Unveiling Media Bias: A Comprehensive Analysis of News Coverage on Trump and Biden
Media bias has long been a topic of debate, with concerns about how journalists and news outlets shape narratives around political figures. In recent years, the discussion has intensified, prompting us to conduct a thorough examination of media bias by delving into a random sample of news articles covering Trump and Biden from 2016 onwards. This study, conducted using our extensive news media database, also involved analysing over 100,000 comments associated with those news media articles to gain deeper insights.
Sentiment analysis, a powerful tool in social media research, employs natural language processing (NLP) and machine learning techniques to discern the emotional tone within written text. By applying our proprietary sentiment sampling techniques to news articles and their associated comments, we aimed to unravel the underlying bias in media coverage.
Positive Sentiment: Reflects a favorable or positive opinion, emotion, or attitude. Values greater than 0 indicate positivity.
Negative Sentiment: Conveys an unfavorable or pessimistic opinion. Values less than 0 suggest negativity.
Neutral Sentiment: Signifies a lack of strong positive or negative emotion, with values close to 0.
Key Findings:
Our analysis, visualized in a Power BI dashboard, highlighted intriguing patterns. In Figure 1, comparing Article vs. Commenter Sentiment, we observed that commenter sentiment was consistently negative (around -0.05) for both Biden and Trump articles indicating little bias between commentaries. However, the striking revelation was that news article sentiment was over ten times more negative when covering Trump than Biden, indicating a significant bias in media coverage favouring Biden.
Figure 2, breaking down sentiment by news organisation, provided additional insights. Fox News, often considered right-wing, showed slightly more positive articles towards Biden despite commenters expressing more favourable attitudes towards Trump. Conversely, left-wing CNBC exhibited a pronounced bias in both articles and comments towards Biden over Trump. However, the overall trend across various news organizations, was a consistently more favourable portrayal of Biden compared to Trump in news articles.
Conclusion:
This unbiased analysis, driven by neutral machine learning algorithms, unveils a noteworthy discrepancy in news coverage. Despite similar sentiment and tone in reader commentaries for both Trump and Biden, news articles consistently favoured Biden. Surprisingly, the bias was not only evident in traditionally right-wing outlets but also observed in left-wing media, challenging preconceived notions. As we navigate through this era of information, it becomes imperative to critically assess media narratives and strive for a more balanced discourse.
You can see the dashboard itself here: instatdata.com/newsmedia
