In the wake of the Black Lives Matter (BLM) protests emerging internationally following the murder of George Floyd, I was interested to see how political leaning influenced different newspapers’ coverage of the topic.
Based on AllSides’ Media Bias Ratings, I sorted 200 articles from 36 newspapers into five categories: far left, left, center, right, and far right. I retrieved the articles based on recent results from the search word “protests” in early- to mid-June. To see the full code, check out my GitHub.
Graphic A: Quantity of text analyzed by political leaning
By conducting a word frequency analysis in Python, I found three interesting patterns. One, regardless of political leaning, the most frequent word in the articles was “police.” Two, only the right leaning articles discuss Antifa with much frequency. Three, across all political leanings, male gender nouns and pronouns are mentioned much more frequently than their female equivalents.
Graphic B: Most Frequent Words by the political leaning of the newspaper
The above chart graphs the top eight most frequent words categorized by the political leaning of the newspapers. It shows that the word “police” is mentioned over 2,000 times and that words like “riots” and “violence” are only frequent in the right wing.
Graphic C: Frequency with which the word “antifa” appears by political leaning
Graphic C shows the frequency with which the word “antifa” appears by the political leaning of the newspaper. Even considering that the right leaning articles contained more text than those of the other categories, right and far right leaning newspapers seem to link Antifa to the protests more than other outlets.
Finally, the chart below demonstrates the frequency of male and female gender nouns and pronouns across the political spectrum. Here, no significant difference appears between the outlets, but all of them discuss men much more than women.
#media-analysis #nlp #blacklivesmatter #data analysis