Blog Post 7: Sentiment Analysis
Hey everyone, and welcome back! This week, I took a new angle in my research — looking at how the tone of a news article might hint at its political bias.
I used sentiment analysis, which measures whether a piece of writing sounds positive, negative, or neutral. When you apply that to political news, the results can be pretty telling. For example, if an article sounds positive about gun rights or negative about immigration policy, that tone could reflect an underlying political perspective.
I ran the sentiment analysis on the same 15 articles I’ve been working with — one per source, per topic — and the results varied a lot depending on both the source and the subject.
Take gun control, for instance:
- Vox and NPR had very negative sentiment scores, which likely reflects criticism of gun violence or support for tighter regulation.
- The Washington Examiner, on the other hand, leaned positive, possibly highlighting personal freedoms or defending the Second Amendment.
Immigration was even more interesting:
- Vox and The Washington Examiner both scored very negatively, but probably not for the same reasons. Their negative tone might be aimed at totally different aspects of the issue. A closer read of the articles would help unpack that.
- NPR’s article stood out with a very positive tone, likely emphasizing a compassionate or reform-minded take.
All this shows that tone alone doesn’t tell the whole story, but when you pair it with context, it can definitely point toward a source’s political lean. Two articles can sound equally negative but be biased in opposite directions — it all depends on what they’re being negative about.
Next week, I’ll be performing another layer of analysis: a human-coded framing analysis to tie everything together. Stay tuned — and thanks for reading!