Gabbard, Booker and Biden get most negative media coverage over last four months
Tulsi Gabbard, Cory Booker and Joe Biden are the 2020 Democratic candidates receiving the most negative media coverage, according to updated findings from the Storybench Election Coverage Tracker at Northeastern University School of Journalism.
An analysis of 5,658 articles published across 28 outlets between June 1 and September 28 shows that media coverage paints Pete Buttigieg, Amy Klobuchar and Elizabeth Warren in the most positive light.
Breaking this out by media bias, which we scored according to Media Bias/Fact Check, we see that right-leaning media organizations consistently treat all candidates in a more negative light.
In both April and June, we found a consistent pattern of media outlets covering female candidates more negatively than their male counterparts. That seems to have broken over the last four months – in terms of overall coverage.
But looking at how individual news organizations cover these select 2020 candidates, we found that many tended to cover the men, on average, more positively than the women – with the exception of The Washington Post, The Atlantic, the Huffington Post and Reuters, most notably.
We intend to continue following media coverage of the 2020 Democratic candidates as their campaigns mature.
Methodology: This analysis was based on articles published by media outlets – including CNN, The New York Times, BuzzFeed, The Daily Caller, Breitbart and Fox News – mentioning one of the following 2020 candidates in the headline: Joe Biden, Cory Booker, Pete Buttigieg, Kamala Harris, Amy Klobuchar, Tulsi Gabbard, Bernie Sanders or Elizabeth Warren. After scraping article text with Python’s newspaper3k package, we used R’s tidytext package to tokenize, remove stop words, and apply sentiment analysis. We scored the full text of news articles about each candidate against a dictionary containing 10,222 words assigned an average positivity score between 1.44 and 8.50. The “score of positivity” we calculate for each candidate is an average of all the scorable words that have been written about them in our dataset.
