Political News Analysis: A New Project Exploring the Sentiment Behind Mainstream Media News Coverage

Two weeks ago, I started a new project intended to estimate how negative – or positive – political news reporting is from major media outlets. To accomplish this, I’m performing a daily sentiment analysis on news articles scraped from, at the moment, two sources: CNN’s politics RSS feed and Fox News’ politics RSS feed.

You can check out the results as the come in each day by clicking on the “Political News Analysis” tab at the top of the blog or by clicking on the link in the sidebar.

So far, as of March 19, 2017, my analysis suggests that both CNN and Fox News generally use more negative words than positive words in their political news coverage. Positive words include, for example, terms like “support,” “good,” and “well,” and negative words include, for example, terms like “allegations,” “concerns,” and “fears.”

Yet although both media outlets show a greater tendency to use negative language, CNN seems to exhibit slightly more negativity than Fox News.

During the past seven days – from Monday March 13, 2017 to Sunday March 19, 2017 – the average political news story at CNN contained 2.47 more negative words  than positive words. Meanwhile over at Fox News, the average political news story contained only 1.88 more negative words than positive words.


What this Analysis Does (and Does Not) Tell Us About Political News Coverage

So if CNN generally uses more negative language in its political news coverage than Fox News, does this mean CNN treats the Trump administration more unfairly than other media outlets? In a word, no.

Let’s take a look at some important caveats and limitations you should keep in mind regarding this analysis.

First, the findings of my “Political News Analysis” are most certainly not a measure of political bias among these major news organizations. So, don’t look at the numbers above as evidence that, for example, CNN is more – or less – antagonistic towards the Trump administration than is Fox News. Based solely on the findings, there’s simply no way at all to know how political partisanship maps onto the average sentiment conveyed by a major news organization. Indeed, it could cut both ways. For instance, on a day when a new scandal rocks the White House, an extremely biased conservative media outlet might use negative language to try to quiet and discredit critics of the President. Alternatively, such an organization might use positive language to express support for the administration or to echo Republican talking points designed to calm fears among the general public.

Secondly, these results should not be viewed as a measure of the media’s bias towards using negative language to cover otherwise neutral political events and news stories. That’s because there’s no guarantee that each media outlet was reporting about precisely the same news stories at a given time on a given day. Moreover, there are at least two reasons why a media outlet might use a lot of negative language to cover an event – either because the news organization is indeed biased to cover stories in a negative light or, more simply, because the majority of stories being covered by that news organization are, themselves, inherently negative and unpleasant and, therefore, deserving of negative language.

Third, and finally, we have to remember to consider the context in which language is used, because sometimes the same words can be used to describe both positive and negative events, as in “He was killed” vs. “He was not killed.” These two sentences contain the same exact negative word, “killed,” even though a very different sentiment is conveyed in each case. In an analysis such as this one, which focuses solely on the number of positive words relative to the number of negative words in political news articles, the context in which words appear is entirely stripped away. Therefore, the present results should not be viewed as anything more than a mere estimate of the average sentiment conveyed by major media outlets.

So, if the findings from my “Political News Analysis” do not necessarily reflect partisan bias or the media’s tendency to use negative language to cover otherwise neutral stories and events, then what do they reflect? What precisely does it mean to say, for example, that CNN and Fox News use about two or three more negative words than positive words in any given news story they publish about politics?

Two possible interpretations come to mind.

First, excessive use of negative language in political news coverage might reflect a bias on the part of the news media to report on negative political events (a form of the old trope, “If it bleeds, it leads.”). If so, then it seems both CNN and Fox News are inclined to report on negative political stories over positive ones, with CNN being be ever so slightly – and perhaps ultimately negligibly – more inclined to do so than Fox News.

Alternatively, the news media might not be biased to report on negative political events at all. Instead, given the hyper-polarized and adversarial nature of American politics today, perhaps it’s simply that most political news stories are inherently negative to some degree. As such, maybe there’s just not that many “positive” political news stories these days for media outlets to cover.


Brian Kurilla is a psychological scientist with a Ph.D. in cognitive psychology. You can follow Brian on Twitter @briankurilla 

Leave a Reply

Your email address will not be published. Required fields are marked *