Last Updated on 31 minutes ago by TodayWhy Editorial
Open any political social media account, watch a campaign ad, or read a fundraising email, and the tone is hard to miss. The other side isn’t just wrong. It’s dangerous, corrupt, an existential threat. This is not your imagination and it is not limited to one party or one country.
So why is political messaging becoming more extreme? Political scientists have been studying this for years, and the honest answer is that no single cause explains it. It’s a feedback loop with several moving parts — some involving the media environment, some involving the incentives politicians face, and some involving psychology that has nothing to do with technology at all. A few popular explanations turn out to be weaker than commonly assumed, and a few less obvious ones turn out to be doing more work than people expect.
Why political messaging is becoming more extreme: the two kinds of polarization
Researchers split this question into two separate phenomena, and mixing them up is where a lot of public debate goes wrong.
Ideological polarization is a gap in policy positions — how far left or right people actually stand on issues like taxes or immigration. Affective polarization is something different: how much people simply dislike, distrust, and feel contempt for the other side, independent of specific policy disagreement.
The data is much clearer on the second kind. Polling from Pew and YouGov shows Democrats’ negative views of Republicans — seeing them as immoral, for instance — rose roughly 18 to 28 percentage points in categories measured since 2016, with Republicans’ negative views of Democrats rising by a similar margin. People aren’t necessarily moving further apart on policy. They are increasingly convinced the other side is bad, not just wrong.
That distinction matters for the question of why political messaging is becoming more extreme, because it points toward a specific driver: messaging built to provoke contempt works even when the underlying policy gap hasn’t widened nearly as much.
There’s a related finding worth sitting with. A 2022 study covered by Knowable Magazine surveyed over 13,000 Americans and found that 49% believed voters from the opposing party held extreme policy positions and talked about politics constantly. The actual figure, measured directly, was 14%. Most people on “the other side” are neither as extreme nor as politically obsessed as partisans assume — but that inflated caricature is exactly the image that extreme messaging is built to reinforce.
Why politicians benefit from sounding more extreme
Here is a fact that runs against how most people think politics works: attacking, not legislating, gets the attention.
Research from Stanford’s Polarization and Social Change Lab found that inflammatory statements from members of Congress — attacking opponents’ character rather than their policies — generate outsized media coverage and social media engagement compared to substantive legislative work. An obscure lawmaker can build a national profile through a single viral, aggressive statement in a way that years of committee work never achieves.
A separate 2024 study analyzing 17 elections between 2016 and 2020 found that when political leaders escalate personal attacks, it measurably increases affective polarization among voters — and the effect is strongest among people who are already strongly partisan. The researchers describe a toxic spiral: aggressive rhetoric radicalizes the public, and a more radicalized public then rewards even more aggressive rhetoric from its leaders. Neither side has much incentive to be the first to de-escalate, because unilateral restraint looks like weakness to your own base while the other side keeps attacking.
Is political messaging becoming more extreme because of social media?
This is where public discussion tends to oversimplify in both directions.
A 2021 review by researchers at NYU’s Stern Center for Business and Human Rights, covering more than 50 academic studies, concluded that social media platforms are “likely not the root causes of political polarization” but that they meaningfully “exacerbate it.” Platform algorithms are typically built to maximize engagement, and engagement correlates strongly with emotionally charged, moralized, and outrage-triggering content — not because platforms are designed to divide people, but because that content keeps people scrolling, which is what sells advertising.
Two pieces of counter-evidence keep this claim honest, though. A 2017 study found that polarization rose most sharply between 1996 and 2016 among Americans over 65 — the demographic least likely to use social media at all. And a separate comparison of eight developed democracies over four decades found some countries saw polarization rise far less than the U.S., or even fall, despite similar social media adoption. Both findings suggest platforms are amplifying something that already had other engines running, rather than starting the fire themselves.
Why two popular explanations don’t hold up as well as people assume
This is the part of the story that surprises most people, including plenty of political commentators.
Gerrymandering is one of the most commonly cited culprits for political extremism, but the research is notably unkind to that theory. Political scientists Nolan McCarty, Keith Poole, and Howard Rosenthal found that polarization increased just as sharply in the U.S. Senate — where district lines are impossible to gerrymander, since senators represent entire states — as it did in the House. States with only one House district, which by definition cannot be gerrymandered internally, show representatives just as partisan as those from heavily redrawn districts.
Primary elections face a similar problem. A review by Protect Democracy of the existing evidence concluded that primary electorates are not meaningfully more ideologically extreme than the broader pool of people who might otherwise pick a party’s nominee — convention delegates, caucus-goers, party officials. The popular theory that a small sliver of hardcore partisan primary voters is dragging candidates to the extremes has less empirical support than its popularity in political commentary would suggest.
None of this means district maps and nomination rules are irrelevant to American politics — they clearly matter for other things, like which party wins which seats. It means they are weaker explanations for why political messaging is becoming more extreme than commonly assumed.
Why media fragmentation changed what “normal” sounds like
Before cable news proliferated and the internet arrived, most Americans got political news from a small number of sources with an incentive to appeal broadly, which meant sanding down the sharpest edges.
The end of the U.S. Fairness Doctrine in 1987, which had required broadcasters to present contrasting views on controversial issues, opened the door to partisan talk radio and, later, cable news channels built around a specific ideological audience rather than a general one. Once people could choose news that already agreed with them, media outlets had a commercial incentive to intensify that agreement rather than moderate it — the same engagement logic that later applied to social media, just with a two-decade head start.
Over time this reshaped what political messaging normally sounds like, and it’s one more reason why political messaging is becoming more extreme even without any single dramatic event driving it. A message that would have seemed shockingly aggressive to a mass audience in 1985 can now be entirely normal to an audience that has spent years inside a single ideological media ecosystem.
Why elite behavior seems to lead, not follow, public opinion
A recurring finding across this research is that elite polarization — how politicians, party officials, and media figures speak — tends to move first, with the public following rather than leading.
Research by economist Matthew Gentzkow and coauthors documented a sharp, measurable rise in explicitly partisan language in U.S. Congressional speeches since the 1990s. Separate research on social media found that posts from political elites expressing hostility toward the opposing side are disproportionately likely to go viral, which creates a direct incentive for elites to keep producing that kind of content.
This matters for why political messaging is becoming more extreme because it shifts the frame. Extreme messaging isn’t simply a reflection of a public that has organically grown more extreme and is demanding it. In large part, it looks like a supply-side phenomenon — political and media elites producing more extreme content because it performs better, with voters absorbing and then rewarding that tone over time.
Why this isn’t unique to any one country or party
It’s tempting to read this as a story about one country’s politics or one party’s behavior. Neither holds up well.
Polarization is rising across multiple democracies, not just the United States, though research consistently finds the U.S. case to be unusually severe compared to peer democracies over the same time period. And within the affective polarization research specifically, the rise in contempt for the opposing side has been documented moving in both directions — this is about the dynamics of the system, competitive elections, fragmented media, algorithmic amplification, viral incentives for elites — rather than a story where only one side’s messaging has changed.
Some academic work does examine whether the pace of change has been symmetric between parties on specific measures, and that research remains actively debated among political scientists rather than settled. Readers interested in that narrower question are better served going directly to the peer-reviewed literature than to any single summary of it, since it is one of the more contested corners of this field.
What actually makes political messaging less extreme
The research on solutions is thinner than the research on causes, but a few findings recur.
The Facebook deactivation study mentioned above found that participants who stopped using the platform for a month showed measurably reduced polarization on policy issues specifically, even though it didn’t reduce animosity based purely on party identity. That suggests exposure to algorithmically amplified content plays some causal role, distinct from party loyalty itself.
Structural reforms aimed at gerrymandering and primaries may still be worth pursuing for other reasons — fairer representation, broader participation — but the polarization research suggests they shouldn’t be expected to meaningfully cool down political messaging on their own, given how weak the evidence linking them to polarization has turned out to be.
Frequently asked questions
Is political messaging actually becoming more extreme, or does it just feel that way?
Both are measurable. Ideological gaps between parties have widened over decades, but the sharper and more consistent trend is affective polarization — how much partisans dislike and distrust the other side — which has risen significantly since 2016 according to Pew and YouGov data.
Do people accurately judge how extreme the other political side actually is?
No. Research finds Americans dramatically overestimate the other party’s extremism — in one large survey, roughly half of respondents believed opposing voters held extreme views and discussed politics constantly, when the real figure was closer to one in seven.
Is social media the main cause of political extremism?
Most researchers say no, but it is a real amplifier. A major review of over 50 studies concluded platforms are unlikely to be the root cause but do meaningfully worsen polarization, partly because engagement-based algorithms favor emotionally charged content.
Does gerrymandering cause more extreme political messaging?
The evidence is weaker than commonly assumed. Polarization rose just as sharply in the U.S. Senate, where gerrymandering is impossible, as in the House, and single-district states show similar polarization to heavily redistricted ones.
Why do politicians keep using aggressive rhetoric if voters say they dislike it?
Because aggressive, personal attacks generate disproportionate media coverage and social engagement compared to policy-focused messaging, giving individual politicians a strong incentive to use it even if it’s unpopular with the electorate as a whole.
Is this happening in other countries too, or just the United States?
Rising polarization has been documented in multiple democracies, though research consistently finds the U.S. has polarized faster and further than most comparable countries over the same period.
Can political polarization be reduced?
Some interventions show measurable effects — reduced social media use lowered issue-based polarization in at least one controlled study — but researchers generally describe polarization as the product of multiple reinforcing systems, which makes single-fix solutions unlikely to work well on their own.