Quantitative evaluation of student mental health policies in China: an LDA-PMC coupled approach
Abstract
IntroductionQuantitative evaluation of student mental health policies remains methodologically underdeveloped, with most existing studies relying on qualitative assessments or manually constructed indicators that introduce subjective bias.MethodsDrawing on Self-Determination Theory and the Biopsychosocial Technological Systems Theory, this study develops a data-driven evaluation framework by integrating TF-IDF keyword extraction, Latent Dirichlet Allocation (LDA) topic modeling, and the Policy Modeling Consistency (PMC) index model. A three-tier scoring system is applied to 22 national and provincial student mental health policy documents issued in China between 2001 and 2025; secondary indicators were scored on a three-tier scale (0, 0.5, or 1), primary indicators were calculated as their means, and the PMC score was computed as the sum of the ten primary indicators.ResultsThe LDA analysis identifies five latent policy topics: curriculum and faculty development, risk prevention and intervention, home-school-community collaboration, resource assurance, and evaluation and innovation; a seed-locked re-estimation with adjacent K = 4 and K = 6 solutions confirmed the robustness of this five-topic solution. These topics directly inform the construction of 10 primary and 44 secondary PMC indicators. The results indicate that overall policy quality is moderate (mean PMC score = 6.56, range: 3.17 to 8.23), with 59.1% of policies rated as excellent (PMC: 7.00 to 8.99), 18.2% as good (PMC: 5.00 to 6.99), and 22.7% as weak (PMC: 0.00 to 4.99). Policies perform well on transparency (indicator score 0.85), domain coverage (0.84), and policy nature (0.83), but exhibit consistent structural weaknesses in legal enforceability (0.36), multi-departmental coordination (0.35), and medium- to long-term planning (0.42). Heterogeneity analysis reveals three patterns: policy documents have become increasingly comprehensive and standardized from the exploratory phase (2001 to 2010) through the deepening phase (2021 to 2025); contrary to H2, provincial-level documents on average scored higher than national-level ones, a pattern largely attributable to the composition of the sample; and eastern and western regions scored similarly high, with central-region documents lower.DiscussionThese findings suggest that while China's student mental health policy framework has matured in scope and standardization, it remains constrained by low legislative hierarchy, fragmented governance, and insufficient institutional support for family-school-community coordination.
Related
- AI-based detection of suicidal ideation in text: model development and evaluation for a student mental health chatbot
- Predictors of engagement in a hub-based social prescribing programme for children and young people with mental health difficulties: a quantitative service evaluation study
- The settings approach to health promotion in schools: examining the role of school policies in fostering student health literacy
- Structural characteristics and policy differences of digital health management policies in China: A quantitative analysis based on the PMC index model
- Post-policy evaluation lags of tobacco and nicotine product policies: a systematic review from the United States, Canada, Australia, and the United Kingdom
BibTeX
@article{li2026quantitative,
title = {{Quantitative evaluation of student mental health policies in China: an LDA-PMC coupled approach}},
author = {Weiwei Li and Xiaoyu Wang and Xia Liu and Baochun Zhao and Hongfeng Zhang},
journal = {Frontiers in Public Health},
year = {2026},
month = nov,
doi = {10.3389/fpubh.2026.1950713},
url = {https://www.frontiersin.org/articles/10.3389/fpubh.2026.1950713},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Quantitative evaluation of student mental health policies in China: an LDA-PMC coupled approach
AU - Weiwei Li
AU - Xiaoyu Wang
AU - Xia Liu
AU - Baochun Zhao
AU - Hongfeng Zhang
PY - 2026
DA - 2026/11/03
JO - Frontiers in Public Health
DO - 10.3389/fpubh.2026.1950713
UR - https://www.frontiersin.org/articles/10.3389/fpubh.2026.1950713
AB - IntroductionQuantitative evaluation of student mental health policies remains methodologically underdeveloped, with most existing studies relying on qualitative assessments or manually constructed indicators that introduce subjective bias.MethodsDrawing on Self-Determination Theory and the Biopsychosocial Technological Systems Theory, this study develops a data-driven evaluation framework by integrating TF-IDF keyword extraction, Latent Dirichlet Allocation (LDA) topic modeling, and the Policy Modeling Consistency (PMC) index model. A three-tier scoring system is applied to 22 national and provincial student mental health policy documents issued in China between 2001 and 2025; secondary indicators were scored on a three-tier scale (0, 0.5, or 1), primary indicators were calculated as their means, and the PMC score was computed as the sum of the ten primary indicators.ResultsThe LDA analysis identifies five latent policy topics: curriculum and faculty development, risk prevention and intervention, home-school-community collaboration, resource assurance, and evaluation and innovation; a seed-locked re-estimation with adjacent K = 4 and K = 6 solutions confirmed the robustness of this five-topic solution. These topics directly inform the construction of 10 primary and 44 secondary PMC indicators. The results indicate that overall policy quality is moderate (mean PMC score = 6.56, range: 3.17 to 8.23), with 59.1% of policies rated as excellent (PMC: 7.00 to 8.99), 18.2% as good (PMC: 5.00 to 6.99), and 22.7% as weak (PMC: 0.00 to 4.99). Policies perform well on transparency (indicator score 0.85), domain coverage (0.84), and policy nature (0.83), but exhibit consistent structural weaknesses in legal enforceability (0.36), multi-departmental coordination (0.35), and medium- to long-term planning (0.42). Heterogeneity analysis reveals three patterns: policy documents have become increasingly comprehensive and standardized from the exploratory phase (2001 to 2010) through the deepening phase (2021 to 2025); contrary to H2, provincial-level documents on average scored higher than national-level ones, a pattern largely attributable to the composition of the sample; and eastern and western regions scored similarly high, with central-region documents lower.DiscussionThese findings suggest that while China's student mental health policy framework has matured in scope and standardization, it remains constrained by low legislative hierarchy, fragmented governance, and insufficient institutional support for family-school-community coordination.
SN - 2296-2565
ER -