| 其他摘要 | The COVID-19 pandemic, a global public health crisis, has not only caused physical ailments among infected college students but has also profoundly affected the psychological well-being of both infected and uninfected student populations—an impact that may linger into the post-pandemic era. In light of the psychological health threats posed by such emergent public health challenges, the implementation of spatial risk management measures for timely interventions emerges as a crucial strategy for mitigating risks. However, existing research indicates that spatial risk assessment models designed to curtail the spread of the pandemic do not consistently mirror psychological health risks. In the era of digital proliferation, the role of media exposure related to the pandemic emerges as a significant determinant of the psychological health risks confronting college students during this crisis. Traditional research on media exposure often treats the consumption of epidemic-related media as a monolithic body, thus potentially overlooking the nuanced, bidirectional effects that various types of epidemic information can have on psychological health, both ameliorative and detrimental. Consequently, this study endeavors to elucidate the correlation between different COVID-19 risk assessment criteria and psychological health risks, delineate the variances in psychological health risks across media exposure subgroups under distinct risk assessment conditions, and unpack the influence of various types of media exposure content on psychological health during the pandemic. Such insights aim to refine our understanding of, and response to, the psychological health challenges posed by public health crises.
This study comprises three distinct investigations. The first study aims to examine whether two different risk assessment criteria under the COVID-19 pandemic align with the psychological health risks of college students. Utilizing a convenience sampling approach, an online survey was administered to 77,197 students across 177 universities nationwide. This survey collected demographic and mental health status data. The results showed that the risk assessment using confirmed COVID-19 cases pinpointed the highest instances of depression and anxiety within the most critical areas, while paradoxically uncovering the lowest levels of these mental health concerns in areas deemed to be of the second-highest risk, thereby demonstrating a distinct ‘marginal zone effect’. A similar pattern was noted in the assessment based on geographic distance, suggesting discrepancies with actual psychological health risks.
The second study investigates the differences in psychological health risks among different media exposure subgroups under spatial risk assessment criteria during the pandemic, aiming to understand whether various risk assessment criteria can identify the association between media exposure and psychological health risks. Latent profile analysis revealed potential categorizations among college students’ media exposure: ‘low media exposure group’, ‘medium media exposure group’, and ‘high media exposure group’. Resilience differed among media exposure subgroups in second-highest risk areas under both risk categorizations. Significant differences in depression and anxiety levels were found among media exposure subgroups in low-risk areas defined by the number of confirmed cases. In contrast, no significant associations in depression and anxiety were observed among media exposure subgroups in geographic distance-based groups, pointing to a lesser recognition by geographic space-based risk categorization of the correlation between media exposure and psychological health risks.
The third segment of the study explored depression and anxiety levels across various COVID-19 risk categories using advanced machine learning methods, including polynomial logistic regression, random forest, support vector machine, and LightGBM. The LightGBM models proved superior in predicting mental health outcomes. The analysis also examined media exposure impact across these risk zones. Findings indicated that in high-risk areas, depression and anxiety were negatively impacted by epidemiological news and the tone of online discourse. However, local news coverage showed a positive effect on anxiety, whereas it negatively correlated with depression in these zones. In contrast, sub-high risk areas saw no significant media effect on depression, but positive online public opinions slightly raised anxiety levels. For sub-low risk groups, negative impacts were observed from online opinion on both mental states, while local news and public health information had a beneficial effect. Interestingly, authoritative commentary tended to reduce anxiety. In the lowest risk areas, media had little impact on depression, but coverage related to the epidemic was linked to increased anxiety. The study underscores the varied influence of media on mental health, with significant effects in high and sub-low risk areas and minimal impact on those in sub-high and low-risk zones.
Through these three studies, this research examines the consistency between two spatial risk categorizations and psychological health risks under COVID-19, and through latent profile analysis and machine learning algorithms, analyzes the effect of media exposure on psychological health risks. This helps us understand the effectiveness of different risk categorizations in identifying the association between media exposure and psychological health risks during emergent public events, offering theoretical significance for supplementing media exposure risk research and providing crucial guidance for media dissemination practices in emergency management measures for psychological interventions during public health crises. |
修改评论