Computable Structures of National Narratives: A Dataset for Generating Governance Legitimacy Models Based on Computational Content Analysis, Emotional Mediation, and Semantic Networks

This dataset supports computational modelling research on national-level political narratives, employing the Computable Structure of National Narrative (CSNN) as its theoretical framework. It systematically demonstrates how national narratives generate governance legitimacy and social cohesion through structured content configuration, affective mediation mechanisms, and semantic network coupling. The dataset is constructed from publicly available national-level political texts, processed through computational content analysis, sentiment analysis, and semantic network analysis. This yields a multi-layered variable system encompassing narrative input variables (such as development, people's livelihoods, culture, discipline, and external governance narratives), mediating variables (emotional resonance, policy perceptibility, governance credibility), latent variables (national confidence, institutional trust), and outcome variables (governance legitimacy, social cohesion), with variable relationships explicitly defined through directed causal pathways. This dataset emphasises theory-driven, mechanism-oriented and reproducible methodologies, serving computational social science, political communication, governance studies, Text-as-Data approaches, and causal inference teaching, replication and extension research. All variable definitions, path specifications and data processing logic are explicitly documented in accompanying explanatory files, facilitating comparative analysis across texts, years and nations.本数据集以国家叙事可计算结构（Computable Structure of National Narrative, CSNN）为理论框架，服务于国家级政治叙事的计算建模研究。本数据集系统阐释了国家叙事如何通过结构化内容配置、情感中介机制与语义网络耦合，生成治理合法性与社会凝聚力。本数据集的数据源为公开可得的国家级政治文本，经计算内容分析、情感分析与语义网络分析流程处理后得到。由此构建出多层级变量体系，涵盖叙事输入变量（如发展叙事、民生叙事、文化叙事、纪律叙事与外部治理叙事）、中介变量（情感共鸣、政策感知度、治理公信力）、潜变量（国家信心、制度信任）与结果变量（治理合法性、社会凝聚力），且变量间的关系通过有向因果路径予以明确定义。本数据集强调以理论为驱动、以机制为导向且可复现的研究方法，可服务于计算社会科学、政治传播学、治理研究、文本即数据（Text-as-Data）方法以及因果推断的教学、复现与拓展研究。所有变量定义、路径规范与数据处理逻辑均在配套说明文件中予以明确记载，便于开展跨文本、跨年份与跨国别的比较分析。