Two headlines can describe the same event. One says a proposal "was criticised by several groups". The other says it "sparked fury". The facts are identical; the emotional charge is not. And the emotional version is more likely to be shared.
What the research shows
In 2017, William Brady, Jay Van Bavel and colleagues published a study in the Proceedings of the National Academy of Sciences analysing hundreds of thousands of messages about contested political topics. They found that each additional moral-emotional word in a message, words like "shame", "evil" or "disgust", was associated with roughly a 20 percent increase in how often it was shared. The effect was strongest within like-minded groups.
Emotion is not the enemy of journalism. Real events are often moving, and reporting on suffering should convey it. The problem begins when emotional language replaces evidence instead of accompanying it.
Signs of emotional framing
- Loaded verbs: "slams", "destroys", "admits", "caught".
- Fear without scale: "a deadly threat", with no numbers about how common or likely it is.
- Moral labels in place of facts: people described as "traitors", "heroes" or "evil" before the evidence is presented.
- Urgency: "share before it's deleted", "they don't want you to know".
- Us and them: framing that invites you to defend a group rather than evaluate a claim.
Why it works on all of us
Strong emotions narrow attention and speed up decisions. That is useful when escaping danger, and unhelpful when deciding whether a post is accurate. Emotional framing does not need to convince your reason; it only needs to act before your reason gets involved.
A simple test
Rewrite the claim in flat, neutral language. "Minister slams report" becomes "the minister disagreed with the report". "Shocking new figures" becomes "new figures". If the claim loses most of its force, the force was coming from the framing, not the facts.
How Honest News flags it
The Distortion Tracker marks "added spin", meaning alarmist verbs, unattributed certainty and speculative conclusions, and names the manipulation tactics it finds, such as False Urgency or Selective Omission. It even does this when you have no original source to compare with: emotional framing is visible in the text itself.
Source: Brady, W. J., Wills, J. A., Jost, J. T., Tucker, J. A., & Van Bavel, J. J. (2017). Emotion shapes the diffusion of moralized content in social networks. PNAS, 114(28), 7313–7318.