Do you have the sense that everything is becoming more negative? Does what you hear on the radio, and what you read in the news, seem altogether gloomier than it did ten years ago? That’s because it is – say two new studies.
The first, by David Rozado and colleagues, tracks ‘sentiment’ in U.S. news headlines since the year 2000. (We’ve covered Rozado’s work before on the Daily Sceptic.)
His team quantified ‘sentiment’ – i.e., how positive or negative the headlines were – using machine learning tools. In simple terms, they used an algorithm to identify strings conveying positive emotions (e.g., ‘good’, ‘benefit’, ‘praises’) or negative emotions (e.g., ‘bad, ‘cost’, ‘attacks’), and then assigned an overall score to each headline.
Results are given in the chart below. There are three lines: one for left-leaning outlets, one for centrist outlets, and one for right-leaning outlets. (The researchers used Allsides ratings to classify outlets.)
There is a clear shift from toward negative sentiment for both left-leaning outlets and right-leaning outlets. For centrist outlets, by contrast, the shift is much less clear.
To home-in on exactly how headlines have been changing, Rozado and colleagues looked at six different categories of emotions: anger; fear; sadness; disgust; joy; and neutral. They found that the first four have all been getting more common, while the last two have been getting less common – at least since 2010.
As to why headlines have been getting more negative, the researchers suggest it may have to do with what goes ‘viral’. Studies have found that statements with strong emotional valence – particularly those that convey animus towards an outgroup – spread fastest on social media. Outlets are presumably tweaking headlines with this in mind.
The second study, by Charlotte Brand and colleagues, tracked ‘sentiment’ in pop-song lyrics since the 1970s. One of their key findings is shown in the image below.
It indicates that “love” has decreased as a proportion of all song lyrics, while “hate” has correspondingly increased. These are just examples: the researchers found a “substantial decrease” in the use of many positive emotion-related words (and a corresponding increase in the use of negative words).
Interestingly, when they ran models predicting whether a particular song lyric was negative, they found that chart position mattered: negative lyrics were more common in songs that reached positions in the chart (i.e., closer to ‘Number 1’).
It’s unclear exactly why this would be the case. One possible explanation is the rise of rap and hip hop in the 1990s, which gave rise to more songs about themes like gang violence and breaking up. Though this is speculative.
What the two studies do suggest is that your average American is now exposed to more negative emotional content in his daily life – assuming he reads the news and listens to an occasional pop song. Faced with this finding, it’s worth reminding yourself: things could always be worse.
Profanity and abuse will be removed and may lead to a permanent ban.