『人民日報』における報道内容の変容

1950年から2020年を対象とした計量テキスト分析

抄録

This article examined the changes in news articles published in People’s Daily from 1950 to 2020 using quantitative text analysis. A semi-supervised learning approach was employed to analyze 24,896 articles about the Chinese Communist Party (CCP). First, we extracted articles related to China based on R package Newsmap. We then set the seed words and used Latent Semantic Scaling to predict the value of each article. The analysis revealed that dogmatic propaganda about the CCP continues to be published in People’s Daily news articles. However, articles about national development had increased since 1989. Furthermore, while self-criticism was common in articles about corruption, the ratio of articles focusing on public opinion supervision had increased since 2003. We concluded that the CCP has changed its method of securing its legitimacy through People’s Daily.