermani’s (2023) analysis of #MahsaAmini on Twitter show how hair-cutting and other gendered acts circulate as iconic protest images across platforms, linking on-the-ground risk to global visibility and transnational solidarity

 

Recent TikTok-based analysis shows that these symbolic repertoires also travel through platform-native formats such as Get Ready With Me (GRWM) videos, makeup-based bruising, haircutting gestures, and hijab-burning clips, creating a shared visual language across the “Woman, Life, Freedom” movement (Walsh, 2024).

 

These repertoires travel widely through remixing, translation, and cross-platform circulation. What remains less understood is how such multimodal strategies adapt to short-video architectures like TikTok, where expressive choices are shaped not only by political risk but also by platform-format norms.

 

Existing research on affective publics, connective action, and platformed visibility offers important insights into how protest communication circulates in networked environments, including under conditions of repression and algorithmic governance. This work explains how emotion, personalization, and platform curation shape participation and attention. However, it provides fewer tools for analyzing how collective expression operates when personalization is unsafe and when participation is organized primarily through reproducible multimodal structures rather than individual narratives. In short-video environments, where circulation is shaped less by textual markers and more by standardized audio-visual templates, the format itself becomes a central organizing unit of participation. This gap motivates the need for an analytical framework that foregrounds format as a mechanism of collective visibility.

 

Visibility Under Constraint

 

Bringing these dynamics together, I describe visibility under constraint as a communicative condition in which participation must be simultaneously politically safe and platform-recognizable. Rather than treating visibility as a straightforward outcome of circulation, this framework foregrounds how appearing online becomes a strategic process shaped by the convergence of state surveillance and platform-format norms.

 

Under this condition, participants navigate two interrelated regimes: The first is a risk regime of state visibility, in which public appearance can trigger surveillance, identification, or punishment. In this context, face-forward testimony, highly personalized expression, and identifiable affect carry heightened danger. Participants, therefore, gravitate toward symbolic, depersonalized, or low-exposure formats that reduce the likelihood of recognition.

 

The second is a platform regime of algorithmic visibility, in which circulation depends on alignment with platform-favored multimodal structures, including recognizable visual symbols, shared audio tracks, and standardized gestures. These elements fit TikTok’s curation logic and are more likely to be surfaced to broader audiences.

 

The empirical findings map onto this dual structure. Symbolic imagery is associated with greater reach, reflecting its platform legibility and reduced personal exposure, while protest audio predicts higher engagement by conveying collective presence and affect without requiring visual identification. Visibility under constraint thus clarifies why different multimodal features produce distinct outcomes depending on whether they prioritize safety or circulation.

 

In the Iranian context, this condition extends beyond TikTok’s internal recommendation logics. Protest videos are often produced with an awareness of transplatform circulation and transnational audiences, including diaspora networks and Western news media that routinely source and recontextualize short-form protest footage, a pattern also observed in TikTok-based conflict communication oriented toward international publics (Kalnes & Bjørge, 2025). Symbolic visuals and legible multimodal formats are therefore shaped not only by platform governance but also by their capacity to travel across linguistic, cultural, and geopolitical boundaries. Visibility under constraint thus operates across multiple layers of mediation, linking platform-specific affordances to broader transnational visibility regimes.

 

Format Publics

 

To explain how protest publics assemble under these conditions, I introduce the concept of format publics, which constitutes the central theoretical contribution of this article.

 

Format publics are collectivities that form through reproducible, low-exposure, and platform-legible multimodal formats—symbolic visuals, shared protest sounds, text overlays, and modular editing patterns. Unlike hashtag publics (Bruns & Burgess, 2015), which organize around textual markers, or imitation publics (Zulli & Zulli, 2020), which center on replication more broadly, format publics identify the unit of participation, the format, and specify its protective function under repression.

 

Format publics represent an adaptation of connective action. In connective action, personalized expression is the organizing unit. Under repression, personalization is unsafe. Instead, participants adopt format-based action frames: standardized audio-visual structures that encode affect, stance, and solidarity while minimizing personal exposure.

 

Through this lens, TikTok protest communication becomes intelligible not simply as user creativity, but as a strategic balancing act between risk management and platform visibility.

 

Existing research on TikTok’s role in the Mahsa Amini protests has primarily examined how protest visuals recontextualize established symbols across long-standing struggles over women’s rights (Walsh, 2024). My study extends this work by analyzing the multimodal repertoires that appear and how they relate to visibility outcomes—reach and engagement—under conditions of political constraint.

 

Format publics are thus a sociotechnical formation: They arise from the interplay between repression, user creativity, and platform architectures that privilege imitable and modular expression. The concept also travels beyond this specific case by offering a way to describe multimodal collective expression on short-video platforms in other repressive or high-risk contexts. Similar dynamics of digitally mediated repression have been documented in other contexts. Reporting on Myanmar, Potkin and Lone (2021) describe an “information combat” in which activists and authorities struggle over visibility, narrative control, and the circulation of protest content across social platforms.

 

Together, these dynamics allow us to anticipate how different multimodal elements should influence both reach and engagement on TikTok during contentious events.

 

Theoretical Expectations

 

This framework yields three expectations for short-video protest communication. These expectations are not predictive claims about TikTok’s algorithm, but analytical heuristics derived from how users navigate political risk and platform norms.

 

First, format alignment predicts reach. Content that adopts recognizable, imitable formats—mainly symbolic visuals—should achieve wider reach because it fits the platform’s recommended expressive grammar.

 

Second, authentic protest audio predicts engagement. Audio that conveys collective presence (chants, speeches, ambient sound) should drive higher interaction, as it signals immediacy and affective urgency.

 

Third, reach and engagement decouple under constraint. The features that maximize circulation are not the same as those that invite interaction, because participants must choose formats that balance safety, resonance, and platform legibility.

 

Research Questions

 

RQ1: What narrative styles and thematic repertoires characterize TikTok videos about the Mahsa Amini movement, and how are they associated with two visibility outcomes: reach and engagement?

 

RQ2: How do multimodal tactics—especially symbolic visuals and different types of audio—shape reach and engagement on TikTok under conditions of political constraint?

 

RQ3: How do hashtag practices (topical tags, generic tags, and hashtag count) relate to reach and engagement in these protest videos?

 

Methods

 

This study uses a mixed-methods design combining systematic content coding, statistical modeling, and qualitative video review.

 

Sampling and Data Collection

 

To examine how TikTok users communicated during the Mahsa Amini protests, I compiled a sampling frame of public TikTok videos posted between September 16 and October 31, 2022, that used bilingual protest-related hashtags in English and Persian (e.g., #MahsaAmini, #WomanLifeFreedom; #مهسا_امینی#, #زن_زندگی_آزادی). A search was conducted using Apify to collect publicly available metadata, including video URLs, basic statistics, captions, and hashtags.

 

After removing off-topic clips and exact or near duplicates, the eligible pool comprised 300 videos. From this pool, I drew a stratified random sample of 145 videos based on seed hashtag and week, using a fixed random seed. If a randomly selected video was unavailable at the time of coding, a replacement was drawn from the same stratum. The unit of analysis is the individual TikTok video.

 

Measures and Codebook

 

Videos were coded on narrative style, main theme, tone, audio elements, visual elements, and hashtag practices. The full codebook and decision rules appear in Appendix A (see Tables A1–A7). Platform-provided metrics (views, likes, comments, shares) were recorded for each video.

 

The coding instrument and analytical procedures follow established guidelines for systematic content analysis (Neuendorf, 2017), with categories derived from prior work on affective publics and digital activism and refined through pilot coding. Two dependent variables were derived:

 

• Reach: natural log of view count
• Engagement rate: (likes + comments + shares)/views

 

Coding and Analytical Strategy

 

The coding instrument was developed iteratively through pilot rounds. Coding followed a structured protocol informed by research on affective publics and digital activism, with all coding decisions documented for consistency. Quantitative analysis was conducted in Python using descriptive statistics, chi-square tests, and OLS regressions with HC3 robust standard errors. Given the modest sample size, regression estimates are interpreted as robust associations rather than precise effect sizes. All coding was conducted by a single coder, with coding decisions documented across multiple rounds for consistency.

 

To assess intracoder reliability, I independently re-coded a random subsample of 40 videos (28% of the data set) two weeks after the initial coding. Reliability was assessed using Cohen’s κ, which is appropriate for nominal variables commonly used in content analysis (Neuendorf, 2017). Across variables, agreement was substantial to almost perfect: Narrative Style (κ = .74), Main Theme (κ = .97), Tone (κ = .95), Audio Elements (κ = 1.00), and Visual Code (κ = .86). These values meet accepted standards for coder stability and indicate that the coding instrument and decision rules were applied consistently.

 

The qualitative component involved reviewing a subsample of videos to identify recurrent multimodal strategies. Particular attention was given to how symbolic imagery and protest audio jointly frame emotion and mobilization. Representative examples are presented in the Findings section.

 

Ethical Considerations

 

All analyzed content was publicly accessible at the time of collection. Identifying details were removed or blurred in reproduced screenshots. No full videos or usernames are shared. These practices follow established guidelines for Internet-mediated research (British Psychological Society, 2021; Townsend & Wallace, 2016).

 

Findings

 

To address the three research questions, I present the results across four empirical areas that reflect the main analytical dimensions of the study: (1) narrative and thematic patterns (RQ1), (2) visual formats and reach (RQ2), (3) audio cues and engagement (RQ2), and (4) hashtag practices (RQ3). Together, these findings show how users navigated visibility under political constraint through distinct multimodal strategies.

 

Narrative and Thematic Patterns

 

Analysis of the 145 videos shows that content production during the Mahsa Amini protests relied mainly on two narrative styles. Reporting was the most common (45.5%), followed by artistic/expressive formats (31.7%). First-person testimony and mobilization appeals were used far less often. See Figure 1 for the distribution of narrative styles.

 

A chi-square test confirmed that narrative styles were not evenly distributed (χ²(4) = 93.24, p < .001).

 

Figure 1. Distribution of narrative styles in TikTok protest videos (N = 145).

 

The thematic patterns reflect a similar trend. News/Update (35.9%) and Call for Solidarity (33.1%) were the most frequent themes, followed by International Outreach (20.7%), as shown in Figure 2.

 

Among the most-viewed videos, International Outreach and Call for Solidarity appeared more often than News/Update, and theme distribution differed significantly between high-visibility clips and the rest of the sample (χ²(5) = 13.21, p = .022; see Figure 3).

 

Figure 2. Distribution of main themes (overall).

 

Figure 3. Theme prevalence among most-viewed videos vs. the rest.

 

Qualitative observations show that reporting videos often combine brief factual captions with protest footage. At the same time, artistic/expressive clips relied on symbolic gestures or montage editing that, in some cases, avoided showing identifiable faces. Testimony was rare and usually anonymized. These patterns suggest that creators used content styles that enabled them to convey protest messages without risking direct surveillance.

 

Visual Formats and Reach

 

The regression model predicting log-transformed views (R² = .21) indicates that symbolic imagery is the strongest predictor of reach (β = 1.58, p = .004). Other visual categories, including protest-scene footage and selfie-style videos, do not significantly predict views (Table 1).

 

Table 1. OLS Regressions of Audio, Visual, and Tone Predictors on Log Views and Engagement Rate (N = 145).

Predictor	b (Log views)	p (Log views)	b (Engagement rate)	p (Engagement)
Symbolic Imagery	1.58	.004	1.55	.36
Original Speech/ Ambient	0.95	.018	3.57	.017
Protest Audio	–0.34	.59	4.18	.032
Voice-Over/ News/ Mixed	0.30	.69	8.60	.10
Protest Scene	0.49	.42	0.10	.97
Selfie Video	–0.61	.20	3.95	.16
Tone (Low/ Neutral/ Reflective)	–0.00	.99	–0.13	.93

 

Qualitative examples help explain why symbolic visuals travel more widely. Videos showing hair-cutting gestures, the burning of a scarf, or makeup-based bruising rely on close-up framing that conveys clear political meaning without revealing identifiable faces. These formats recur across many clips in the data set, forming a recognizable visual repertoire that is relatively safe to reproduce and aligned with TikTok’s short-video grammar (see Figures 4–6).

 

Overall, symbolic visuals functioned as low-exposure, platform-aligned formats that enabled wider circulation during the protests, in contrast with more personalized or face-forward visual styles (see Figures 4–6).

 

Figure 4. Symbolic hair-cutting gestures. A close-up showing a lock of hair being cut as a protest symbol (personal communication, September 19, 2022).

 

Figure 5. Burning of a scarf as a protest symbol. This symbolic act appeared in multiple stylistic variations across the data set, functioning as a reproducible visual format that conveyed dissent (personal communication, September 20, 2022).

 

Figure 6. Makeup-based bruising as representational imagery. A close-up of makeup simulating bruising, used to represent state violence. This format reflects expressive symbolism (personal communication, October 11, 2022).

 

Audio Cues and Engagement

 

The model predicting engagement rate (R² = .18) shows that original speech or ambient sound (β = 3.57, p = .017) and protest audio (β = 4.18, p = .032) are significant predictors of higher engagement (Appendix B, Figure B1). In contrast, symbolic imagery—which was strongly associated with reach—does not significantly predict engagement (Appendix B, Figure B2).

 

Qualitatively, many clips recorded from balconies, windows, or behind crowds capture collective chants, rhythmic slogans, or ambient protest noise. Even when the visuals are distant or static, these sound cues seem to invite viewers to comment or share and often generate dense interaction relative to the video’s view count (see Figure 7).

 

In addition to protest chants and ambient sound, two songs appeared across the sample: Another Love (Odell, 2012) and Unstoppable (Sia, 2016). In this context, these tracks serve primarily as stylistic or emotional framing devices rather than direct indicators of on-the-ground protest activity. They do not exhibit statistically significant associations with reach or engagement in the models, but they form part of the broader sonic repertoire through which creators signal solidarity or mood without revealing personal identity.

 

Taken together, these patterns suggest that while symbolic visuals help videos circulate widely, audio that conveys collective presence and ambient protest activity plays a central role in generating interaction and affective resonance.

 

Figure 7. Distant protest scene emphasizing collective audio. A screenshot from a protest video. Chants and collective shouting dominate the audio (personal communication, October 10, 2022).

 

Hashtag Practices and Their Limited Role

 

Hashtag practices showed only modest associations with visibility outcomes. Videos used an average of six hashtags, and roughly one-quarter included generic visibility tags such as #fyp or #viral. Hashtag count showed no meaningful relationship with reach: Both a simple correlation (ρ = .02, p = .79) and an OLS model indicated no significant association with log views (b = 0.02, p = .712). In contrast, hashtag count displayed a small, but statistically significant relationship with engagement rate (b = 0.27, p = .013). The presence of generic hashtags did not predict either reach or engagement (Table 2).

 

A qualitative review indicates that hashtags were used primarily to mark solidarity and to situate videos within the broader protest vocabulary rather than to optimize visibility. Most captions repeated well-known protest tags, and few clips used hashtags strategically and experimentally. This is consistent with TikTok’s short-video architecture, where audio and visual formats carry more influence over circulation than textual metadata. Taken together, these patterns indicate that hashtag practices play a comparatively limited role in shaping visibility during the protests.

 

Overall, hashtag practices accounted for a comparatively limited share of visibility compared with multimodal formats. These patterns are summarized in Table 2, with additional descriptive plots in Appendix B (Figures B3–B5).

 

Table 2. OLS Regressions of Hashtagging Practices on Log Views and Engagement Rate (N = 145).

Predictor	B (SE)	t	P	B (SE)	t	P
Constant	11.66 (0.45)	25.9	<.001	8.58 (0.52)	16.5	<.001
Hashtag Count	0.02 (0.05)	0.37	.712	0.27 (0.10)	2.52	.013
Generic Hashtag Dummy	–0.15 (0.44)	–0.45	.744	–0.13 (0.14)	–0.09	.926

Note. B = unstandardized regression coefficient; SE = standard error; t = t statistic; p = significance value; N = 145. Columns 2–4 represent Log Views; columns 5–7 represent Engagement Rate.

 

Summary Interpretation

 

Across all four empirical areas, the results show that protesters used multimodal strategies that balanced political risk with the expressive norms favored by TikTok. Symbolic visuals supported reach, protest audio increased engagement, and hashtags played only a limited role. These patterns align with the theoretical concept of visibility under constraint and demonstrate how protest publics assembled through format publics—reproducible, low-exposure expressive units that allow individuals to participate collectively while managing exposure to both state surveillance and platform curation.

 

Discussion

 

The findings of this study show how protesters navigated TikTok’s visibility structures amid political repression. Instead of relying on face-forward narration or personalized storytelling, users adopted multimodal strategies that balanced the risks of being seen with the platform’s expressive norms. These strategies reveal how protest communication adapts when visibility must be both politically safe and algorithmically recognizable.

 

Symbolic visuals were the strongest drivers of reach. Acts such as hair cutting, scarf burning, or makeup simulating bruising served as highly legible units of expression that circulated widely. These visuals condensed political meaning into formats that fit TikTok’s short-video grammar and avoided the personal risks of appearing on camera. In this sense, symbolic imagery did not simply “represent” dissent; it provided a safe and scalable format through which dissent could travel.

 

Audio functioned differently. Protest chants, ambient street noise, and recorded speech predicted higher engagement even when these clips did not reach large audiences. Authentic soundscapes conveyed immediacy and collective presence, inviting viewers to respond in the comments or share. These patterns indicate that under constraint, affective cues shift from individual expression to collective sound, distributing emotion through multimodal traces rather than through identifiable bodies.

 

Together, these results clarify how visibility under constraint operates as a layered process shaped by two forces: the threat of state visibility and the demands of platform visibility. Symbolic visuals help content circulate safely, while authentic audio fosters interaction without requiring identity disclosure. The decoupling of reach and engagement reflects the different creative choices users must make when aiming for visibility without exposure.

 

These dynamics also help define the concept of format publics. Unlike affective publics, which rely on distributed emotion, or connective action, which relies on personalized frames, format publics assemble through standardized multimodal structures that minimize personal risk. These structures provide a shared repertoire of gestures, sounds, and visual templates that allow participants to signal alignment without revealing themselves. Format publics, therefore, describe a type of collective expression that emerges when personalization is dangerous, and platform architectures reward repeatable, template-like forms.

 

The findings also show that hashtag practices play a limited role in shaping visibility. Hashtags mainly served symbolic and organizational purposes rather than functioning as mechanisms for discovery. This aligns with the platform’s design, where audio and visual formats outweigh textual metadata in determining circulation.

 

Overall, the results demonstrate how protest communication on TikTok adapts to the combined pressures of political repression and platform governance. Under constraint, visibility becomes an outcome of multimodal formatting rather than personal narration. Symbolic visuals and collective sound become the primary carriers of affect and mobilization, enabling publics to appear collectively while reducing individual exposure. Consequently, this study contributes a concept that explains how format-driven protest communication operates within short-video environments and suggests new avenues for examining digital activism under repressive conditions.

 

Conclusion

 

This study examined how TikTok was used for protest communication during the early months of the Mahsa Amini movement in Iran. By analyzing 145 videos, the study showed that protesters relied on multimodal formats that balanced political risk with the platform’s expressive norms. Symbolic visuals were strongly associated with higher reach, while authentic speech, ambient sound, and protest chants generated higher engagement. Hashtag practices played only a limited role in shaping visibility. These patterns indicate that visibility was achieved not through personal narration or face-forward performance, but through reproducible multimodal structures that allowed users to appear without exposing themselves to direct surveillance.

 

The concepts of visibility under constraint and format publics help explain these dynamics. Visibility under constraint describes the layered environment in which users must remain safe while still becoming visible to the platform’s recommendation system. Format publics capture how collective expression forms through symbolic gestures, shared sounds, and modular editing patterns rather than through personalized stories. In this concept, formats are not decorative, but functional—they allow individuals to participate in protest communication while managing their exposure to both state power and platform governance.

 

Although this study focuses on a single protest wave and a single platform, its insights raise broader questions about digital activism in repressive contexts. As short-video platforms continue to shape how political communication unfolds, understanding how users navigate their multimodal affordances becomes increasingly important.

 

Future research could examine whether similar symbolic and sonic strategies appear in other movements, or how evolving platform algorithms change the formats available for safe expression.

 

Overall, the findings show that protest publics on TikTok assemble through strategies that distribute emotion, solidarity, and political meaning across reproducible audio-visual formats. Under repressive conditions, these formats become the primary carriers of visibility, enabling collective expression while limiting individual risk. Thus, this study offers a theoretical and empirical framework for understanding how digital activism adapts to the combined pressures of political constraint and platform-format logics.

 

Practical Implications

 

These findings have several implications for scholars and practitioners who study or work with digital activism under repressive conditions. First, the strong performance of symbolic visuals suggests that reproducible, low-exposure formats can help movements communicate effectively while limiting personal risk. This highlights the importance of understanding how platform formats shape the visibility of political expression. Second, the influence of authentic protest audio on engagement indicates that sound plays a meaningful role in generating responsiveness. For researchers and open-source investigators, this underscores the value of preserving audio-rich content during crises. Finally, the limited role of hashtags suggests that strategies relying heavily on textual metadata may be less effective on short-video platforms, where multimodal templates shape circulation more strongly than tagging practices.

 

Future Research Directions

 

This study opens several avenues for future research on the interplay between political repression, multimodal expression, and platform logics. One direction is to examine how protest formats evolve over more extended periods and whether symbolic and sonic strategies shift as platform algorithms or political conditions change. Comparative work across movements or countries could also clarify how format publics emerge in different cultural or geopolitical settings.

 

In addition, research on other short-video platforms may reveal how different recommendation systems influence the expressive choices available to activists. Together, these directions can help build a broader understanding of how multimodal protest communication adapts under constraint.

 

Limitations

 

This study has several limitations that should be acknowledged. The analysis is based on a sample of public TikTok videos collected via protest-related hashtags, so content circulated through private sharing or alternative discovery pathways is not captured. The study also infers platform influence from observable patterns rather than direct access to TikTok’s algorithmic processes. While this approach is appropriate for studying visibility under constraint, it cannot identify causal mechanisms inside the recommendation system. Finally, the coding was conducted by a single researcher, which may limit the assessment of intercoder reliability, although coding decisions were documented systematically and reviewed across multiple rounds. These limitations do not undermine the findings but indicate areas where future research could build on and extend this work.

 

 

References

 

Abidin, C. (2021). Mapping internet celebrity on TikTok: Exploring attention economies and visibility labours. Cultural Science, 12(1), 77–103. https://doi.org/10.5334/csci.140

 

Bennett, W. L., & Segerberg, A. (2013). The logic of connective action: Digital media and the personalization of contentious politics. Cambridge, UK: Cambridge University Press.

 

Bishop, S. (2019). Managing visibility on YouTube through algorithmic gossip. New Media & Society, 21(11–12), 2589–2606. https://doi.org/10.1177/1461444819854731

 

boyd, d. (2010). Social network sites as networked publics: Affordances, dynamics, and implications. In Z. Papacharissi (Ed.), A networked self: Identity, community, and culture on social network sites (pp. 39–58). New York, NY: Routledge.

 

Brighenti, A. (2007). Visibility: A category for the social sciences. Current Sociology, 55(3), 323–342. https://doi.org/10.1177/0011392107076079

 

British Psychological Society. (2021). Ethics guidelines for internet-mediated research. https://www.bps.org.uk/guideline/ethics-guidelines-Internet-mediated-research

 

Bruns, A., & Burgess, J. (2015). Twitter hashtags from ad hoc to calculated publics. In N. Rambukkana (Ed.), Hashtag publics: The power and politics of discursive networks (pp. 13–28). New York, NY: Peter Lang.

 

Bucher, T. (2018). If . . . then: Algorithmic power and politics. Oxford, UK: Oxford University Press.

 

Castillo-Esparcia, A., Caro-Castaño, L., & Almansa-Martínez, A. (2023). Evolution of digital activism on social media: Opportunities and challenges. Profesional de la Información, 32(3), 1–16. https://doi.org/10.3145/epi.2023.may.03

 

Cervi, L., Tejedor, S., & Marín Lladó, C. (2021). TikTok y el nuevo lenguaje de la comunicación política [TikTok and the new language of political communication: The case of Podemos]. Cultura, Lenguaje y Representación, 26, 267–287. https://doi.org/10.6035/clr.5817

 

Cunningham, S., & Craig, D. (2019). Creator governance in social media entertainment. Social Media + Society, 5(4), 1–11. https://doi.org/10.1177/2056305119883428

 

Danesh, A., & Athari, S. H. (2024). Cyber activism in Iran: A case study. Social Media + Society, 10(3), 1–12. https://doi.org/10.1177/20563051241279258

 

Duffy, B. E., & Meisner, C. (2023). Platform governance at the margins: Social media creators’ experiences with algorithmic (in)visibility. Media, Culture & Society 45(2), 285–304. https://doi.org/10.1177/01634437221111923

 

Faris, D. M., & Rahimi, B. (Eds.). (2015). Social media in Iran: Politics and society after 2009. Albany, NY: SUNY Press.

 

Freelon, D., McIlwain, C. D., & Clark, M. D. (2016). Beyond the hashtags: #Ferguson, #Blacklivesmatter, and the online struggle for offline justice. Washington, DC: Center for Media & Social Impact. https://doi.org/10.2139/ssrn.2747066

 

Gillespie, T. (2018). Custodians of the Internet: Platforms, content moderation, and the hidden decisions that shape social media. New Haven, CT: Yale University Press.

 

Gitomer, A., Atienza-Barthelemy, J., & Foucault Welles, B. (2024). Youth activism and (de)personalized remix on TikTok. Social Movement Studies, 24(2), 215–233. https://doi.org/10.1080/14742837.2024.2415672

Izadi, D., & Dryden, S. (2024). Woman/Life/Freedom: The social semiotics behind the 2022 Iranian protest movement. Discourse, Context & Media, 60, 1–11. https://doi.org/10.1016/j.dcm.2024.100803

 

Kalnes, Ø., & Bjørge, N. M. (2025). “So, we have occupied TikTok”: Ukrainian women in #ParticipativeWar. Media, War & Conflict, 18(2), 233–250. https://doi.org/10.1177/17506352241307010

 

Kaye, D. B. V., Zeng, J., & Wikström, P. (2022). TikTok: Creativity and culture in short video. Cambridge, UK: Polity Press.

 

Kermani, H. (2023). #MahsaAmini: Iranian Twitter activism in times of computational propaganda. Social Movement Studies, 24(2), 257–267. https://doi.org/10.1080/14742837.2023.2180354

 

Kreutz, J., & Makrogianni, A. A. (2024). Online repression and transnational social movements: Thailand and the #MilkTeaAlliance. Political Research Exchange, 6(1), 1–20. https://doi.org/10.1080/2474736X.2023.2299120

 

Lee, J., Abidin, C., & Duguay, S. (2023). Introduction to the special issue of “TikTok and social movements.” Social Media + Society, 9(1), 1–8. https://doi.org/10.1177/20563051231157452

 

Navarro, C. (2023). “Hair for freedom” movement in Iran: Interreligious dialogue, digital feminism, and social media activism. Religions, 14(5), 1–13. https://doi.org/10.3390/rel14050602

 

Neuendorf, K. A. (2017). The content analysis guidebook (2nd ed.). Thousand Oaks, CA: SAGE Publications. https://doi.org/10.4135/9781071802878

 

Odell, T. (2012). Another love [Song]. On Long way down [Album]. New York, NY: Columbia Records.

 

Papacharissi, Z. (2014). Affective publics: Sentiment, technology, and politics. New York, NY: Oxford University Press.

 

Potkin, F., & Lone, W. (2021, November 1). “Information combat”: Inside the fight for Myanmar’s soul. Reuters. https://www.reuters.com/world/asia-pacific/information-combat-inside-fight-myanmars-soul-2021-11-01/

 

Rahimi, B. (2011). The agonistic social media: Cyberspace in the formation of dissent and consolidation of state power in postelection Iran. The Communication Review, 14(3), 158–178. https://doi.org/10.1080/10714421.2011.597240

 

Sia. (2016). Unstoppable [Song]. On This is acting [Album]. New York, NY: RCA Records.

 

Sreberny, A., & Khiabany, G. (2010). Blogistan: The internet and politics in Iran. London, UK: I.B. Tauris.

Townsend, L., & Wallace, C. (2016). Social media research: A guide to ethics. Aberdeen, UK: University of Aberdeen.

 

van Dijck, J., & Poell, T. (2016). Social media and the transformation of public space. Social Media + Society, 2(3), 1–5. https://doi.org/10.1177/2056305115622482

 

Walsh, T. (2024). TikTok as a site of social protest in Iran’s Gen-Z uprising: Visuality, social media, and the diaspora following the death of Mahsa Amini. Discourse & Society, 35(5), 625–650. https://doi.org/10.1177/09579265241234351

 

Zulli, D. J., & Zulli, D. (2020). Extending the internet meme: Conceptualizing technological mimesis and imitation publics on the TikTok platform. New Media & Society, 24(8), 1872–1891. https://doi.org/10.1177/1461444820983603

 

 

Appendix A. Codebook

 

Table A1. Narrative Style.

Category	Conceptual Definition	Operational Criteria/Decision Rules	Typical Indicators/Examples
Reporting	Informational or news-like storytelling documenting events. Grounded in affective reporting (Papacharissi, 2014).	Used when a video summarizes events, shows on-screen captions with facts, or uses reporter-style narration.	“Protests erupted in Tehran tonight . . .”; text overlays with times & locations (personal communication, September 25, 2022).
First-person Testimony	Personal narration or lived experience. Reflects connective action frames (Bennett & Segerberg, 2013).	Speaker uses “I” or “we”; selfie framing; recounts personal experience or emotion.	“I saw the police beating . . .”
Artistic/Expressive	Creative or symbolic expression through art, montage, or performance rather than direct speech.	Select when artistic editing, dancing, lip-sync, or visual metaphor dominate > 60 % of the video.	Hair-cutting montage set to Another Love (Odell, 2012); symbolic slow-motion gestures.
Call for Support/Mobilization	Explicit appeal for viewers to act or share; linked to connective mobilization.	Presence of imperative verbs (“share,” “join,” “spread”), or a caption urging participation.	“Please repost this to support Iranian women.”
Other/Hybrid	Mixed or rare formats combining multiple of the above.	Use only when no single mode dominates.	Interview, stitched duets combining news + music.

 

Table A2. Main Theme.

Category	Definition	Inclusion Cues	Example
News/Update	Announces protest developments or information.	Mentions dates, locations, and numbers.	“Internet is down in Shiraz . . .” (personal communication, October 7, 2022).
Call for Solidarity	Appeals for unity or collective emotion.	“We stand together,” collective chanting.	Montage with “Woman Life Freedom.”
International Outreach	Addresses diaspora or global audience.	English captions; #IranProtests.	“Dear world, hear our voice.”
Eyewitness Documentation	Raw on-scene video recorded by participant.	Minimal editing, ambient protest noise.	Street-level footage of chants.
Memorial/Remembrance	Tribute to victims or martyrs.	Portraits, candlelight, slow music.	Video honoring Mahsa Amini.

 

Table A3. Tone.

Category	Definition	Observable Cues	Example
High-Arousal/Emotive	Expresses strong emotion: anger, hope, defiance, celebration.	Loud voice, chanting, rapid cuts, bright colors, upbeat music.	Shouting “Zan Zendegi Azadi” with energetic edits.
Low/Neutral/Reflective	Calm, factual, or mournful; restrained affect.	Slow pacing, somber music, grayscale filter, tears, candlelight.	Hair-cutting montage with slow piano.

 

Table A4. Audio Elements.

Category	Definition	Indicators	Example
Music	Pre-existing or protest song dominates; no spoken narration.	Recognizable track underlies visuals.	Use of Sia’s “Unstoppable” as soundtrack (Sia, 2016)
Original Speech/Ambient	Recorded at the filming site.	Street noise, voices, chants.	Interview clip with background protest.
Protest Audio	Collective chants or slogans.	“Woman Life Freedom,” rhythmic slogans.	Crowd chanting sequence.
Voice-over/ News/Mixed	Edited narration or broadcast fragment.	External voice-over, news anchor voice.	TV audio layered on protest images.

 

Table A5. Visual Elements.

Category	Definition	Inclusion Cues	Example
Symbolic Imagery	Politically charged acts or visuals represent resistance.	Hair-cutting, scarf   \burning, color symbolism, and makeup representing bruises	A young woman burns a scarf.
Selfie Video	Creator addresses the camera directly.	Eye-level framing, direct gaze	“I want the world to know . . .” (personal communication, September 25, 2022).
Protest Scene	Collective protest footage.	Rallies, placards, crowds, chanting	Street march footage.
Other/Media Reuse	News footage, text slides, memes.	Screenshots, subtitles	Compilation of international media clips.

 

Table A6. Hashtagging Practices.

Variable	Definition	Operationalization	Example
Hashtag Count	Number of hashtags per caption.	Count each term preceded by “#”.	“#MahsaAmini #IranProtest #FYP” = 3.
Generic Hashtag Presence	Presence (1)/ absence (0) of generic visibility tags.	#fyp, #viral, #trending, #explore, #foryou, #foryoupage, #duet.	“#fyp” → 1.

 

Table A7. Visibility Metrics.

Variable	Definition	Formula
Views	total number of video views in the collection.	—
Likes, Comments, Shares	Interaction counts.	—
Engagement Rate	Relative interaction intensity.	(Likes + Comments + Shares) ÷ Views
Log Views	Natural log of (Views + 1) for scale normalization.	ln (Views + 1)

 

Appendix B—Statistical Figures

 

Figure B1. Mean log views by audio elements.

 

Figure B2. Mean log views by visual code.

 

Figure B3. Distribution of hashtag count.

 

Figure B4. Log views and engagement rates by the presence of generic hashtags.

 

Figure B5. Scatterplots of hashtag count with log views and engagement rate.

 

 

Copyright © 2026 (Fatemeh Oudlajani). Licensed under the Creative Commons Attribution Non-commercial No Derivatives (by-nc-nd). Available at https://ijoc.org.

https://doi.org/10.65476/nt5fcf85

 

[1] This research received no external funding. The author declares no conflict of interest. This study was exempt from IRB review as it relies solely on publicly available TikTok content.