A Dictionary-Based Textual Analysis of Domestic Support for Government Leaders

• In Event: Political Methodology Virtual Poster Session

Abstract

The purpose of this paper is to construct a database that uses a dictionary-based textual analysis of newspaper articles to measure domestic support of government leadership. The increasingly significant changes in the political and social environment worldwide have attracted much attention to the rise of populism, nationalism, and anti-globalization. Domestic citizens responses to the external and internal changes shape the international community. To understand the political consequences of the growing strong nationalist leaders and populists across countries, it is necessary to develop a high-quality and accessible database of domestic citizens attitudes towards government leadership, as measured by popular social media.

Traditional studies of public opinions of government leaders have mainly concentrated on survey-based approaches. However, survey studies are expensive, time-consuming, and rely on self-report measures. Additionally, data quality may be affected by variations in respondent motivation (e.g., compensation, forced participation). Therefore, an alternative data source is necessary. To fill this gap, I propose a framework of textual analysis to measure domestic popular support of government leadership.

Firstly, I demonstrate how to extract text information from social media. Next, I identify a framework to construct a reusable dictionary for text mining. Then, I employ a dictionary-based approach to analyze the texts and build a database using news articles, from the New York Times for the US and the South China Morning Post for China. Finally, I will conclude this paper with some important implications of this approach and a discussion of the challenges of textual analysis.

This project makes significant contributions to the existing literature. Firstly, to meet the increasing demand for text mining, this reusable dictionary is conducive to analyzing unstructured data of large-scale texts for researchers. Additionally, the database in this project can not only be used to illustrate information about public attitudes towards the government, but also examine the relationship between public opinion and other political and social issues. In future studies, I will extend this database to incorporate domestic popular attitudes to the government in other countries.

Author

• Jingjing An, Claremont Graduate University