From leadership research to district action: Trust dynamics in social media crises: a topic–aspect–sentiment analysis of public evaluations during a school food safety incident

AI-assisted editorial implementation guide. Proposed district actions are separate from the original authors’ findings and recommendations. Independent peer review of this guide is not asserted.

Read source Search paper agent CC BY 4.0

Evidence and transfer

Education-related source: verify which participants, institutions and outcomes were actually studied before district adoption. Source publication type: research-article. License and identity checks do not establish study quality or causality.

Original authors’ abstract, reproduced under the source license:

School food safety crises often trigger intense discussion on social media and raise public concerns about institutional governance and public accountability. Yet existing studies have primarily focused on overall sentiment trends, providing limited insight into how public evaluations differ across specific governance issues or how sentiment is expressed toward these issues. Taking the 2024 spoiled meat incident at a middle school in Yunnan Province as a case, this study analyzes 11,016 Weibo posts and comments through a Topic–Aspect–Sentiment framework that combines LDA topic modeling, aspect extraction, and manual sentiment coding. The analysis examines how public evaluations were distributed across governance issues and how they evolved over the course of the crisis. The findings show that public evaluations were concentrated around governance issues related to food safety violations, response conduct, responsibility attribution, and accountability fairness rather than being evenly distributed across issues. Negative evaluations tended to be more concentrated during periods when response actions were perceived as inadequate or when responsibility was viewed as poorly assigned, with negative evaluations concentrated more heavily around these governance dimensions than around the food safety incident itself. During the remediation stage, relatively more favorable evaluations were observed in discussions surrounding regulatory interventions and corrective measures, although negative evaluations remained predominant overall during the observation period. In the later stages of the crisis, sentiment toward procedural legitimacy and accountability outcomes became relatively more balanced, reflecting continued public discussion concerning whether governance problems were being addressed through visible and credible procedures. By linking sentiment analysis to issue-specific evaluations, this study offers a more detailed account of how public evaluative patterns vary across governance issues in high-responsibility crises. The Topic–Aspect–Sentiment framework offers a practical approach for identifying evaluative signals in social media discussions and for examining how public evaluations shift across different issues and stages of a crisis. The findings also provide practical implications for information disclosure, accountability communication, response management, and long-term governance strategies in school food safety crises.

Study methods

Despite substantial advances in understanding the institutional accountability dimensions of school food safety crises, public sentiment expression on social media, and the role of such expression in shaping institutional trustworthiness judgments, gaps remain in explaining how public sentiment around school food safety crises contributes to evaluations of institutional trustworthiness. First, although prior research has established the theoretical framing of school food safety crises as preventable crises, little systematic attention has been paid to how publics develop differentiated patterns of sentiment concentration around distinct governance issues in social media contexts. Existing discussions of attribution pressure have largely remained at the macro level, with limited examination of issue-specific sentiment structures within individual crisis events. Second, although crisis communication scholarship recognizes the evaluative function of sentiment expression, methodological applications remain largely confined to aggregate polarity statistics, with insufficient integration of sentiment orientation and issue structure. Third, computational communication research broadly faces the limitation of unclear evaluative targets, as automated sentiment classification often struggles to capture context-specific public judgments of institutional trustworthiness. By contrast, manual coding offers greater sensitivity to context and crisis-specific discourse, thereby helping address this methodological gap and improving interpretive validity in fine-grained analyses of issue–sentiment structures. These three gaps converge on one central question: how does sentiment distribution across specific governance issues relate to public evaluations of institutional trustworthiness in school food safety crises, and how can this relationship be systematically examined within a single crisis? To address this question, the present study develops a Topic–Aspect–Sentiment framework that uses LDA to identify core topic structures while incorporating manual sentiment coding to preserve contextual interpretive depth. This design enables a systematic examination of how issue structure, sentiment distribution, and perceived institutional trustworthiness interact within a single crisis event.

Limitations

Despite substantial advances in understanding the institutional accountability dimensions of school food safety crises, public sentiment expression on social media, and the role of such expression in shaping institutional trustworthiness judgments, gaps remain in explaining how public sentiment around school food safety crises contributes to evaluations of institutional trustworthiness. First, although prior research has established the theoretical framing of school food safety crises as preventable crises, little systematic attention has been paid to how publics develop differentiated patterns of sentiment concentration around distinct governance issues in social media contexts. Existing discussions of attribution pressure have largely remained at the macro level, with limited examination of issue-specific sentiment structures within individual crisis events. Second, although crisis communication scholarship recognizes the evaluative function of sentiment expression, methodological applications remain largely confined to aggregate polarity statistics, with insufficient integration of sentiment orientation and issue structure. Third, computational communication research broadly faces the limitation of unclear evaluative targets, as automated sentiment classification often struggles to capture context-specific public judgments of institutional trustworthiness. By contrast, manual coding offers greater sensitivity to context and crisis-specific discourse, thereby helping address this methodological gap and improving interpretive validity in fine-grained analyses of issue–sentiment structures. These three gaps converge on one central question: how does sentiment distribution across specific governance issues relate to public evaluations of institutional trustworthiness in school food safety crises, and how can this relationship be systematically examined within a single crisis? To address this question, the present study develops a Topic–Aspect–Sentiment framework that uses LDA to identify core topic structures while incorporating manual sentiment coding to preserve contextual interpretive depth. This design enables a systematic examination of how issue structure, sentiment distribution, and perceived institutional trustworthiness interact within a single crisis event.

A 90-day district inquiry

Days 1–15 · Appraise. Read the source methods, recruitment, comparator, outcomes and limitations. Record how the study setting differs from your district and which claim the design can support.
Days 16–30 · Design. Choose one school improvement question with teachers. Protect a short shared planning cycle, agree an instructional change and use a common rubric for examining student work. Agree participation, privacy protections, baseline and stopping rules before starting.
Days 31–60 · Observe. Collect the pre-agreed measures, keep a log of implementation changes and invite confidential voluntary feedback. Preserve existing support for staff and learners.
Days 61–90 · Decide. Review implementation fidelity, teacher time and independently scored student work against a recorded baseline. Examine access differences and competing explanations before attributing changes to leadership. Make a dated decision to stop, revise or extend the inquiry.

Board decision brief

Present the local problem, source setting, evidence limits, intervention, responsible owner, total cost, measured changes and uncertainty. Separate staff perceptions, implementation measures and student learning outcomes.

Attribution and reuse

Yaling Bai; Jiatong Tang; Baosheng Zhang. Trust dynamics in social media crises: a topic–aspect–sentiment analysis of public evaluations during a school food safety incident. Frontiers in Psychology (2026-08-05). https://doi.org/10.3389/fpsyg.2026.1895872

Prepared 2026-11-07. Source authorship and DOI are preserved. The abstract and source passages are attributed extracts; this editorial guide is offered under the same CC BY 4.0 license.