Media Representation of Autism in Chinese Official News: A Topic-Modeling-Assisted Discourse Study

Global Media Representation of Autism: Dominant Narratives and Stereotypes

Internationally, research has consistently shown that media representations of autism often perpetuate stereotypes and can contribute to public misunderstanding. A common theme is the medicalization of autism, where individuals are frequently depicted as dependent on clinical interventions, cures, or treatments (Baroutsis et al., 2021; Ellis & Gerard, 2015; Hamilton, 2019). This is often coupled with portrayals of individuals with ASD as socially pathological, burdensome, objects of pity, or even dangerous (Holton et al., 2014; Jones & Harwood, 2009; Loftis, 2015; Osteen, 2008; Sarrett, 2011). The autistic savant trope, popularized by films like Rain Man, is another pervasive stereotype, where exceptional abilities are highlighted, often at the expense of portraying the diverse realities of ASD (Belcher & Maich, 2014; Draaisma, 2009; Fontes & Pino-Juste, 2021; Lin & Ma, 2023; Sarrett, 2011). Such representations can reinforce what Sarrett (2011) terms a “scientific reality of autism” (p. 143), often leading to themes of fragmentation and the “imprisonment of the normal child” (p. 145).

Studies across various Western countries have revealed specific framing patterns. In the U.S., research has examined the autism-vaccine controversy (Clarke, 2008; Mann, 2019), the use of “loss frames” emphasizing adverse outcomes (Holton et al., 2014), and the prevalence of episodic framing in television news (Kang, 2013). McKeever (2013) noted an increase in policy-related frames and a focus on human interest stories. Similarly, Wendorf Muhamad and Yang (2017) found that U.S. newspapers often frame autism through personal interest and attribution of responsibility. Clarke (2012) also highlighted how American women’s magazines adopted subjective “tragedy frames” compared to more objective “medical and science frames” in gender-neutral publications.

In the U.K., Huws and Jones (2011) found that newspaper representations (1999–2008) often featured stereotypical depictions and notably lacked autistic voices, a concern echoed by Baroutsis et al. (2021) in their analysis of Australian newspapers. Karaminis et al. (2023), in a recent large-scale corpus-based study of the British press (2011–2020), found that coverage emphasized adversities, used negative language, and focused on boys, though with a gradual shift towards more difference-based descriptions. Australian media research by Jones and Harwood (2009) initially identified portrayals of autistic individuals as dangerous or unloved, while more recent work by Baroutsis et al. (2021) pointed to a dominant focus on cures and treatments, often singling out individuals. Studies from other countries like Canada (Lanovaz et al., 2015; Wolbring & Mosig, 2017), Finland (Pesonen et al., 2021), and Vietnam (Nguyễn, 2023) further illustrate the global nature of these representational challenges.

Media Representation of Autism in the Chinese Context

While research on autism representation in Chinese media is less extensive, existing studies have revealed a distinctly discursive landscape. Bie and Tang (2015) and Tang and Bie (2016) conducted foundational analyses of Chinese-language newspapers (2003–2012), identifying a dominant “familial responsibility” frame. This frame places the primary burden of care on families, with limited societal or governmental acknowledgment, often accompanied by themes of shame and stigma (ibid.). Their work highlighted the underrepresentation of government voices in Chinese media compared to Western counterparts (Bie & Tang, 2015).

More recent research has explored new media and utilized advanced methodologies. Li et al. (2024) and Zhou et al. (2023), analyzing data from two social media platforms of Weibo and Zhihu, respectively, found that while public attitudes are becoming more emotionally positive, cognitive understanding of autism often remains underdeveloped and stereotypical. The discourse has also progressed to include discussions on language preferences and the cultural nuances of terminology, reflecting a growing awareness of identity and discursive rights (Lao et al., 2024). Multimodal analyses have also become prominent. Pan (2023) examined video PSAs, noting the prevalent “Children with Autism are Stars” metaphor, while Lin and Ma (2023) analyzed images in an official Chinese newspaper, finding portrayals of deficit and savantism. These studies suggest that while there is an increase in autism coverage, the representations can still perpetuate stereotypes or focus on limited aspects of the autistic experience. However, as Lin and Ma (2023, p. 337) note, “the literature on news’ representation of autism is quite sporadic in China,” and a significant gap exists in the analysis of official English-language Chinese news media, which is crucial for understanding how China communicates about ASD to an international audience.

Theoretical Framework: Framing Theory

This study is theoretically grounded in Framing Theory, which provides a robust lens for understanding how news media construct and disseminate representations of social issues like ASD. Framing, as conceptualized by Entman (1993), involves “select[ing] some aspects of a perceived reality and mak[ing] them more salient in a communicating text, in such a way as to promote a particular problem definition, causal interpretation, moral evaluation, and/or treatment recommendation” (p. 52). Frames, therefore, are not just about what is said but also about what is omitted, how information is organized, and which elements are emphasized, thereby shaping audience understanding and interpretation (Gamson & Modigliani, 1989; Nguyễn, 2023; Pan, 2023).

The power of framing lies in its ability to define problems, diagnose causes, make moral judgments, and suggest remedies (Entman, 1993). In the context of health communication and disability representation, media frames can significantly influence public perception, attitudes, and even policy responses. For instance, Iyengar (1991) distinguished between episodic frames, which focus on individual case studies or discrete events, and thematic frames, which place issues in a broader social, political, or economic context. This distinction is crucial, as episodic framing tends to lead audiences to attribute responsibility to individuals, whereas thematic framing encourages attribution to societal or systemic factors—a particularly relevant consideration for understanding disability narratives (Holton et al., 2014; Kang, 2013).

Previous research on media representation of autism has frequently employed framing analysis. Studies in Western contexts have identified various dominant frames, such as the “human interest frame,”“policy frame,” and “science frame” (McKeever, 2013), or “loss frames” versus “gain frames” (Holton et al., 2014). Wendorf Muhamad and Yang (2017) also utilized framing to analyze how U.S. newspapers portrayed autism, identifying distinct ways the issue was contextualized. The selection of sources plays a critical role in this framing process, as “accredited sources” often become the primary definers of topics, shaping the narrative and influencing public understanding (Hall et al., 1978; Nguyễn, 2023).

In the Chinese media context, Bie and Tang (2015) found that the absence of governmental voices in news coverage on autism led to a framing that located the challenges primarily within the private domain of families, rather than as a matter of public or social policy. This highlights how the selection and framing of information, including the voices represented or omitted, can direct public attention and accountability. Lin and Ma (2023), analyzing visual representations in an official Chinese newspaper, also implicitly touched upon framing by examining how images construct particular “gazes” and “social relations” that shape the portrayal of autism.

By applying Framing Theory to the China Daily corpus, this study aims to identify the dominant ways in which autism is presented to an international and policy-sensitive audience. The topic-modeling-assisted discourse analysis will help systematically uncover the salient keywords and broader thematic clusters that constitute these frames. Understanding these frames is critical because, as Entman (1993, p. 232) stated, “the character, causes, and consequences of any phenomenon become radically different as changes are made in what is prominently displayed.” This study, therefore, seeks to elucidate the specific frames used in China Daily’s discourse on autism and discuss their potential implications for public understanding and the social construction of ASD within an official Chinese English-language media narrative.

Discourse Analysis and Topic Modeling

Topic modeling bears resemblance to certain techniques utilized in Corpus Linguistics (CL) and serves as a significant adjunct to discourse analysis, offering a comprehensive content map of large document collections, thereby uncovering subtle yet systematic patterns and trends within the data. Critical Discourse Analysis (CDA) can further enhance this by facilitating a more in-depth and methodical qualitative analysis, transcending superficial exploratory outcomes and advancing towards a deeper understanding and explanation.

Combining CL and CDA has a long history, with various computer-based techniques employed to handle large datasets (Mautner, 2009). This integration has gained more attention recently under the term Corpus-assisted Discourse Studies (CADS), which encompasses approaches merging discourse analysis with CL methods (Cheng, 2013; Partington, 2006; Stubbs, 1994; Wodak & Meyer, 2009). These can be either methodological syntheses (Baker et al., 2008) or combined for triangulation purposes (Baker & Levon, 2015). CL involves empirical linguistic analysis using corpora to identify probabilities, trends, and patterns (Teubert & Krishnamurthy, 2007). Methods in CL range from simple frequency calculations to more complex qualitative analyses of word collocations and semantic patterns (Baker et al., 2008; Baker, 2023). Although not yet central to mainstream CDA, the potential of CL methods is increasingly recognized, with recent CDA studies incorporating them (Baker & McEnery, 2005; Nartey & Mwinlaaru, 2019; Orpin, 2005).

While CADS often begins with predefined keywords or concepts, this study incorporates topic modeling, a machine learning technique that offers a more inductive, corpus-driven approach to identifying thematic structures. Topic modeling, particularly Latent Dirichlet Allocation (LDA), identifies recurring word clusters to discover latent themes within a large collection of texts (Blei et al., 2003). Its application in media discourse analysis has grown significantly, providing a powerful tool for systematically analyzing extensive textual data.

For instance, Törnberg and Törnberg (2016) pioneered the combination of CDA and topic modeling to analyze the discursive links between Islamophobia and anti-feminism on a large internet forum. They argued that this synergy allows researchers to map the “dynamic terrain in which meaning contests occur” (Steinberg, 1999, p. 748), which they define as broader “discursive fields” (Törnberg & Törnberg, 2016, p. 408). This approach was instrumental in revealing how gender equality was used as a discursive strategy to criticize Islam. Subsequent studies have effectively applied this synergy to various media contexts. Jin (2023), for example, used topic modeling to examine the portrayal of DiDi’s delisting in Anglo-American media, revealing how the event was framed not merely as a business issue but as a threat to the Western-dominated financial order. Similarly, other research has used topic-modeling-assisted approaches to trace the “diachronic discursive construction” of political principles in news reports (Chen & Wang, 2023). These studies demonstrate the method’s utility in identifying prominent themes and interpreting them within broader socio-political and ideological frameworks.

By building on this established methodological framework, this study adopts the procedure advanced by Törnberg and Törnberg (2016). Topic modeling is first used to automatically and inductively structure the text to find recurrent thematic patterns. This is followed by a qualitative CDA to analyze the resulting topics in greater depth, focusing on both the keywords and the associated news reports to identify the broader discursive fields. This combination allows for a robust analysis that is both data-driven and theoretically grounded.

Procedure

To search for relevant terms, we adopted the keywords used by Karaminis et al. (2023) to ensure comprehensive retrieval of related articles from China Daily using the Factiva database. The search terms included: (autism* OR autistic* OR Asperger* OR aspergic* OR Aspie? OR ASD OR neurodiver* OR [“on the spectrum” NOT “on the spectrum of”]). The corpus was constructed using articles from China Daily retrieved via the Factiva database, covering the period from 1981 to 2023. This range reflects the full temporal coverage of China Daily available through Factiva, ensuring that our analysis captures both historical and contemporary representations of autism over more than 4 decades. It is noteworthy that this comprehensive and contemporary keyword set is not identical to those used in earlier foundational studies on this topic (Jones & Harwood, 2009; Kang, 2013), which often employed narrower search terms such as “autism” or “autistic.” This methodological distinction is intentional, as it allows our study to capture the more recent and nuanced discourse surrounding neurodiversity. All files were downloaded in RTF format (N = 441). We conducted a close reading of the remaining texts to ensure that each article focused meaningfully on autism-related themes such as diagnosis, education, therapy, family experiences, social inclusion, or public awareness. This process resulted in a refined dataset of 93 articles (409,400 tokens) that explicitly address autism-related issues, which formed the basis for our subsequent analysis.

After separating the metadata (author, date, etc.) from the titles and passages into different documents, the text files were preprocessed using a Python script (removing unrelated stopwords and lemmatizing). The preprocessed text was then imported into Python using the Gensim library to perform topic modeling. Perplexity and coherence scores were calculated to determine the appropriate number of topics (N = 7). The identified topics and categorized words are shown in Figure 1. In the present study, we selected concordances guided by keywords generated from topic modeling for detailed concordance analysis. This approach aims to provide a deeper understanding of the media representation of autism and autistic people beyond the context of individual linguistic items.

OPEN IN VIEWER

Figure 1.

Perplexity and coherence scores of the dataset.

Topic Modeling Results

As shown in Figure 2, the seven topics identified by the model cluster into three distinct thematic groups. Group A (Topics 1 and 5) focuses on the lived experience of autism within a social and educational context. Group B (Topics 4, 6, and 7) highlights the systems of support and professional interventions available. Group C (Topics 2 and 3), while sharing keywords with Group B, emphasizes the human agents of care and the collective aspiration for improvement.

OPEN IN VIEWER

Figure 2.

Topic distribution of the modeling results.

As shown in Table 1, this group contains a mix of keywords that ground the discourse in the everyday realities of individuals with autism. The prevalence of terms like “child,”“son,”“parent,” and “family” frames autism not as an isolated medical condition, but as a deeply familial experience. This narrative is further situated within an institutional context through keywords such as “school,”“teacher,”“student,” and “education.”

OPEN IN VIEWER

Table 1.

Keywords of Topics 1 and 5.

Topics	Keywords
1	School, said, say, year, autism, people, teacher, autistic, work, child, student, special, life, time, education, son, skill, one, social, public, first, training, china, center, job, boy, help, star, could, wang
5	Child, said, autism, autistic, education, social, special, also, disorder, year, school, son, year-old, parent, people, behavior, mother, help, one, boy, training, family, difficulty, interaction, according, many, still, brain, teacher, student

Source. Prepared by the authors based on the data (from the Factiva database) processed through Python.

This topic group constructs autism as a multifaceted lived experience. By focusing on both family dynamics and the educational system, the discourse moves beyond a purely clinical perspective. The keywords suggest that media coverage in China Daily frequently portrays autism through personal stories, highlighting the challenges and triumphs within the home and the classroom. This frame emphasizes the need for understanding and support in the two most crucial environments for a child’s development, suggesting a holistic and human-centric portrayal.

The keywords in this group, presented in Table 2, point strongly toward the formal structures of care and treatment. Terms such as “professional,”“therapy,”“treatment,”“care,” and “center” dominate this discourse. This highlights a narrative focused on the organized, institutional response to autism. The inclusion of keywords like “support,”“program,” and “activity” further reinforces the idea of structured interventions designed to assist individuals with ASD.

OPEN IN VIEWER

Table 2.

Keywords of Topics 4, 6, and 7.

Topics	Keywords
4	Said, autistic, autism, help, child, son, year, also, many, work, parent, training, school, people, one, say, China, center, treatment, family, cup, education, province, better, special, study, two, world, dog, could
6	Said, child, autistic, people, autism, support, China, also, parent, time, professional, care, therapy, say, teacher, family, son, boy, dog, year-old, hope, activity, many, need, program, may, year, like, mother, others
7	Child, said, autism, autistic, music, say, parent, care, people, family, many, year, like, training, treatment, therapy, teacher, center, school, take, service, also, need, life, learn, find, daughter, love, doctor, china

Source. Prepared by the authors based on the data (from the Factiva database) processed through Python.

This topic group frames the response to autism as a systematic and professional endeavor. The discourse emphasizes that addressing the needs of autistic individuals requires organized systems of support, including therapeutic services and specialized programs. This portrayal suggests that the media is communicating the importance of professional expertise and dedicated facilities in providing effective interventions, positioning the solution to the challenges of autism within a structured, service-oriented framework.

While the keywords in this group (Table 3) show significant overlap with Topic Group B, the subtle interpretive focus shifts toward the human element and the goal of progress. Here, keywords like “parent,”“teacher,” and “family” are not just contextual but represent the primary agents of care. This is coupled with aspirational terms such as “help,”“hope,” and “better.”

OPEN IN VIEWER

Table 3.

Keywords of Topics 2 and 3.

Topics	Keywords
2	Said, autistic, autism, help, child, son, year, also, many, work, parent, training, school, people, one, say, china, center, treatment, family, cup, education, province, better, special, study, two, world, dog, could
3	Said, child, autistic, people, autism, support, China, also, parent, time, professional, care, therapy, say, teacher, family, son, boy, dog, year-old, hope, activity, many, need, program, may, year, like, mother, others

Source. Prepared by the authors based on the data (from the Factiva database) processed through Python.

This topic group emphasizes the human side of the support network and the collective desire for positive outcomes. The narrative highlights the collaborative effort between families and professionals (“parent,”“teacher”) in seeking and providing care. The prominence of “hope” and “better” suggests a forward-looking, optimistic frame that focuses on progress and the potential for improvement in the lives of autistic individuals. This discourse is less about the institutions themselves and more about the people within them working toward a common, hopeful goal.

In summary, these three topic groups reveal a comprehensive narrative framework in China Daily’s coverage of autism. The discourse begins by grounding the condition in the Lived Experience of families and schools (Group A), then details the Systems of Support and interventions available (Group B), and finally highlights the human Agents of Care and their shared Aspirations for Improvement (Group C). This framework suggests a media portrayal that is not only informative but also empathetic and action-oriented, constructing autism as a complex human condition that requires a coordinated and hopeful response from all sectors of society. While this keyword analysis identifies the core actors and concepts, it does not reveal how they are portrayed. The next phase addresses this by examining the specific contexts in which these keywords are used.

Qualitative Analysis

Guided by the keywords identified through topic modeling and our close reading of the texts, we selected the keywords listed in Table 4 for concordance analysis in AntConc (see Figure 3). These keywords were chosen because they frequently appeared in contexts relevant to the representation of autism, including discussions of identity, treatment, and public attitudes.

OPEN IN VIEWER

Table 4.

Keywords for Concordance Analysis.

Keywords selected	Child, son, daughter, family, mother, parent, life
	Treatment, therapy, rehabilitation, school, education, teacher, doctor, center
	Help, assist, hope, support, future, better

Source. Prepared by the authors based on the data (from the Factiva database).

OPEN IN VIEWER

Figure 3.

A screenshot of concordances.

The qualitative analysis of news reports from China Daily demonstrates a portrayal of autism that, while acknowledging the inherent challenges, predominantly emphasizes narratives of familial dedication, professional intervention, and societal support, all framed within an overarching discourse of hope and the pursuit of a “better future.”

A prominent theme emerging from the China Daily data is the central role of the family, particularly mothers, in the narrative surrounding autism. Keywords such as family, mother, parent, son, daughter, and child frequently appear, grounding the experience of autism within the domestic sphere. For instance, the narrative of “Cao’s son” (“It is now home to 21 people with autism, including Cao’s son. Cao noticed that her son was a little different in 2002.”) immediately frames the mother as the primary observer and the one who initiates the journey of understanding and seeking help. This is further exemplified by statements like, “Like many parents of autistic children, Cao embarked on a thorny road of trying to help her son become ‘normal’,” which portrays the parental role as one of active struggle, dedication, and a quest for solutions, often involving significant personal sacrifice (“She went to Beijing and Hong Kong to learn how to educate autistic kids.”).

The discourse often highlights the emotional labor and resilience of parents. Sentences such as “The mother burst into tears when her son went to hug her and whispered ‘thank you’ to him” and “As a parent myself, it’s pleasing and meaningful to help autistic children with my professional knowledge” (Wang said) showcase not only the challenges but also the profound emotional rewards and motivations driving parental actions. While fathers are mentioned (e.g., “The father of Swedish teenage climate change activist Greta Thunberg…”), mothers are more frequently positioned as the primary caregivers and advocates for their autistic children (e.g., “On the other side of the screen sat Haorui and his mother. Holding a card showing an apple, the mother repeated the word over and over again…”). This portrayal underscores a familial responsibility framework, where the family unit, led by determined parents, is central to navigating the complexities of autism and seeking a “better” life for their children.

The China Daily coverage extensively features education, therapy, and rehabilitation as key pathways for the development and societal integration of individuals with autism. Keywords like “school,”“education,”“teacher,”“treatment,”“therapy,”“rehabilitation,” and “center” are highly prevalent, often appearing in contexts that emphasize structured support and professional intervention. For example, “Three of the six staff at Star’s Home are special education teachers who help the residents study and form healthy living habits,” and the establishment of specialized facilities like “Star’s Home” or a “rehabilitation center” (“Two suspects have been criminally detained after a police investigation into a rehabilitation center…”) signify the importance placed on institutional and professional support.

These interventions are often framed as challenging yet hopeful endeavors. Phrases like “Rehabilitation therapy can be challenging. But there is always hope.” capture this duality. The discourse suggests that through dedicated efforts from professionals and institutions, positive outcomes are achievable. Even when early interventions are described as difficult (e.g., Cao’s initial “wrong path to ‘cure’ her child”), the narrative often pivots towards more effective approaches centered on understanding and tailored support, such as “be taught according to their own character to help them adapt to adult life.” The media, therefore, constructs these interventions not merely as corrective measures for a “disorder” but as enabling systems that foster learning, skill development, and adaptation, ultimately aiming for a “better life” and “future” for individuals with autism.

A consistent and significant narrative thread within the China Daily reports is the discourse of support, hope, and the aspiration for a “better” future for individuals with autism and their families. This is evidenced by the frequent co-occurrence of keywords like “help,”“assist,”“hope,”“support,”“future,” and “better.” The reporting often highlights community initiatives, professional assistance, and policy measures aimed at improving the lives of those affected by autism. For instance, statements like “Star’s Home can give them the necessary companionship and help them live happier lives” and “The future looks bright” directly project optimism and the positive impact of support systems.

The concept of “help” is multifaceted, ranging from direct intervention (“help them adapt to adult life,”“help the residents study”) to broader societal efforts (“attract more public attention,”“help more children and more families in the future”). This suggests a portrayal where challenges are acknowledged, but the focus is often on the solutions and the collective responsibility to provide support. The keyword “hope” is particularly salient, appearing in contexts of parental perseverance (“But there is always hope.”), professional dedication (“She has brought hope to more than 200 families”), and even in the aspirations of individuals with autism themselves (“The reason I love my job is that I hope I can take care of my parents in the future,” Ying said). This pervasive framing of hope and support constructs a narrative that, while not ignoring the difficulties, emphasizes progress, societal compassion, and the potential for improved outcomes, thereby potentially fostering a more positive and encouraging public perception of autism.

While many narratives are centered on the actions of parents and professionals, there are instances within the China Daily reports where autistic individuals are portrayed with agency and their voices are directly or indirectly represented. The keyword “child (and son/daughter)” is overwhelmingly present, often positioning them as the recipients of care and intervention (e.g., “The special needs teacher was giving an online class to an autistic child”). However, there are also glimpses of their active participation and perspectives. For example, the report mentioning Ying, a resident at Star’s Home, includes a direct quote: “‘I’ve never been late for my job. The reason I love my job is that I hope I can take care of my parents in the future’, Ying said.” This quote gives Ying agency, portraying him not just as someone with autism, but as a responsible individual with aspirations.

Similarly, descriptions of children “learning valuable lessons” or how work can “help develop their social interaction abilities” suggest a capacity for growth and active engagement. The mention of “autistic children are encouraged to join in” (Yip’s project) further implies active participation rather than passive reception. However, it is also noticeable that many descriptions of autistic individuals’ behaviors are from the perspective of others (e.g., “Cao noticed that her son was a little different”; “He always crumbled every biscuit before eating”). The term “children of the stars” is also used, which, while potentially poetic, can also create a sense of otherness or distance. Therefore, while China Daily does offer instances of portraying agency and voice, the dominant framing still often positions autistic individuals, particularly children, as requiring guidance, care, and intervention from family members and professionals, with their direct perspectives being less frequent than those of their caregivers or support providers.