Claim
The incentive is engagement, and anger engages
An advertising funded platform is paid for attention, so its ranking systems are tuned to whatever keeps people scrolling and responding. Outrage does that better than agreement, and better than nuance, because an angry man replies, shares and comes back to check the replies.
Nobody needs to intend this. A system that optimizes for response will discover outrage on its own, the way a river finds the lowest ground. That is why the same mechanism shows up on platforms with completely different politics, and why the content that reaches you has usually been selected for how it makes you feel rather than for whether it is true.
Evidence
Moral emotion measurably spreads further
William Brady and colleagues published a study in PNAS in 2017 analyzing several hundred thousand messages on social media about contested political topics. Each additional moral or emotional word in a message was associated with roughly a 20% increase in its rate of sharing.
They also found the spread was concentrated inside ideological groups rather than across them, which is the part worth sitting with: the material that travels furthest is the material that confirms what a group already believes about its opponents. Molly Crockett made the related argument in Nature Human Behavior the same year, describing how digital environments lower the cost of expressing outrage and raise its rewards.
Claim
Learn the physical tell
The useful signal is in your body before it is in your reasoning. A tightening in the chest, a hot feeling behind the eyes, the urge to read the replies rather than the article, the sentence forming in your head that you want to say to someone.
That state is the product working as designed. It is not evidence about the world and it is not evidence that the thing is important. Naming it while it is happening is most of the defense, because outrage depends on being experienced as urgency rather than recognized as a state you are in.
Action
Run the four-question check on the next thing that angers you
Before you share or reply. Who benefits if I am angry about this. Is this the most extreme example of a category, and how common is the category, which is day 3. What would I need to see to know it is true, which is day 1. And will this matter in a week.
Then apply the twenty-four hour rule: if it still seems worth responding to tomorrow, respond tomorrow. Most of it does not survive the night. Nothing you gain by replying in the first hour is worth what you spend, and the reply you write in that hour is the one you will regret.
Example
Duncan noticed he was having the same evening twice a week: forty minutes on his phone, a raised heart rate, and an argument with a stranger under a post.
He made one change. He moved to reading a single long article each morning, from an outlet whose funding he had actually checked, and removed the two apps where the arguments happened. His description six weeks later was that he felt less informed for about ten days and then realized he knew more, because he was finishing pieces instead of reacting to fragments of them. He also says he has no idea what anyone he went to school with thinks about anything now, and that this has improved his life.
Caution
Disengagement is not the same as judgment
The cheap version of this day is to disconnect from everything and treat that as wisdom. It is not: some things genuinely warrant anger, and a man who has trained himself out of all response has made himself useless in the situations where response is the only thing that helps.
The goal is not calm, it is directed attention. Pick two or three issues you will actually follow properly, with primary sources and base rates, and be honestly uninformed about the rest. That is a defensible position and it beats both the man who is furious about everything and the man who has opted out and calls it detachment.
Aside
The argument that this framing is convenient
There is a real objection: telling people their political anger is an artefact of a recommendation system is a comfortable way to avoid the content of the anger. Sometimes a man is angry because something is genuinely wrong, and explaining the mechanism of outrage to him is a way of not engaging with it.
The distinction I would defend is between the object and the timing. The incentive analysis explains why you encountered this thing now, at this intensity, in this order. It says nothing about whether the thing is bad. So use it to choose what to pay attention to, and never as an argument against a claim. If someone tells you your concern is just the algorithm, that is not a rebuttal either.