Scroll for two minutes and count the tone. Argument, outrage, threat, feud, collapse, scandal. Then try to remember the last piece of content you saw that was built around people getting along, and notice how much harder that is. It is tempting to explain this as somebody's decision -- a room where the cynical choices get made. The truthful explanation is duller and more useful than that, and it has been measured repeatedly: in a market where attention is the product, conflict outperforms harmony, so the catalog of what gets made bends toward conflict. Nobody has to intend it. The incentive does the work.
Start with the mechanism, not the motive
This site has a method post about exactly this kind of reasoning, and the discipline is worth restating. You do not need a secret room or a coordinated plot to explain most of what shows up here. You need incentives, arithmetic, and enough time for both to compound. A producer choosing between two pitches, an editor choosing a headline, a musician's label choosing a single -- none of them has to be cynical. Each is responding, reasonably, to what the last thing that worked looked like. Do that ten thousand times and you get a catalog nobody designed, that nonetheless has a very clear shape.
The numbers, because this part is measured
This is not a vibe. It has been tested in ways you can check. In a series of randomized trials covering roughly 105,000 variations of news stories that generated 5.7 million clicks, each additional negative word in a headline raised the click-through rate by about 2.3 percent, while positive words lowered it -- and this held even though positive words were slightly more common overall [1]. A separate 17-country study across six continents measured people's physiological responses to real video news and found reactions to negative content were stronger on average, which is a plausible reason audience-seeking news skews the way it does worldwide [2].
- Each extra negative word in a headline: about 2.3 percent more clicks.
- Positive words in a headline: fewer clicks.
- Stronger physiological reactions to negative news, replicated across 17 countries.
- None of that requires anyone to intend a negative product. It just describes what the market pays for.
The same thing happens to arguments, not just news
It runs through social platforms too, and specifically through moral language -- the vocabulary of outrage and blame. In an analysis of 563,312 messages about polarizing issues, each additional moral-emotional word was associated with roughly 20 percent more sharing, and that extra spread stayed largely inside political groups rather than crossing between them [3]. A later pre-registered replication of 849,266 posts, plus a meta-analysis of 27 studies, put the effect at about 13 percent per word -- real, robust, and meaningfully smaller than the original figure [4]. Both numbers belong here. Quoting only the larger one would be doing the thing this post is complaining about.
Why unity is a harder sell, structurally
Consider what an attention market can actually work with. Conflict has built-in structure: two sides, stakes, an unresolved question, a reason to come back tomorrow. It generates comment, which generates more distribution. Harmony has none of that machinery. A story about people understanding each other resolves, and a resolved story gives you nowhere to go. So it is not that unity is unpopular -- people say they want more of it, and mean it. It is that unity is a worse fit for a business model that gets paid by the click and the argument. That difference is structural, and it explains the imbalance without anyone needing to be a villain.
The concession that matters -- this is not an argument against sad songs
Be clear about the limit here, because the sloppy version of this argument is genuinely wrong. There is no doubt that acknowledging our feelings in certain songs and lyrics helps us process things. Music that names grief, anger or heartbreak is doing something valuable and old, and a person going through something hard is often served better by a song that tells the truth than by one that insists everything is fine. The whole point of the companion post on emotional vocabulary is that naming a hard feeling helps. So the objection is not to darkness in art. The objection is to a diet -- to what happens when the proportion is set by what holds attention rather than by what a person needs, and the same note is served at every meal.
The politics version is the same machine
It is worth noticing that this is not a media problem with a political exception. The same incentive runs through political communication, which is why the loudest voices in that arena are so reliably the most divisive ones, and why the reasonable version of almost any position travels so much worse than the furious version. Once you have seen the mechanism in headlines and in shares, you cannot unsee it in a campaign. It is the same market paying for the same thing.
What to do about it, which is not "stop reading the news"
The useful response is not to disengage, and it is definitely not to conclude that everything is manipulation. It is to know what you are standing in. When something reaches you at high emotional volume, the volume is information about distribution, not about importance -- it tells you the item was well shaped for travel. That is worth a beat of separation before you react or share. Beyond that, the leverage is in what you deliberately go and find, because an attention market will only ever hand you what performs. Nothing about the mechanism stops you choosing differently. It just means the choosing has to be active.
- Loud does not mean important. Loud means well shaped for travel.
- Notice the beat between feeling something and passing it on. That beat is where the leverage is.
- What you seek out on purpose is the only part of your feed the incentive does not choose.
- Judge the mechanism, not the people inside it. Most of them are responding to the same market you are.
Key Takeaways
- In an attention market, conflict outperforms harmony, so the catalog bends that way. No coordination is required -- the incentive does the work.
- Measured: across ~105,000 headline variations and 5.7 million clicks, each extra negative word raised click-through by about 2.3 percent while positive words lowered it.
- A 17-country study found physiological reactions to negative news were stronger on average, which helps explain why audience-seeking news skews negative worldwide.
- Moral-emotional language spreads further -- about 20 percent more shares per word in the original study, about 13 percent in a larger pre-registered replication and meta-analysis.
- Unity is a structurally worse fit for the business model: conflict has sides, stakes and a reason to return, while a resolved story goes nowhere.
- This is not an argument against sad songs. Acknowledging feelings in music genuinely helps us process things. The objection is to a diet set by what holds attention.
- Loud is information about distribution, not importance. The only part of your attention the incentive does not choose is what you go and find on purpose.
Sources
Powered by CellWell.life- 1.Negativity drives online news consumption (Nature Human Behaviour, 2023)
- 2.Cross-national evidence of a negativity bias in psychophysiological reactions to news (PNAS, 2019)
- 3.Emotion shapes the diffusion of moralized content in social networks (PNAS, 2017)
- 4.Estimating the effect size of moral contagion in online networks: a pre-registered replication and meta-analysis (2025)
- 5.Rethinking the Negativity Bias (2018)
Related