Outrage can draw attention to a real problem. It can also make a post easy to react to before a reader has considered the evidence. A high share count alone tells us little about how an audience feels afterward.
Research on political social media offers a warning about what engagement can reward. It does not establish that every angry post loses followers, damages trust or has the same effect in a different subject.
What gets shared
Rathje and colleagues studied posts by US news outlets and members of Congress on Facebook and Twitter. Posts referring to political opponents were more likely to be shared, an observational association stronger in their data than several other language measures. The result concerns political accounts and particular platforms; it does not prove that hostility caused each share or predict the effect for a technology publication.
A separate study by Brady and colleagues combined social media observations with experiments to examine how social feedback relates to expressions of moral outrage. It suggests a reinforcing loop: attention can shape later expression. That is a study of expression and social feedback, not a measurement of subscriber loss.
What readers said after the feed
In a preregistered audit, Milli and colleagues compared Twitter’s engagement-ranked feed with a reverse-chronological baseline. The ranked feed amplified emotionally charged content hostile to political out-groups. Participants reported feeling worse about their political out-group after exposure and did not prefer the political tweets selected by the algorithm.
This is a specific result about political content and reported attitudes and preferences in the study. It does not measure whether a particular publisher lost subscribers, whether every reader felt worse, or how today’s ranking system behaves.
Engagement and satisfaction are different questions
A click or share is an action; it is not a complete account of a reader’s experience. People may share for many reasons, including criticism or a desire to warn someone. Those are possible explanations, not motives established by an individual share.
A publisher can ask two questions at once: did readers engage, and did the material help them understand something? The first is easier to count. The second requires closer attention to feedback and evidence.
Make room for justified anger
Reporting harm sometimes calls for strong language. The practical boundary is accuracy: describe what happened, link to evidence, distinguish facts from inference and give the reader relevant context. Avoid treating an opponent as a substitute for an argument.
-
Check whether a strong claim has a source and an appropriate qualification.
-
Compare reactions with visits and meaningful reader feedback, without treating correlation as proof of cause.
-
Ask what a reader can understand or do after reading, beyond reacting in the moment.
Outrage may attract attention, but attention is not the same as approval or understanding. The evidence supports examining that distinction, not a universal claim that anger always costs an audience.
Sources
AI-generated editorial illustrations do not depict study participants.
