From Private Trouble to Collective Concern: A Corpus-Based Analysis of Intimate Partner Violence in China News Media

Abstract

Intimate partner violence (IPV) remains understudied in China despite its public health significance. Previous research lacks comprehensive analysis of how Chinese media frames this issue, creating a gap in understanding the sociocultural factors shaping public discourse. This study employs corpus-based framing analysis of 603 news articles (435,581 words) from major Chinese newspapers spanning 2012–2022, a period encompassing significant legal developments including the 2016 Domestic Violence Law. We analyze how IPV is framed through examination of keyword frequencies, collocation patterns, and concordance analysis. Our findings reveal that IPV is predominantly framed as matrimonial conflict and family dispute rather than criminal violence requiring state intervention. We argue that framing IPV as a ‘family issue’ operates as a spatial containment strategy, relocating violence to the domestic sphere while rerouting intervention into administrative/civil channels rather than criminal accountability spaces. Our findings reveal significant imbalances in stakeholder representation, with government and legal voices dominating the public discourse domain while community support organizations are marginalized. Source attribution patterns produce uneven zones of legitimacy, where state actors occupy authorized public space while survivors’ experiences remain confined to private, silenced domains. This research enhances the understanding of IPV media coverage in China while highlighting the need for more inclusive public discourse.

1. Introduction

Intimate partner violence (IPV) represents a significant global health concern, with the World Health Organization reporting that approximately 40.3% of women worldwide experience IPV during their lifetime (WHO 2021). This alarming prevalence has prompted extensive research across diverse cultural contexts. At the same time, how IPV is publicly recognized and institutionally handled varies substantially across sociocultural and media environments. Within this global landscape, China presents a particularly important case for investigation, as public awareness of IPV remains limited (Tam et al. 2016) and institutional recognition inadequate (Yuan and Hesketh 2021). Despite China’s 2016 Anti-Domestic Violence Law recognizing IPV as a public concern requiring multi-institutional state intervention and establishing civil protection mechanisms, Chinese news media may continue to frame IPV as a private family matter requiring informal resolution. While the 2016 law does not create a distinct criminal offense of domestic violence, it mandates coordinated response from police, courts, and administrative agencies, shifting IPV from purely private to public institutional terrain. This study investigates whether media framing has adapted to this legal repositioning of IPV as requiring formal state response.

This study investigates this potential disjuncture through large-scale corpus analysis, examining whether and how media framing reinforces spatial containment of IPV within domestic and administrative domains rather than criminal justice space. By detecting systematic patterns across different newspapers over 12 years (2012–2022), we reveal how media representation may lag behind or resist legal frameworks positioning IPV as a public crime.

Research indicates that only 51.2% of the Chinese population recognized IPV as a significant societal issue as recently as 2016 (Tam et al. 2016). This statistic is particularly troubling when contrasted with IPV’s widespread occurrence in Chinese communities. This limited awareness extends beyond the general public to institutional frameworks, particularly in rural areas where law enforcement often demonstrates inadequate understanding of IPV dynamics and severity (Yuan and Hesketh 2021), reflecting how cultural and institutional factors shape responses to IPV.

The consequences of this insufficient awareness are multifaceted and interconnected. First, it contributes to systematic underreporting and inadequate intervention, perpetuating cycles of violence that might otherwise be interrupted through appropriate institutional responses (Yuan and Hesketh 2021). Second, limited understanding reinforces social stigma, potentially intensifying the isolation experienced by survivors who may feel their experiences are delegitimized or misunderstood. Third, these combined factors create significant barriers to effective prevention strategies and support systems. Recognizing these complex dynamics, public health researchers have emphasized the urgent need for enhanced IPV awareness within China’s sociocultural context (Tiwari et al. 2018).

In this environment, mass media plays a crucial role in shaping public discourse about IPV in China (Cao et al. 2023). Media representations not only reflect existing social attitudes but also construct frameworks through which the public interprets and responds to IPV. The research problem, therefore, is not “China as an underrepresented context” in general, but the discursive mechanisms through which IPV becomes publicly legible (or remains domesticated and privatized) in Chinese news coverage.

China’s media landscape operates under a distinctive institutional arrangement. News organizations function within a framework of state ownership and editorial coordination while also responding to market dynamics and audience preferences (Stockmann 2013). This dual-characteristic system differs from Western media models and creates a unique environment for examining how social issues like IPV are represented in public discourse. Accordingly, we examine how Chinese newspapers recurrently frame IPV and how they allocate discursive authority across actors (who is mentioned vs. who is quoted). The hybrid media system in China presents complex dynamics, where journalists navigate both political expectations and commercial pressures, often prioritizing sensational coverage to attract audiences, rather than focusing on public interest reporting that could better serve social awareness (Fenton 2011).

This unique media environment makes China an especially important site for investigating how IPV is represented in public discourse. By examining Chinese media coverage of IPV, this study contributes to broader scholarly understanding of how cultural, political, and economic factors interact to shape social issue communication in non-Western contexts. More specifically, our corpus-based design enables us to identify recurring patterns across multiple outlets and over time, rather than relying on single-outlet or small-sample evidence.

2. Literature Review

2.1. Media Framing and IPV

To understand how IPV is represented in Chinese media, this study draws on media framing theory, which provides a robust analytical framework for examining how social issues are communicated. Media framing theory establishes that how issues are presented in media fundamentally shapes public understanding and response to social phenomena (Scheufele 1999). By selecting certain aspects of perceived reality and making them more salient, frames promote particular problem definitions, causal interpretations, moral evaluations, and treatment recommendations (Entman 1993). This process is neither neutral nor arbitrary but reflects broader power structures and cultural values.

Contemporary scholarship has identified two critical dimensions of framing particularly relevant to IPV coverage: issue construction (how the problem is defined, causally attributed, morally evaluated, and resolved, following Entman’s (1993) four framing functions) and source attribution (whose voices are amplified and granted authority). Complementing these framing dimensions, discourse-analytical approaches examine the linguistic and narrative structures through which these frames are enacted, including vocabulary choices, grammatical patterns of actor representation, and narrative sequencing (Lloyd and Ramon 2017; van Leeuwen 2008). The present study uses framing analysis to identify dominant problem frames and discourse analysis to examine how these frames are linguistically constructed and how authority is distributed across actor groups. These dimensions provide useful analytical categories for examining how Chinese media constructs IPV narratives within its specific cultural context.

Recent scholarship has extended framing analysis to consider how media discourse produces and manages space as well as meaning. Drawing on Lefebvre’s (1991) foundational insight that space is socially produced rather than merely given, media scholars have examined how news coverage locates social problems within particular spatial domains—private/domestic, public/civic, or state/institutional—with significant consequences for accountability and intervention (Rodgers et al. 2014; Couldry and McCarthy 2004). This spatial dimension of framing is particularly salient for IPV, where the boundary between ‘private’ household space and ‘public’ civic space has historically determined whether violence is treated as a personal matter or a collective concern requiring formal intervention (Pain 2014). Accordingly, this study examines not only what frames dominate Chinese IPV coverage, but also where violence, responsibility, and legitimate response are discursively positioned.

Framing serves as an essential mechanism in journalism, functioning as a central organizing principle that guides event interpretation and narrative construction (Nelson and Clawson 1997). When covering complex social issues like IPV, journalists necessarily select certain aspects to emphasize while downplaying others, creating interpretive packages that help audiences make sense of unfamiliar or uncomfortable topics. This framework directs argumentative dimensions while highlighting core issues within public discourse (Gamson and Modigliani 1989). For IPV coverage specifically, framing choices can significantly impact whether violence is understood as an individual problem or a systemic social issue requiring collective response.

Building on these theoretical foundations, scholars have investigated media representations of IPV across various cultural contexts, revealing several critical patterns.

First, research has demonstrated how media coverage significantly shapes both public perception and policy development. Lloyd and Ramon’s (2017) longitudinal discourse analysis of domestic violence in two UK national daily newspapers revealed notable variations in how victims were portrayed across publications, with some reinforcing victim-blaming narratives.

Second, research has identified mechanisms through which perpetrator responsibility becomes systematically minimized. Isaacs and Mthembu (2018) revealed three predominant frameworks: pathological individualization (framing violence as aberrant individual behavior), justice system focus (emphasizing procedural aspects over human impact), and emotional appeals that may oversimplify complex dynamics. These patterns demonstrate how media framing can subtly reinforce problematic perspectives even while ostensibly treating IPV as a serious issue.

Third, digital platforms have introduced new dimensions to IPV discourse. Ebadijalal and Weisi’s (2022) study of Iranian social media identified six distinctive thematic frameworks and significant gender-based differences in perception, highlighting how discourse patterns evolve across media platforms.

Collectively, these studies demonstrate how media representations of IPV reflect and reinforce broader cultural values, institutional structures, and power dynamics. Examining Chinese media representations requires careful attention to these contextual factors while drawing on established methodological approaches.

2.2. Source Attribution in IPV Coverage

Complementing framing analysis, source attribution analysis represents a crucial dimension in understanding IPV media representation (Bullock and Cubert 2002; Seely and Riffe 2021). Examining who speaks and who is spoken about in media coverage reveals power hierarchies that significantly impact public understanding of social issues. Cross (2010) defines sources as narrative contributors providing informational context, diverse perspectives, and eyewitness accounts. For IPV coverage specifically, source selection decisions significantly influence whether survivors’ experiences are centered or marginalized in public discourse.

Empirical studies have revealed systematic patterns of institutional dominance in IPV coverage. Research consistently demonstrates how official sources—police, courts, and legal experts—control media narratives. Bullock and Cubert’s (2002) analysis demonstrated a clear hierarchical structure in source utilization, with factual statements (51.4%), police sources (16.7%), and court documentation (5.4%) emerging as predominant information sources. This reliance on institutional sources often results in narratives that privilege legal perspectives over survivor experiences and broader social contexts, reflecting journalistic tendencies to prioritize sources perceived as objective and authoritative (Gans 2004; Vasterman and Ruigrok 2013). Such patterns potentially marginalize more diverse perspectives that could provide richer understanding of IPV’s social dimensions.

Research also reveals a critical paradox in victim representation: while victims are frequently discussed in coverage, they rarely have opportunity to speak for themselves. Seely and Riffe’s (2021) examination of newspaper coverage revealed significant disparities in voice representation, with victim and family voices comprising only 1.6% and 2.1% of citations respectively, while police sources dominated coverage. This imbalance frequently resulted in fact-centric reporting that neglected broader sociological implications and survivor perspectives. The disconnect between being discussed and having voice demonstrates how media coverage can simultaneously center and silence certain groups, creating visibility without agency. This pattern raises important questions about whose knowledge is valued in public discourse about IPV and how these valuations shape public understanding.

Collectively, these studies establish source attribution as a critical mechanism through which power operates in media discourse about IPV. By examining who speaks and who is spoken about in Chinese media coverage, this study contributes to understanding how authority is constructed in Chinese public discourse about sensitive social issues within its specific cultural and institutional context.

2.3. Research Gaps and Current Study

While existing scholarship has substantially advanced our understanding of media framing in IPV coverage, several critical gaps remain that this study addresses. First, the Chinese context remains severely underrepresented in English-language scholarship on IPV media representation (Berns 2004; Bullock 2007; Carlyle et al. 2008; Donovan and Hester 2015; Meyers 1996; Simons and Morgan 2017). Given China’s unique sociocultural dynamics and distinctive media environment, this gap limits our understanding of how cultural and institutional factors influence IPV discourse across diverse contexts.

Second, existing studies often focus on small-scale samples or single media outlets, limiting their ability to identify systematic patterns across broader media landscapes. Examining multiple outlets over extended timeframes can reveal patterns that smaller studies might miss. Building on this insight, our study is designed around cross-outlet pattern identification rather than single-case illustration.

This study addresses these gaps through a systematic examination of IPV media coverage in Chinese news discourse. Media coverage in this study refers to news articles published in major Chinese newspapers, and does not include television or social media outlets. Given the traditional nature of newspaper reporting in China, these outlets are seen as central to shaping public discourse, although the digital media landscape has increasingly become a source of alternative information. By investigating how IPV is discursively constructed in major Chinese newspapers, this research illuminates how these representations reflect and shape broader public understanding. Through corpus analysis and framing analysis of media texts spanning 2012 to 2022, this study provides new insights into IPV portrayal within China’s complex media landscape.

Following Phelps et al. (2014), who define voice as “activities that encourage reflection, discussion, dialogue and action” (p. 34), this research distinguishes (i) actor presence (who is represented/mentioned) from (ii) discursive voice (who is quoted or otherwise attributed speech/authority) in Chinese news coverage of IPV. This approach distinguishes between being discussed and having an actual voice in media discourse, recognizing that having a voice is prerequisite to having power and agency in shaping public understanding. By examining both representation and voice, this study provides a nuanced analysis of power dynamics in Chinese media discourse about IPV.

Accordingly, this study proposes the following research questions:

RQ1a: Which social actors are most frequently represented/mentioned in Chinese news coverage of IPV between 2012 and 2022?

RQ1b: Which social actors are granted voice (quoted authority) and how is quoted authority distributed across actor groups?

RQ2a: What are the dominant frames in Chinese news media coverage of IPV from 2012 to 2022, and which frames remain marginal or absent?

RQ2b: What spatial logics do these frames enact (domestic containment, administrative regulation, or criminal justice intervention)?

By addressing these questions, this study contributes empirical insights about Chinese media representations of IPV and methodological approaches for analyzing media discourse about sensitive social issues in diverse cultural contexts. The findings may inform more effective communication strategies for IPV awareness and prevention in China.

3. Theoretical Framework

3.1. Theoretical Integration: Framing, Discourse, and Space

This study adopts a corpus-assisted discourse-analytic approach that combines framing analysis, actor representation, and spatial theory as sequential analytical lenses. Framing theory (Entman 1993; Scheufele 1999) guides how we interpret recurrent patterns as problem definition, causal attribution, moral evaluation, and treatment recommendation; discourse analysis (van Leeuwen 2008) specifies how these frame functions are linguistically enacted through actor representation and voice allocation; spatial theory (Lefebvre 1991; Couldry and McCarthy 2004; Rodgers et al. 2014) supports interpreting their combined effect as symbolic “spatial positioning” of violence and intervention. Empirically, these lenses are operationalised through corpus evidence—keywords/collocates/concordance for frame patterning, quoting-verb concordances for voice attribution, and lexical clusters indexing domestic, criminal, and administrative domains for spatial containment.

3.2. Framing Theory: Problem Definition and Solution Frames

Framing theory, initially developed by Goffman (1974) and later extended by Entman (1993) and Scheufele (1999), argues that media do not simply mirror reality but actively construct meaning through the selection and salience of specific aspects of reality. Media frames help define problems, identify causal agents, make moral judgments, and suggest solutions. This concept is crucial for understanding how issues like IPV are represented in the media and, importantly, how public perceptions of IPV are shaped by framing mechanisms.

3.3. Discourse Analysis and Actor Representation

The relationship between framing and actor representation is central to this study. As framing theory highlights, who gets to speak and how actors are represented in the media are not neutral but are shaped by power structures. For instance, in the context of IPV, governmental actors, such as the Women’s Federation or judges, are often granted more authority to speak on the issue than survivors or non-state actors. This hierarchical representation in the media constructs IPV as a legal or governmental issue rather than a personal or societal issue requiring broader public engagement.

To systematically operationalize these power dynamics in our linguistic analysis, we turn to specific frameworks of actor representation. This study draws specifically on van Leeuwen’s (2008) categories of activation/passivation and nomination/categorisation. Activation refers to the representation of social actors as active agents performing actions, while passivation represents them as the recipients or beneficiaries of others’ actions. Nomination refers to the identification of actors by name or individual identity, while categorisation identifies actors by role or function (e.g., ‘victim’, ‘official’, ‘perpetrator’). In the corpus analysis, these categories are operationalised through concordance examination of quoting verbs (who is the grammatical subject of speech attribution) and through the frequency of categorising labels applied to different actor groups. Specifically, we examine whether survivors appear primarily as grammatical objects of others’ actions (passivated) or as grammatical subjects exercising agency (activated), and whether governmental actors are predominantly nominated as authoritative speakers while survivors are categorised as passive recipients. This operationalisation connects directly to Entman’s (1993) framing dimensions: actor representation shapes both causal attribution (who is responsible) and treatment recommendation (whose expertise determines solutions).

3.4. Spatial Theory and Its Operationalisation

As discussed in the Introduction, the concept of spatial containment is key to understanding how IPV is framed in Chinese media. Another critical theoretical lens in this study is Lefebvre’s theory of the social production of space (1991), which examines how spaces are socially constructed rather than simply existing as neutral backdrops. Lefebvre’s framework provides a foundation for understanding how IPV discourse in media is spatially contained within domestic and administrative spheres, rather than being framed as a criminal justice issue in public spaces. According to Lefebvre, media discourses actively produce and define social spaces through the ways they represent issues, such as IPV, within particular domains. For example, IPV is often positioned within the private realm (family, marital space) rather than the public realm (courtrooms, police stations). This spatial containment limits the scope of intervention and makes the problem of IPV seem like an issue to be handled privately, thus reinforcing gendered power dynamics and social norms about the privatization of domestic violence. This theoretical concept allows us to track how Chinese news media privileges private, familial solutions (e.g., mediation, divorce) over formal legal responses (e.g., prosecution, police involvement). The absence of criminal justice terms in much of the coverage indicates how public responsibility for IPV is actively undermined.

To further understand how these spatial boundaries are discursively constructed and maintained in news reporting, we rely on media geography scholarship. This study draws on Rodgers et al. (2014) and Couldry and McCarthy (2004) to understand how spatial frameworks in media discourse direct public attention to private realms while marginalizing public spaces where accountability and formal intervention could occur. Couldry and McCarthy (2004) argue that media do not merely report on spaces but actively produce symbolic spatial boundaries, determining which issues ‘belong’ in public or private domains, and which actors have legitimate presence in those domains. Rodgers et al. (2014) extend this to show how news coverage of urban social problems constructs jurisdictional imaginaries: by repeatedly locating a problem within a particular institutional or geographic domain, coverage naturalises that domain as the appropriate site of intervention. Applied to IPV, this means that consistent use of domestic/marital vocabulary does not simply describe where violence occurs but actively produces the domestic sphere as the natural and legitimate boundary of both the problem and its resolution.

These spatial constructions are operationalised in the present study through three analytical categories derived from both Lefebvre’s spatial triad and the corpus lexicon: domestic/private space (identified through family and marital vocabulary), criminal/public accountability space (identified through prosecution and policing vocabulary), and administrative/civil governance space (identified through protection order and mediation vocabulary). The relative frequency and collocational patterns of these lexical clusters provide empirical evidence of which spatial domain is discursively dominant. By tracking the lexical markers associated with domestic space (e.g., 家庭 [jiating, family], 婚姻 [hunyin, marriage], 夫妻 [fuqi, couple/spouses]), we can see how these markers contribute to the privatization of IPV in media discourse, further reinforcing the gendered dimensions of domestic violence and hindering progress toward broader legal reform. The spatial containment of IPV in Chinese news coverage reframes the issue as one of family conflict or marital dispute, often minimizing its severity and the urgency for formal legal action.

4. Materials and Methods

This study employs a systematically constructed database of Chinese newspaper reports spanning 2012–2022 (see Appendix B for newspaper characteristics). This temporal scope deliberately aligns with significant legislative developments, including the implementation of Domestic Violence Law in 2015 and the 2020 Civil Code, enabling examination of how legal framework changes influence media discourse. The sampling framework utilizes a robust random constructed week approach, generating 14 days of coverage annually across 11 years. This methodology effectively controls for cyclical coverage patterns while ensuring comprehensive temporal representation. The investigation leverages WiseNews’s extensive archive of over 1200 Chinese newspapers, employing theoretically grounded keyword selection protocols following Yu (2023).

4.1. Corpus Construction and Analytical Scope

The final corpus, comprising 603 articles totaling 435,581 words, achieves sufficient size for quantitative and framing analysis of targeted discourse. This design follows standard corpus linguistic principles of using frequency, collocation, and concordance evidence as empirical entry points for discourse interpretation (Sinclair 1991; Baker 2006; Partington et al. 2013; Hunston 2022).

4.2. Data Analysis

We employed a corpus-assisted discourse-analytic design to reduce selective or impressionistic interpretation by grounding claims in reproducible lexical and attribution patterns. Specifically, we combined (i) corpus techniques (frequency, keyword, collocation, concordance) for pattern identification with (ii) qualitative coding for interpreting frame functions and voice allocation in context (Breeze 2011). Our method allows for the systematic and efficient identification of thematic content and frames, leveraging textual analysis software to empirically extract central ideas and reduce subjectivity—an approach validated by previous scholars (Matthes 2009; Miller 1997; Touri and Koteyko 2014; Wiedemann 2013).

4.3. Operationalizing Spatiality in Corpus-Based Framing

Following Lefebvre’s (1991) distinction between conceived space (representations of space, i.e., how institutions and discourses define and plan space) and perceived space (spatial practice, i.e., how space is used and experienced), we treat news discourse as a form of conceived space production. Media texts do not report neutral facts about where IPV occurs; they construct symbolic representations of domestic, legal, and administrative space that shape public understanding of where the problem ‘belongs’ and who has legitimate authority within it. The three spatial domains identified below—domestic/private, criminal/public, and administrative/civil—are therefore not descriptive categories derived post hoc from the data, but theoretically grounded analytical categories derived from Lefebvre’s framework and then populated through corpus evidence. We interpret the lexical patterns identified through our corpus analysis as indicators of spatial positioning—examining how violence, intervention, and authority are discursively located across domestic, criminal, and administrative zones. This approach follows media geography scholarship that treats space as discursively produced (Couldry and McCarthy 2004). Specifically, we track three ‘spaces’ via lexicon and concordance analysis:

(1) Domestic/private space: identified through home/family/marriage/dispute lexicon, such as jiating (family, 家庭), hunyin (marriage, 婚姻), fuqi (couple, 夫妻), jiufen (dispute, 纠纷), and tiaojie (mediation, 调解);

(2) Criminal/public accountability space: identified through police/investigation/prosecution/sentencing lexicon, such as xingshi (criminal, 刑事), gongsu (public prosecution, 公诉), panxing (sentencing, 判刑), daibu (arrest, 逮捕), and gongan (public security, 公安);

(3) Administrative/civil governance space: identified through mediation/divorce/protection orders/institutional programs lexicon, such as fayuan (court, 法院), lihun (divorce, 离婚), baohuling (protection orders, 保护令), Fulian (Women’s Federation, 妇联), and shequ (community, 社区).

Our spatial analysis is interpretive rather than geographically empirical: we examine how discourse positions violence and intervention across symbolic spatial domains (domestic, criminal, administrative), not physical locations.

4.4. Software and Tools

We used AntConc (Version 3.5.0), a corpus linguistic software, for concordance generation and text analysis (Anthony 2017). AntConc was chosen for its robust features in concordance generation and text analysis, facilitating detailed examination of word frequency and keyness. To objectively evaluate the IPV dataset, it was necessary to compare it with a reference dataset, generating a keyword list to identify salient lexical items. Keywords were defined as words that occur frequently in one corpus compared to another, identified through statistical tests considering word frequency and corpus size. AntConc’s default settings were used to conduct log-likelihood tests for keyness. Following Baker et al. (2008), we report keyword extraction settings (statistic, threshold) and treat keywords as starting points for concordance-based interpretation rather than as self-standing evidence of frames. The reference news dataset consisted of news reports published from 1 January 2012 to 31 December 2022.

Collocation and concordance procedures: For collocation, we used a symmetric window (e.g., ±5 tokens) and retained collocates above a minimum frequency threshold; association was evaluated using AntConc-supported measures and cross-checked through concordance inspection. For concordance analysis, we generated concordance lines for key node terms and sampled lines using a stratified procedure across outlets and years to avoid over-representing any single outlet or period.

5. Results

With regard to RQ1, this study identifies seven major sources: victims, perpetrators, family members, professionals (including lawyers and doctors), governmental officials, community supporter, and journalists. The analysis reveals that victims are the most visible in media coverage of IPV in China, while governmental officials are the most vocal. In this study, “visibility” refers to being mentioned/represented in the news narrative, whereas “voice” refers to being quoted or attributed speech/interpretive authority. In the case of references to victims, several specific terms in Chinese were used to describe victims, such as “受害人 (Victim)”, “被害人 (Victim)”, and “受害者 (Victim)”. These terms appeared a total of 1091 times. Overall, references to victims appeared 4420 times in the corpus, making victims the most frequently discussed group, as depicted in Figure 1.

The second most mentioned source in the corpus is the perpetrator. Words related to perpetrators, such as “他” (he) and “某” (someone), were also examined. The total occurrence of these terms is 2451.

To examine who was quoted most frequently, we analyzed the concordance lines of six popular quoting verbs in Chinese: “说” (speak), “表示” (express), “解释” (explain), “强调” (stress), “介绍” (introduce), and “认为” (believe). There were 3690 instances of these quoting verbs in the corpus. After eliminating unrelated lines and manually checking all the lines, we obtained 1361 relevant lines, categorized into seven groups, including quotes by victims, perpetrators, family members, and others.

Despite the high frequency reference to of victim, their voices were barely heard in media accounts. Of all the quoting verbs associated with victims, there were only 233 instances out of 1361 total quoting verb lines, as demonstrated in Figure 2.

A similar pattern is observed with the perpetrators. While the perpetrators rank second in the visibility counts, their voices only appeared 84 times in the corpus.

The prominence of governmental officials in the media was evident. “Fulian” (the All-China Women’s Federation) appeared 972 times and was the 12th most frequent word in the corpus. Their voice was also the most prominent in the quoting sample. Among governmental officials, “Fulian” was the most cited group, followed by “faguan” (judge), which appeared 439 times, while “jingcha” (police) was the least cited, appearing only 56 times. In a manual review of the quoting sample, only a small number of concordance lines attribute interpretive authority to police, suggesting that police voice is comparatively muted relative to quasi-state governance actors (e.g., ACWF) and the judiciary.

In contrast, references to professionals, such as “lüshi” (lawyers), “yanjiu” (researchers), and “yisheng” (doctors) (566 times), were much fewer than those to governmental officials (1467 times), as demonstrated in Figure 1. Similarly, these professionals’ voices were less quoted in the corpus compared to governmental officials, as demonstrated in Figure 2. References to social workers were infrequent, with only 187 mentions of “gong yi” (non-profit organization) in the corpus. However, their voice was more prominent than that of family members and perpetrators, as depicted in Figure 2.

Finally, among the examined sources, references to (jizhe) journalists formed the smallest group among the actor groups, being mentioned only 40 times in the corpus. Their voice was also the least represented among the quoting concordance lines.

5.1. Zones of Legitimacy in Source Attribution

These patterns reveal an uneven geography of voice in IPV discourse. State and quasi-state actors, particularly the Women’s Federation and judiciary occupy what we term zones of public legitimacy: discursive spaces where speakers are authorized to define problems, propose solutions, and speak on behalf of others. Survivors, by contrast, remain spatially confined to zones of private experience—visible as objects of discussion but rarely granted the discursive position of authorized speakers. This spatial distribution of voice reinforces the broader containment of IPV within domestic space: survivors’ knowledge remains ‘inside’ while institutional knowledge circulates ‘outside’ in public discourse.

5.2. Framing IPV Primarily as a Family Issue

Chinese media discourse overwhelmingly frames IPV as a private family matter requiring internal resolution rather than a social or criminal issue. While terms like “jiating” (家庭, family), “baoli” (暴力, violence), and “jiabao” (家暴, domestic violence) predictably appear frequently given the topic, the family frame becomes evident through how IPV is constructed as marital conflict and positioned within private domestic space.

IPV is linguistically constructed through intimate relationship terminology rather than perpetrator/victim or criminal language. In Appendix A, terms like “zhangfu” (丈夫, husband, 1043 tokens), “fuqi” (夫妻, couple, 688 tokens), and “qizi” (妻子, wife, 595 tokens) feature prominently, positioning violence within reciprocal spousal relationships rather than as unidirectional criminal acts. This symmetrical framing obscures power dynamics and perpetrator accountability. Most tellingly, IPV discourse is dominated by marriage breakdown vocabulary rather than violence intervention language. “Hunyin” (婚姻, marriage) appeared 879 times and “lihun” (离婚, divorce) 1559 times, demonstrating that IPV is primarily constructed as a marital problem leading to separation rather than a violent crime or public health crisis requiring intervention. Concordance analysis reveals divorce appears as the predominant outcome of IPV rather than criminal prosecution or systematic support.

The framing is further reinforced through terms reflecting interpersonal disputes rather than systematic violence. “Jiufen” (纠纷, dispute, 615 tokens), “maodun” (矛盾, conflict, 383 tokens), “zhengchao” (争吵, quarrel), and “polie” (破裂, breakup) frame IPV as mutual disagreement rather than criminal abuse. This language positions violence within normalized conflict patterns rather than as criminal behavior requiring intervention. Additionally, “tiaojie” (调解, mediation, 677 tokens) appears frequently, emphasizing informal, private resolution within family or community networks. Concordance patterns show mediation positioned as the preferred intervention before legal action, drawing on cultural frameworks that prioritize family harmony and relationship preservation. Specifically, the discourse invokes principles such as 家和万事兴, “when the family is harmonious, everything prospers”) and 家丑不可外扬 (“family shame should not be made public”), which naturalize private resolution and create pressure to avoid publicizing family problems that would bring loss of face (丢脸) or damage to family reputation (面子). However, as Hershock and Ames (2006) demonstrate, this represents selective deployment of certain cultural traditions to support social stability and family-based resolution, rather than comprehensive adherence to tradition. Alternative cultural resources exist in contemporary China—including socialist gender equality ideology (妇女能顶半边天 “women hold up half the sky”), legal rights discourse (依法治国 rule of law), and international human rights frameworks—but these are marginalized in favor of family harmony framing (Faure and Fang 2008). This selective emphasis suggests strategic framing choices rather than cultural determinism.

This framing aligns with cultural imperatives prioritizing family cohesion but may undermine violence severity, leading to underreporting and weak deterrence (Chan 2012; Yuan and Hesketh 2021). While the family frame may reduce stigma for help-seeking within private networks, it simultaneously privatizes violence and limits recognition of IPV as requiring formal intervention systems.

5.3. Spatial Containment: Locating Violence in the Domestic Sphere

The family/marriage/dispute lexicon identified above does not merely ‘frame’ IPV—it spatially places violence within the domestic sphere. This discursive placement operates as spatial containment: by repeatedly locating violence in the home (as marital conflict, family dispute, or household matter), media discourse restricts the imaginable terrain of intervention to private, familial settings. The dominance of terms like 家庭 (family), 婚姻 (marriage), and 夫妻 (couple) constructs a spatial boundary around violence that naturalizes domestic resolution and makes public/criminal intervention appear as intrusion into private space. This containment strategy has material consequences: it narrows the horizon of accountability and positions formal legal action as exceptional rather than standard response to violence.

5.4. Rarely Framed as a Criminal Legal Issue

Greater legal specificity is needed to avoid overgeneralization. The Anti-Domestic Violence Law (effective 2016) primarily provides a civil–preventive governance toolkit (e.g., protection orders and multi-agency coordination) rather than automatically “criminalizing” IPV as a single offense category. By contrast, criminal liability typically applies when specific acts meet criminal law thresholds (for example, serious physical harm or other prosecutable offenses), which means “IPV as crime” is not a default legal framing for every reported incident. Accordingly, our claim is not that IPV is “not a crime” legally, but that criminal prosecution framing is comparatively backgrounded in news discourse relative to civil/administrative management and family dispute narratives.

While legal terminology appears in the corpus, “falü” (法律, law, 1366 tokens), “fayuan” (法院, court, 1661 tokens), and “faguan” (法官, judge, 439 tokens), collocational analysis reveals these terms appear predominantly in civil divorce proceedings rather than criminal prosecution contexts. The primary legal reference is “Hunyinfa” (婚姻法, Marriage Law), positioning legal intervention within family law rather than criminal justice frameworks. Legal language clusters around divorce settlements, custody disputes, and property division rather than criminal accountability. Terms like “panjue” (判决, judgment, 285 tokens) and “caiding” (裁定, ruling, 301 tokens) appear predominantly in civil rulings rather than criminal sentencing contexts.

Criminal justice vocabulary is conspicuously absent or underrepresented. Terms indicating criminal accountability, “xingshi susong” (刑事诉讼, criminal litigation), “gongsu” (公诉, public prosecution), or “panjing” (判刑, criminal sentencing) appear infrequently in the corpus. Even when “fanzui” (犯罪, crime, 403 tokens) appears, concordance analysis shows it describes extreme cases or conditional scenarios (“if it constitutes a crime”) rather than the standard frame for understanding IPV. Police-related terms appear, like “baojing” (报警, call police, 343 tokens) and “paichusuo” (派出所, police station, 265 tokens), but concordance patterns reveal police positioned as mediators and referral points rather than criminal investigators. References describe incident reporting and administrative response rather than evidence collection, criminal investigation, or arrest procedures.

Post 2016 coverage includes references to “renshen anquan baohu ling” (人身安全保护令, personal safety protection orders). Yet these appear as civil protective measures rather than criminal sanctions. This systematic underrepresentation of criminal framing may reflect cultural preferences for family resolution (Faure and Fang 2008), but it means that even the criminal prosecution pathway that exists through Criminal Law (Articles 260 and 234 on family abuse and intentional injury) remains largely invisible in public discourse. While the 2016 Anti-Domestic Violence Law itself emphasizes civil and administrative mechanisms, the near absence of references to available criminal provisions potentially undermines accountability for severe violence and limits public understanding of the full range of legal responses available to IPV survivors (Berns 2004; Tiwari et al. 2018).

5.5. Spatial Deemphasis of Criminal Accountability

The comparatively low salience of criminal prosecution vocabulary suggests a systematic spatial deemphasis of public accountability. Criminal space, the domain of investigation, prosecution, sentencing, and state punishment, appears in the corpus as conditional and exceptional rather than as a routine response script for IPV. When criminal terminology does appear, it is typically hedged (‘if it constitutes a crime’) or displaced by civil/administrative alternatives. This pattern enacts what we term administrative rerouting: rather than moving IPV into the public space of criminal investigation and prosecution, media discourse channels it into civil divorce, mediation, and bureaucratic remedies, zones that manage conflict without necessarily naming violence as a public wrong requiring criminal accountability.

5.6. Virtually Absent: Public Advocacy and Community Support Frame

Public advocacy and grassroots community approaches are systematically excluded from IPV media discourse, rendering non-governmental support systems and survivor-centered advocacy invisible.

While community-related terms appear in the corpus, “shequ” (社区, community, 662 tokens) and “zuzhi” (组织, organization, 603 tokens), concordance analysis reveals these describe state administrative structures, rather than grassroots, support the networks. “Shequ” consistently pairs with “juweihui” (居委会, neighborhood committee) and “paichusuo” (派出所, police station), indicating government jurisdictions rather than community-based advocacy. These terms describe state presence at local level, not independent community organizing. Similarly, “xuanchuan” (宣传, promotion/publicity, 539 tokens) appears predominantly in time-limited, government-organized campaigns rather than ongoing advocacy efforts. Concordance patterns reveal phrases such as “反家暴日宣传活动” (Anti-Domestic Violence Day promotional activities), “妇联组织宣传” (Women’s Federation organized promotion), and “发放宣传资料” (distributing promotional materials), indicating episodic, state-led awareness events rather than sustained public advocacy campaigns. “Xuanchuan” consistently appears with “Fulian” (妇联, Women’s Federation, 972 tokens) as the organizer, positioning awareness-raising as government function rather than civil society advocacy.

Independent civil society terminology is virtually absent from the corpus. “Gongyi” (公益, non-profit/public welfare) appears only 187 times, and concordance analysis shows even these references often describe government-initiated programs rather than independent NGO work. Terms indicating grassroots organizing, “mingjian zuzhi” (民间组织, grassroots organizations), “zhiyuanzhe” (志愿者, volunteers), or explicit references to NGOs, are extremely rare. Manual concordance review reveals fewer than 20 lines clearly describing independent, non-governmental advocacy organizations working on IPV issues.

Most critically, survivor agency and empowerment language is virtually non-existent in IPV media discourse. Terms indicating empowerment (“fuzhuan”, 赋权), peer support (“xianghu zhichi” 相互支持), or survivor-led advocacy do not appear in top keywords and are extremely rare in the full corpus. This absence reveals how IPV discourse systematically excludes survivor voices and community-based support models, constructing IPV response as exclusively managed by family systems or government institutions. This systematic exclusion reflects media environments where state-driven solutions dominate public discourse (Stockmann 2013), obscuring alternative support mechanisms available through community organizations and survivor networks. The pattern reinforces a binary framework, IPV is either a private family matter or a government concern, erasing grassroots advocacy and limiting public understanding of IPV as requiring multi-sector, survivor-centered responses.

5.7. Community as State-Administered Terrain

The lexical patterns above reveal that ‘community’ in Chinese IPV discourse is not constructed as a survivor-centered public arena but as state-administered terrain. When community appears, it designates government jurisdictions (neighborhood committees, police stations, Women’s Federation local branches) rather than grassroots, peer-based, or survivor-led spaces. This spatial construction matters because it forecloses the possibility of community as an autonomous site of collective action, mutual support, or advocacy independent of state institutions. The near-total absence of empowerment vocabulary (赋权), peer support language (相互支持), and NGO references confirms that the ‘community’ terrain available in discourse is already occupied and administered by state actors, leaving no discursive space for survivor-centered alternatives.

6. Discussion

This study reveals a key issue in Chinese IPV media coverage: despite the 2016 Anti-Domestic Violence Law establishing IPV as a public concern requiring multi-institutional state intervention, media coverage persistently frames IPV as a private family problem requiring informal resolution. Significantly, media discourse goes beyond the law’s civil and administrative emphasis to actively marginalize IPV within domestic resolution space. While the 2016 law did not criminalize IPV as a distinct offense, it mandated formal institutional response through protection orders, police intervention, and multi-agency coordination. Media coverage, however, emphasizes family mediation and reconciliation over even these civil/administrative mechanisms, suggesting a lag behind the legal framework’s repositioning of IPV from private to public terrain. This finding supports previous research showing that legal reforms do not automatically change how the public understands IPV: IPV remains under-recognized, especially in rural areas where legal awareness is limited (Yuan and Hesketh 2021).

Read spatially, the ‘family issue’ frame functions as discursive containment: violence is repeatedly located in the home, which narrows the horizon of accountability and makes public/criminal intervention appear secondary or excessive. This spatial containment strategy confines violence to the domestic sphere and limits public accountability, explaining why legal reforms may coexist with weak criminalization in everyday media narratives.

This pattern can be understood as a discursive governance arrangement rather than a simple “lack of legal awareness.” News routines and institutional constraints encourage solution scripts that are administratively manageable and culturally acceptable (mediation, reconciliation, divorce, protection orders), while criminal prosecution scripts are treated as exceptional. In this sense, legal reform and discursive domestication can coexist because news discourse repeatedly routes IPV into civil–administrative and domestic spaces where punishment is not the default narrative endpoint.

A key finding is that, despite the existence of legal tools like the Domestic Violence Law, IPV is still often framed as a family issue, not a criminal one. Media coverage tends to focus on family mediation and reconciliation, reflecting deep-rooted cultural values prioritizing family cohesion (Faure and Fang 2008; Hershock and Ames 2006). This finding echoes Chan’s (2012) observation that in Chinese discourse, IPV is often viewed as a marital conflict, not as a gendered crime that requires legal action. However, the “mild tone” of IPV coverage is not only cultural; it is also produced through journalistic conventions. Three mechanisms are particularly relevant:

(1) Source routines: heavy reliance on official and quasi-official voices (e.g., ACWF, judiciary) privileges governance talk (education, mediation, administrative programs) over survivor-centered narratives and over criminal investigation detail.

(2) Event narrative scripting: IPV is often narrated as a domestic “incident” embedded in marriage/family breakdown, which makes divorce/mediation the narratively coherent outcome and makes prosecution appear disruptive or excessive.

(3) Risk-avoidant framing: sensitive social issues are more readily reported through institutionally sanctioned vocabularies (family dispute, mediation, stability-oriented solutions) than through accusatory criminal language, especially in routine news reporting.

These routines intersect with cultural norms (harmony, privacy, face) to stabilize a “manageable” discursive model of IPV: visible but domesticated, governable but rarely punitive.

This pattern also indicates administrative rerouting: rather than moving IPV into the public space of criminal investigation and prosecution, media and legal discourse channel it into civil divorce, mediation, and bureaucratic remedies—zones that manage conflict without necessarily naming violence as a public wrong. The predominance of civil/administrative vocabulary over criminal justice terminology demonstrates how IPV is systematically redirected away from spaces of criminal accountability.

A second important finding is the state’s dominant role in IPV coverage, especially through voices like the All-China Women’s Federation (ACWF). While government representatives frequently speak on IPV, criminal enforcement is rarely discussed. This supports Fenton’s (2011) argument that in authoritarian states like China, the media often present the state as a problem-solver, focusing on mediation and social services instead of criminal justice. As Strömberg (2015) notes, powerful groups, including the state, can shape media narratives, leading to coverage that focuses on state-managed solutions rather than criminal action. This reflects a hybrid governance model, where the state is present but avoids emphasizing legal enforcement.

Source attribution thus produces uneven zones of legitimacy: state and quasi-state actors occupy the discursive space of public reason-giving, while survivors’ experiential knowledge is kept largely outside the public sphere, appearing as an object of narration rather than an authorized voice. This spatial distribution of voice—where institutional speakers are positioned in ‘public’ discursive space while survivors remain in ‘private’ experiential space—reinforces the broader containment of IPV as a domestic rather than public concern.

Another important pattern is the lack of victim voices in media coverage. Although victims are frequently mentioned, they are rarely quoted or given a chance to tell their stories. Seely and Riffe (2021) found a similar pattern, where institutional voices dominate in media coverage of IPV, and victims are sidelined. This echoes Lloyd and Ramon (2017), who found that, even in the UK, media often frame victims in a way that reflects victim-blaming attitudes. In Chinese media, victims are discussed by others but are not portrayed as active participants in their own stories. This shows that, even when victims are visible, their voices are often silenced.

Lastly, the study highlights that the civil society and grassroots advocacy in IPV coverage is almost absent. Independent NGOs and survivor-led groups are almost completely absent from IPV discourse in China, a possible reason of which could be the media, which, as argued by Breeze (2011), can actively exclude alternative voices. Most references to “community” in the media refer to state-run organizations, such as neighborhood committees, rather than independent NGOs. This matters analytically because it narrows the discursive menu of “legitimate responses” to either domestic/private resolution or state-administered programs, leaving little space for survivor-led empowerment or peer-based infrastructures.

The near absence of grassroots advocacy further shows that ‘community’ is not constructed as a survivor-centered public arena but as a state-administered terrain, where support is articulated through campaigns, committees, and institutional programs rather than collective or peer-based infrastructures. This spatial construction of community forecloses alternative sites of collective action and mutual support independent of state institutions.

Cultural conventions are therefore not merely “invoked”; they function as framing resources that make certain solutions narratively and morally coherent. The emphasis on family harmony and mediation fits with Faure and Fang’s (2008) observation that Confucian values, such as family cohesion, often take precedence over legal accountability in China. These cultural values are used to justify mediation as the preferred solution, but they also overshadow alternative responses to IPV, such as legal action or community-based support. This finding echoes Hershock and Ames (2006), who argue that Chinese media selectively use Confucian values to support social stability, often at the expense of stronger legal protections for vulnerable people.

7. Conclusions

Overall, Chinese IPV news discourse does not merely describe violence, it places it—in a manner that domestically contained, administratively managed, and publicly voiced through state legitimacy. This spatialization helps explain why legal reforms may coexist with weak criminalization in everyday media narratives: the problem is repeatedly relocated to spaces where punishment and public accountability are not the default response. The near absence of grassroots advocacy and survivor voice suggests a form of discursive abandonment: while the state is present through campaigns and institutional programs, survivors are left without peer-based support systems or spaces for collective empowerment within public discourse.

This study fills an important gap in the literature by providing one of the first large-scale analyses of IPV media coverage in China. It highlights how legal reforms coexist with ongoing framing of IPV as a family issue, reinforcing the tension between legal changes and traditional cultural practices. Additionally, it shows how media framing, by controlling who gets to speak and who is quoted, marginalizes survivor voices and alternative responses to IPV.

This study is limited to (1) focus on newspapers only, excluding television, radio, and social media; (2) reliance on published coverage, which may not capture unpublished stories or editorial decisions; (3) corpus-based approach identifies patterns but cannot explain individual journalist motivations; (4) focus on mainstream outlets may miss alternative or regional media voices. These limitations are now explicitly acknowledged and connected to suggestions for future research.

These findings have important implications for policymakers and media professionals. Legal reforms alone are unlikely to be effective without changes in media practices. The media should focus more on the criminal aspects of IPV, include survivor voices, and present criminal accountability as a legitimate response. These changes could help reshape public discourse and improve awareness of legal protections for survivors.

Future research should focus on (1) examining digital and social media IPV coverage; (2) comparative studies with other Asian contexts; (3) reception studies examining how audiences interpret IPV media frames; (4) longitudinal analysis of how coverage changes with continued legal reforms; (5) investigation of journalist practices and newsroom decision-making regarding IPV stories.