Beyond Individualism: Unpacking the Discourse of Student Mental Health Through Corpus Linguistics and Text Analysis

The intersection of student mental health and digital media representation has become a critical focal point for researchers, clinicians, and educational administrators. As academic pressures mount and institutional demands intensify, the way mental health issues are framed in online discourse significantly shapes public perception and policy responses. Recent scholarly investigations into the linguistic patterns within media coverage and the construction of psychological text corpora have revealed profound insights into how student distress is communicated, interpreted, and often misunderstood. By utilizing corpus linguistics, framing theory, and discourse analysis, experts can dissect the mechanisms through which student mental health is portrayed, moving beyond superficial narratives to uncover the structural and individualistic biases embedded in digital communication. This synthesis of methodologies allows for a deeper understanding of the emotional and cognitive dimensions of student well-being, providing a robust framework for analyzing text data to improve mental health interventions.

The prevailing narrative in online media often reduces complex mental health challenges to individual failings rather than systemic issues. A comprehensive analysis of news articles published between 2021 and 2025 reveals a distinct pattern in how student mental health is represented. The data indicates that students are predominantly framed as a vulnerable demographic struggling under academic pressure, while educational institutions are ambiguously positioned. They are portrayed simultaneously as the source of immense pressure and, in some contexts, as providers of psychological support. This duality creates a fragmented discourse where the structural causes of student distress are often obscured by an individualistic lens. The media coverage tends to focus on the symptoms and personal struggles of students, rather than addressing the institutional or societal factors contributing to the crisis. This individualistic framing has significant implications for how help is sought and provided, potentially stigmatizing students who are perceived as unable to cope with their environment.

The Mechanics of Media Framing and Discourse Analysis

To understand the representation of student mental health, one must first examine the theoretical underpinnings of media framing. Framing theory, as articulated by Entman (1993), posits that media selects specific aspects of a perceived reality and makes them more salient in a communicating text. In the context of student mental health, the "selection" process determines whether the issue is presented as a personal character flaw or a structural failure. When combined with Critical Discourse Analysis (CDA) principles derived from Fairclough, this approach allows researchers to deconstruct the power dynamics and ideologies hidden within news texts. The study of 30 news articles from the online media outlet Kumparan utilized these frameworks to reveal that the discourse is heavily skewed toward individualistic explanations.

The methodological rigor applied in these studies relies heavily on corpus linguistics. This field, championed by scholars like Baker (2006) and McEnery & Hardie (2012), provides tools to analyze large volumes of text data objectively. By using software like AntConc, researchers can perform word frequency analysis, identify collocations (words that frequently appear together), and generate keyword in context (KWIC) indexes. These techniques move the analysis beyond subjective interpretation, allowing for a data-driven understanding of how terms like "anxiety," "stress," and "burnout" are used in relation to students. The results consistently show that the language used in media coverage emphasizes the student as the locus of the problem. For instance, articles frequently pair words related to "pressure" and "schedules" with "student," but rarely link them to "systemic issues" or "institutional responsibility." This linguistic pattern reinforces a narrative where the student is the sole actor responsible for managing their mental health, often ignoring the environmental stressors.

The implications of this individualistic framing are profound. When mental health issues are framed as individual problems, the solution is implicitly placed on the shoulders of the student. This can lead to increased self-blame and reduced willingness to seek institutional support. Conversely, when the media does highlight the role of the campus, it is often ambiguous. Campuses are described as both the source of pressure (through rigorous schedules) and the potential savior (through counseling services). This ambiguity in the discourse creates a confusing landscape for students trying to navigate their mental health. The media, therefore, plays a dual role: it highlights the crisis while simultaneously obscuring the structural roots, potentially hindering effective policy changes.

Constructing Psychological Corpora: Methodology and Data Collection

While media analysis deconstructs public perception, the creation of specialized text corpora offers a direct window into the lived experience of mental health. The development of the PsihoRo corpus for the Romanian language exemplifies a rigorous approach to gathering mental health data that avoids the ethical and analytical pitfalls of mining social media. Unlike social media data, which is often collected without explicit consent and relies on suppositions made by the collector, this methodology prioritizes ethical data collection through informed consent.

The PsihoRo project represents a significant advancement in the field of Natural Language Processing (NLP) for psychology. It addresses the scarcity of open-source psychological corpora in non-English languages. The data collection strategy involves a structured survey divided into three parts. The core of this survey consists of six open-ended questions. These questions are carefully designed to elicit deep, narrative responses from participants. Three of these questions focus on positive experiences, while the other three target negative experiences, providing a balanced view of the participant's mental state. This approach ensures that the resulting text corpus captures a wide spectrum of emotional and cognitive expressions related to depression and anxiety.

To ensure the validity of the corpus, the survey integrates standardized clinical screening tools. Participants are required to complete the PHQ-9 (Patient Health Questionnaire-9) and GAD-7 (Generalized Anxiety Disorder-7) scales alongside their written responses. This combination allows for a direct correlation between the linguistic features of the text (the open-ended answers) and the clinical severity of the condition (the screening scores). The study gathered responses from 205 respondents, creating a foundational resource for analyzing the mental health of the Romanian population. Although the sample size may appear modest, the depth of the qualitative data and the integration with quantitative metrics provide a robust dataset for text mining and emotion detection.

This methodology offers a pragmatic alternative to the often flawed data collection from social media. By using open-ended questions and standardized surveys, researchers can bypass the noise and bias inherent in unstructured social media posts. The resulting corpus allows for sophisticated text analysis, including the use of the Romanian LIWC (Linguistic Inquiry and Word Count) tool. This enables researchers to detect emotional language, identify topic modeling patterns, and analyze specific linguistic markers associated with depression and anxiety. The data suggests that an individual's language is a reliable reflection of their mental health status. When verbal responses are combined with quantitative measures like PHQ-9 and GAD-7, the comprehension and assessment of mental health risks are significantly enhanced for both clinical practice and research contexts.

The following table outlines the structural components of the PsihoRo data collection method:

Linguistic Markers of Distress and Resilience

The analysis of text corpora reveals that the language used by individuals experiencing mental health challenges is distinct and analyzable. In the context of student mental health, the linguistic markers identified in media and clinical corpora provide a granular view of distress. In the Kumparan media analysis, the word "pressure" and "busy schedules" are heavily collocated with "student." This linguistic association reinforces the narrative that the student's environment is overwhelming, yet the discourse fails to explicitly link this to institutional failings. Instead, the focus remains on the student's ability to cope.

In contrast, the PsihoRo corpus analysis of 205 respondents provides a more clinical perspective. The text analysis reveals that individuals with higher scores on PHQ-9 and GAD-7 use specific linguistic patterns. These patterns include higher frequencies of words related to negative emotion, physical symptoms, and cognitive rigidity. The integration of the open-ended responses with the screening scores allows researchers to map the linguistic signatures of depression and anxiety. For example, texts associated with high depression scores may exhibit a higher density of words related to hopelessness, isolation, and fatigue, while texts associated with anxiety might show higher frequencies of words related to worry, uncertainty, and future-oriented fear.

The utility of these linguistic markers extends beyond academic research. In clinical settings, analyzing a patient's language can serve as an early warning system. By identifying specific vocabulary and syntactic structures associated with mental health issues, clinicians can detect risks before they escalate. This approach bridges the gap between qualitative narratives and quantitative diagnostics. The ability to detect these markers in text allows for a more nuanced understanding of the patient's internal state, complementing traditional assessment methods.

Furthermore, the corpus analysis highlights the importance of context. The distinction between "positive" and "negative" questions in the PsihoRo survey is crucial. It allows for the detection of resilience markers as well. Texts answering positive questions from individuals with low symptom scores often contain language of hope, agency, and social connection. Conversely, the negative questions elicit language of distress, providing a comprehensive picture of the student's psychological landscape. This dual approach ensures that the corpus is not solely focused on pathology but also captures the spectrum of human experience, including coping mechanisms and positive psychological traits.

Structural Ambiguity and the Individualistic Trap

A critical insight emerging from the synthesis of media framing and corpus data is the persistent ambiguity surrounding the role of institutions. In the Kumparan study, the campus is described as a paradoxical entity. On one hand, it is the source of "never-ending assignments" and "busy schedules," creating a high-stress environment. On the other, it is portrayed as the provider of "psychological support" and "psychoeducation." This dual representation creates a confusing narrative where the institution is both the problem and the solution.

This ambiguity often leads to an "individualistic trap." The media framing tends to present student mental health as a personal failure to cope, rather than a consequence of the structural pressures of the academic environment. When news articles discuss "burnout" and "stress," the language focuses on the student's reaction rather than the systemic causes. This framing can inadvertently victim-blame, suggesting that students who struggle are simply lacking in resilience or time management skills.

The implications of this individualistic framing are significant for policy and practice. If the problem is defined as individual, the solution is expected to be individual. This limits the scope of intervention to personal counseling or self-help strategies, potentially neglecting necessary structural reforms such as reducing assignment loads, improving faculty-student communication, or redesigning academic calendars. The corpus data from the media analysis supports this: the frequency of words like "stress" and "pressure" is high, but these are rarely linked to "university policy" or "academic reform." Instead, they are linked to the student's "schedule" and "feelings."

The PsihoRo corpus provides a counter-narrative by focusing on the individual's internal experience through open-ended questions. However, even this method, while ethically sound, relies on the individual's self-reporting. The integration of PHQ-9 and GAD-7 scores with text analysis helps validate the self-reports, but it does not directly address the external structural factors. The challenge remains in moving the discourse from an individual focus to a systemic one.

The Role of Open-Source Resources in Global Mental Health Research

The creation of open-source corpora like PsihoRo represents a vital step toward democratizing mental health research, particularly for non-English speaking populations. The lack of such resources in languages like Romanian has historically limited the ability to conduct robust NLP studies on mental health in these regions. By providing a publicly available dataset, the PsihoRo project enables researchers worldwide to analyze the linguistic markers of depression and anxiety in the Romanian language.

The methodology of using open-ended questions combined with standardized screening tools sets a new standard for data collection. This approach overcomes the ethical and quality issues associated with social media mining. Social media data often lacks context and consent, leading to "suppositions made by the collectors." The PsihoRo method ensures that every data point is accompanied by informed consent and a clinical baseline, making the data highly reliable for training machine learning models or conducting discourse analysis.

The availability of such corpora is crucial for developing localized mental health interventions. Language is deeply cultural, and linguistic markers of distress in one culture may not translate directly to another. For example, the specific words or phrases Romanian students use to describe anxiety might differ significantly from those used by American or Indonesian students. Having a dedicated corpus allows for the development of culturally sensitive NLP tools and diagnostic aids.

Furthermore, these resources facilitate cross-cultural comparisons. By comparing the Kumparan media analysis (focused on Indonesian student stress) with the PsihoRo clinical corpus (focused on Romanian individual distress), researchers can identify universal patterns of mental health language as well as culture-specific nuances. This comparative analysis can inform global mental health strategies, ensuring that interventions are tailored to specific linguistic and cultural contexts.

Synthesis: From Data to Actionable Insights

The convergence of media framing analysis and clinical corpus construction offers a powerful toolkit for addressing student mental health crises. The Kumparan study highlights the need to shift the media narrative from an individualistic focus to a structural one. If the media continues to frame student mental health as a personal failing, it perpetuates stigma and ignores the institutional pressures that drive the crisis. The data shows that students are represented as vulnerable, but the media rarely holds institutions accountable for creating the conditions of "never-ending assignments" or "busy schedules" that lead to burnout.

Simultaneously, the PsihoRo corpus provides a robust, ethical method for gathering and analyzing mental health data. By combining open-ended narratives with PHQ-9 and GAD-7 scores, researchers can identify precise linguistic signatures of depression and anxiety. This data is essential for developing AI-driven diagnostic tools and for understanding the nuanced ways students express their distress. The integration of these two streams of research—media discourse analysis and clinical text mining—creates a comprehensive picture of the student mental health landscape.

The actionable insights derived from these studies point toward a necessary shift in how mental health is discussed and addressed. First, there is an urgent need to reframe the public discourse. Instead of asking "How can the student cope better?", the question should be "How can the institution reduce the structural causes of stress?" Second, the development of open-source corpora enables the creation of sophisticated, language-specific tools for early detection and intervention.

The following table summarizes the key comparative insights derived from the two primary studies:

Conclusion

The analysis of text corpora related to student mental health reveals a complex interplay between media representation and clinical reality. The media coverage of student mental health predominantly adopts an individualistic frame, portraying students as vulnerable victims of their own inability to cope, while ambiguously positioning educational institutions as both the cause of stress and the potential savior. This framing obscures the structural nature of the problem, potentially hindering effective policy solutions. Conversely, the creation of clinical text corpora, such as the PsihoRo project, offers a rigorous, ethically sound methodology for capturing the lived experience of mental illness. By combining open-ended narrative responses with standardized clinical screening tools, researchers can extract valuable linguistic markers that correlate directly with depression and anxiety.

The synthesis of these two research streams underscores the critical need to move beyond individualistic narratives. To effectively support student mental health, the discourse must evolve to recognize and address the systemic pressures of academic life. Simultaneously, the development of open-source, language-specific corpora provides the empirical foundation necessary for developing culturally responsive diagnostic tools and interventions. The integration of media framing analysis and clinical text mining offers a powerful lens through which to understand the multifaceted nature of student mental health, paving the way for more effective, evidence-based support systems. As the volume of digital communication grows, the ability to accurately analyze this text data will be paramount for shaping policies that protect student well-being and foster resilient learning environments.

Sources

1. Representation of Student Mental Health in Online Media: Linguistic Corpus Analysis Based on Framing and Critical Discourse
2. PsihoRo: Depression and Anxiety Romanian Text Corpus