The linguistic landscape of “controversial”: sentiment and theme distribution insights

• Main Author: Grishechko, Elizaveta G.
• Format: Article
• Language: English
• Published: Penerbit Universiti Kebangsaan Malaysia 2024
• Online Access: http://journalarticle.ukm.my/23581/1/Gema%20Online_24_1_5.pdf; http://journalarticle.ukm.my/23581/; https://ejournal.ukm.my/gema/issue/view/1648

The language used to frame controversial topics on social media has profound implications for public discourse and opinion formation, warranting a close examination of their sentiment and thematic distribution. This study investigates the sentiment and themes associated with controversial topics by analyzing Reddit posts containing the token “controversial” in their titles on three news-related subreddits, aiming to bridge a gap in existing literature by focusing on platform-specific sentiment analysis with an emphasis on content typology. A mixed-methods NLP approach instrumented via Python was employed, combining VADER-supported sentiment analysis and a qualitative content analysis using n-grams to identify and categorize themes. The sentiment analysis results indicated that most of the content had neutral sentiment, which testifies to the predominantly fact-based approach to presenting information with lack of strong emotional connotations. However, the overall compound sentiment scores were negative, which suggests a strong negative undertone in the framing of controversial topics. The theme distribution analysis revealed that Politics and Legislation was the most predominant theme, followed by Technology and Surveillance, Social Issues and Controversies, Health and Medicine, and Environment and Energy. This distribution attests to a range of societal issues that generate controversy on social media platforms. Study findings can be used by content creators and social media analysts to track online content sentiment, guide content moderation practices, and improve audience engagement. By demonstrating the potential of NLP techniques, this study also contributes to the fields of media research and language technology, which can encourage a better scholarly evaluation of online discourse.