By Michael Phillips | Riptide
Charlie Kirk was shot dead at Utah Valley University on September 10, 2025. A year later, Tyler Robinson has pleaded not guilty and is headed to trial on an aggravated murder charge, with prosecutors seeking the death penalty. Turning Point USA is holding a memorial on the same campus where he was killed. And the anniversary has revived a claim that’s been repeated so often over the past twelve months it now sounds like settled fact: that America is living through an unprecedented surge in political violence.
The perception is nearly universal. In a Pew survey conducted after Kirk’s assassination, 85 percent of Americans said they believed political violence was increasing. The harder question is whether that perception, the contemporary incident counts, and the historical record are all measuring the same thing. The claim isn’t baseless. But “unprecedented” is doing a lot of work it can’t support, and at least one death regularly folded into the narrative isn’t political violence at all. Here’s what the receipts actually show.
The count
“Those measures can all illuminate the same risk environment, but their totals aren’t interchangeable.”
The clearest data point comes from Michael Jensen, director of research at American University’s Polarization & Extremism Research & Innovation Lab (PERIL). By his tracking, the U.S. saw more than 520 incidents of terrorism and targeted violence in the first six months of 2025 alone — roughly a 40 percent increase over the same period in 2024. Those incidents produced 96 deaths and 329 injuries, touching every state but Montana and Wyoming.
Princeton’s Bridging Divides Initiative (BDI), using data from the START consortium, recorded a 34.5 percent increase in terrorism and targeted-violence events during the first eight months of 2025 compared with the same period in 2024 — while cautioning that estimates vary substantially depending on methodology. Threats and harassment spiked in the aftermath of high-profile attacks — the U.S. Capitol Police logged 14,938 threat-assessment cases involving members of Congress, their families, staff, and the Capitol complex in 2025, up from 9,474 the year before and 8,008 in 2023. That’s the third straight annual rise, and BDI notes the single largest spike in threats against local officials came in September 2025 — the month Kirk was killed, with roughly half of that spike tied directly to reactions to his death.
A note on definitions: There is no single national “political violence” counter. START’s Terrorism and Targeted Violence database is deliberately broad — it includes terrorism alongside premeditated hate crimes, school and workplace targeted violence, public mass attacks, and some grievance-driven violence. BDI separately tracks threats and harassment against public officials. Capitol Police threat-assessment cases include concerning communications and behavior that may never become violent acts. CSIS, discussed below, uses a narrower terrorism-specific dataset. Those measures can all illuminate the same risk environment, but their totals aren’t interchangeable — which matters for how the roll call below should be read.
Across several major datasets, the indicators moved in the same worrying direction: terrorism and targeted violence increased, threats against officials rose sharply, and harassment intensified. But those datasets are measuring overlapping — not identical — phenomena, and BDI’s own Roudabeh Kishi has said it’s difficult to say definitively whether political violence itself increased in 2025, precisely because researchers define the category so differently.
The roll call — and what doesn’t belong on it
Line up the incidents that get invoked when people talk about “a year of political violence,” and the list is genuinely brutal: former Minnesota House Speaker Melissa Hortman and her husband, killed in their home in June 2025 — a political assassination targeting Democratic elected officials and their families, according to federal prosecutors. Kirk that September. And in 2026, the fatal shootings of Renee Nicole Good and Alex Pretti during federal immigration-enforcement operations in Minnesota, which the Carnegie Endowment and BDI both treat as part of the current risk environment but categorize separately — as violence from government actors, a distinct and reportedly growing category tied to ICE operations rather than the lone-attacker pattern of the others.
UnitedHealthcare CEO Brian Thompson’s killing outside a Manhattan hotel in December 2024 illustrates the definitional problem directly rather than settling into the roll call cleanly. His murder qualifies under the broad “targeted violence” frameworks that produce headline numbers like Jensen’s 520 incidents. But whether an ideologically tinged attack on a corporate executive belongs in the narrower category of political violence — which Carnegie defines as violence intended to affect democratic institutions, political actors, or people’s political participation — depends entirely on which research definition is doing the counting.
“If a senator’s fatal aortic dissection gets absorbed into the same ‘political violence’ narrative as an assassination in front of 3,000 people, the term stops meaning anything.”
One name that does not belong on either list, despite how often it gets swept into the same sentence: Sen. Lindsey Graham. Graham died on July 11, 2026; preliminary findings from the D.C. medical examiner attributed his death to an aortic dissection caused by arteriosclerotic cardiovascular disease. There was no indication of foul play, no suspect, no investigation. It was a natural death that happened to trigger a messy Senate succession fight in South Carolina, not an act of violence.
That distinction matters for reasons beyond accuracy. If a senator’s fatal aortic dissection gets absorbed into the same “political violence” narrative as an assassination in front of 3,000 people, the term stops meaning anything — and it hands ammunition to anyone who wants to dismiss the real data as hysteria. Scalpel, not sledgehammer.
The reaction problem is real — and it isn’t one-sided
“The honest version of this story isn’t ‘the left celebrates death and the right doesn’t.’”
None of that means the rhetoric surrounding Graham’s death is beside the point. It’s a different story than “political violence,” but it’s a real one: how people publicly react to a political opponent’s death has gotten uglier, and the receipts on that are just as documentable as the incident counts above.
Within hours of Graham’s death, Omar Suleiman, a Texas imam and president of the Yaqeen Institute for Islamic Research, posted “Bye Lindsey… May you live an eternity in ruins for the ruins you helped create in Gaza.” St. Bonaventure University professor Ibrahim Zabad posted a message ending “Rest in Hell” before deleting it and apologizing roughly twelve hours later. Indiana University adjunct Elizabeth Booker Houston delivered a cake bearing Graham’s photo and the word “congratudolences,” prompting several Indiana Republican lawmakers to demand the university remove her. The Young Turks host Ana Kasparian posted “Good riddance.” Those are real, named, on-the-record instances — not a vibe.
What complicates the “the left did this” framing is that the backlash wasn’t confined to the right, and the celebration wasn’t confined to the left. Ana Navarro — a Republican political commentator and longtime Graham friend — called celebrating his death “inhumane” and said it was “really lacking empathy” on The View. And the swapped-jerseys test cuts hard the other way, too: this is the same twelve months in which Donald Trump wrote “Good, I’m glad he’s dead. He can no longer hurt innocent people!” after Robert Mueller’s death in March, and invoked “Trump Derangement Syndrome” in attacking Rob Reiner after Reiner’s death in December 2025. If citing individuals gloating over Graham’s death counts as evidence of a rhetoric problem, intellectual consistency requires counting the president’s own eulogies-as-attacks the same way.
The honest version of this story isn’t “the left celebrates death and the right doesn’t” — it’s that the norm against publicly celebrating a political opponent’s death has weakened across the spectrum, from anonymous accounts up through a sitting president. That’s worth documenting. It’s a different claim than “Graham’s death was itself an act of political violence,” and the two shouldn’t be merged.
Where the trackers disagree
Even among researchers who see an elevated threat environment, there’s a live methodological fight over how much of it can be attributed to different ideological categories. A CSIS analysis released in September 2025 argued that left-wing terrorist attacks and plots were on pace to outnumber far-right incidents for the first time in more than three decades — a claim built on just five left-wing attacks or plots through July 4, 2025 (CSIS separately classifies Luigi Mangione’s killing of Brian Thompson as one of those left-wing incidents, which only underscores how much a single classification call can move the count).
Jensen — the same PERIL researcher behind the broader incident count above, working with fellow researcher Amy Cooter — argues that drawing a historical trend from five incidents is statistically fragile, particularly when the classification of only a handful of additional cases could reverse the result. They point out that CSIS’s own dataset, running from 2016 through mid-2025, attributes 13 deaths to left-wing attacks, 82 to jihadist attacks, and 112 to right-wing attacks — a longer baseline that cuts against reading a seven-month swing as a historic reversal. FactCheck.org’s review of the same period CSIS highlighted found the underlying numbers were five left-wing incidents with zero deaths against one right-wing incident with two deaths — small enough in absolute terms that headline percentage comparisons obscure more than they reveal.
“The argument isn’t over whether today’s violence is serious. It’s over whether ‘unprecedented’ describes the historical record at all.”
Historical context cuts against “unprecedented” much harder than a vague gesture at the 1960s and ’70s can convey. Even BDI, which concludes that political violence increased in 2025, notes that available measures remained below the peaks reached in 2020 and early 2021 — the historical comparison gets more lopsided the farther back you go. Dartmouth’s Polarization Research Lab, using its own narrower definition, counts more than 120 incidents of political violence in 1970 alone, against 10 in 2025. An FBI retrospective counted roughly 641 terrorist incidents between 1971 and June 1975, including 166 bombings, 120 fire bombings, and 118 shootings; an FBI bulletin from December 1970 reported that bombings during one six-month stretch were averaging more than three per day. By that measure, the argument isn’t over whether today’s violence is serious. It’s over whether “unprecedented” describes the historical record at all. None of these datasets can tell us how much of the public’s heightened perception comes from greater violence and how much comes from greater exposure to it. A smartphone video, social-media clip, or livestream can put millions of people effectively at the scene of an attack within minutes — a visibility that historical incident totals don’t measure. What’s arguably different now isn’t the body count so much as the visibility: a shooting in front of a phone-camera crowd of thousands travels differently than a wire story from 1975 ever did.
What the receipts actually support
“A smartphone video, social-media clip, or livestream can put millions of people effectively at the scene of an attack within minutes.”
Strip out the noise and here’s what survives: multiple independent trackers using different methodologies agree that targeted violence and threats against officials rose substantially from 2024 into 2025, and researchers continued to describe the risk environment as elevated into 2026. That’s real, and it’s the honest headline coming out of Kirk’s anniversary.
What doesn’t survive scrutiny is the flattened version of that headline — the one where every high-profile death of a political figure in the last year gets stacked into a single body count regardless of cause, and where a single contested report gets cited as proof of a “most violent year in decades” without noting researchers in the same field call the underlying numbers too thin to support that claim. The evidence says several measures are moving in the wrong direction. It doesn’t say every name people are citing belongs on the chart.
“The evidence says several measures are moving in the wrong direction. It doesn’t say every name people are citing belongs on the chart.”
Sources: Reuters and Utah County Attorney’s Office charging documents on the Sept. 1, 2026 ruling and not-guilty plea in State v. Robinson; TPUSA/UVU memorial event listing; Association of Health Care Journalists reporting on PERIL/START data (Michael Jensen); Bridging Divides Initiative, “Key Political Violence and Resilience Trends From 2025” and its September 2025 local-officials threat-spike data snapshot; U.S. Capitol Police 2025 threat-assessment case data; U.S. Department of Justice statements on the Hortman killings; Carnegie Endowment for International Peace, “Political Violence in the U.S.”; START’s Terrorism and Targeted Violence (T2V) database documentation; CSIS, “Left-Wing Terrorism and Political Violence in the United States”; Just Security’s response to the CSIS report; FactCheck.org, “What We Know About Political Violence in America”; Scientific American’s reporting on Dartmouth’s Polarization Research Lab and the post-Kirk Pew survey; FBI historical retrospectives on 1970s-era terrorism; Reuters on the D.C. medical examiner’s preliminary findings on Sen. Lindsey Graham’s death; the Daily Signal, Fox News, Indiana Public Media, and Mediaite on public reactions to Graham’s death; Us Weekly on Ana Navarro’s comments; Reuters on Robert Mueller’s death; the Washington Post on Donald Trump’s response to Rob Reiner’s death.
Keep This Reporting Free
If this work matters to you, please consider supporting it.
Your contribution helps fund independent reporting across our entire network.
Discover more from RIPTIDE
Subscribe to get the latest posts sent to your email.
