Editorial illustration.
Editorial illustration.
In Depth · United States · National

Three Datasets, One Spring: How the 2024 Campus Protests Were Counted

ACLED, the Crowd Counting Consortium and Princeton’s Bridging Divides Initiative each tracked the encampment wave. Their headline figures are often quoted side by side, but they count different things. This is a guide to what each number means, followed by what national polling says about the public’s view of the protests.

Summary

  • ACLED counted “more than 1,360 student demonstrations” from October 7, 2023 to May 3, 2024, over 94% of them pro-Palestinian, and wrote that 97% “remained peaceful.” Its unit is an event aggregated by place, day and actor.
  • The Crowd Counting Consortium logged 3,704 protest-days at 525 schools, K–12 included, in 317 cities. Encampments made up 1,639 of those days (44%), and the dataset recorded 3,645 arrests. Its unit is the protest-day: a camp standing for 30 days counts 30 times.
  • Princeton’s Bridging Divides Initiative re-coded both datasets: about 1,150 encampment events at nearly 150 colleges from April 17 to May 12, 2024. In 95% of them there were “no reports of encampment protesters engaging in physical violence or destructive activity.” Police made arrests at about 9%.
  • ACLED found that when students demonstrated unopposed, police intervened against pro-Palestinian events “more than five-and-a-half times as often” as pro-Israel ones.
  • Gallup’s February 2025 poll found 46% of Americans sympathized more with Israelis, the lowest in 25 years of tracking, and 33% with Palestinians, the highest. Pew found large majorities of Americans saying calls for violence against Jews or Muslims should not be allowed.

Different units, different totals

The most common error in coverage of the 2024 protests is to compare figures from different datasets as if they measured the same thing. BDI’s issue brief, which drew on both ACLED and the CCC, quotes each project’s own definition.

ACLED aggregates. “Events that share the same date, actors, a proximate location, and event type will be aggregated when it cannot be clearly discerned that the events were discrete and independent,” its methodology says. Five demonstrations a few blocks apart in Atlanta on the same topic might become one ACLED event.

The CCC disaggregates by day. “When events span multiple days, as many of the recent student encampments have, we create a separate entry for each day,” its researchers wrote. “Technically, then, our overall count is of protest-days rather than separate events, but in the vast majority of cases the two are the same.” That rule matters a great deal for encampments: Stanford’s first camp, the CCC notes, “stood in White Plaza for more than 120 days.”

The CCC count also includes “colleges, universities, K-12 schools, and school district offices.” A “school protest” in its data is not necessarily a university protest.

What each found

ACLED. Student demonstrations made up “more than 40% of all US demonstration activity related to the conflict” since October 2023. California was “the main student protest battleground.” Police intervention against student demonstrations “increased by more than eight times in April compared to March.” On the asymmetry question, ACLED wrote: “in cases where student demonstrators have gathered unopposed, police have intervened against pro-Palestine demonstrations more than five-and-a-half times as often as pro-Israel demonstrations.” The ratio is descriptive. It does not control for the type of event, and pro-Palestinian student events, more than 94% of the total, included the encampments that campus rules treated differently from one-off rallies.

CCC. Its May 2024 overview found the encampment phase did not start at Columbia: camps “began popping up in mid-October 2023, not mid-April.” It recorded 3,645 arrests across roughly 3,700 protest-days, “nearly one per protest day,” with UCLA highest at more than 250. On violence, BDI quotes the CCC as saying that in “more than 10,000 pro-Palestine actions” since October 7, “we have only seen property damage at 128 of them and police injuries at 13,” and on school campuses, “property damage is 30 and police injuries 6.” Most property damage “involved graffiti or similar defacement.”

BDI. Re-coding the spring wave for its own purposes, BDI found about 1,150 encampment demonstration events at nearly 150 campuses in at least 35 states, Puerto Rico and Washington, DC, between April 17 and May 12. About 60 events included reports of violent or destructive activity: “physical confrontations with law enforcement, physical confrontations with counter-demonstrators, the throwing of projectiles (e.g. bottles, cans) at law enforcement, or damage to occupied buildings.” At least 13%, “over 150 events,” drew counter-protesters, and nearly 40 of those saw violence or confrontation between groups. BDI also counted “at least 5 cases” of car rammings or attempted rammings. It reported that “at least 20 schools in more than 10 states have come to agreements with” protesters.

What “peaceful” means here

All three projects code what was reported, and the codes concern physical violence and property damage. They do not measure harassment, intimidation or the content of chants, which is the subject of most complaints by Jewish students and of the federal investigations that followed. A protest coded peaceful can still be the subject of a Title VI complaint. Both things can be true, and an honest summary will say so.

The public’s view

Gallup’s annual World Affairs survey in February 2025 found that “the 46% expressing support for Israel is the lowest in 25 years of Gallup’s annual tracking,” while “the 33% of U.S. adults who now say they sympathize with the Palestinians is up six percentage points from last year and the highest reading by two points.” The partisan gap was wide: Republicans 75% to 10% toward Israelis, Democrats 59% to 21% toward Palestinians.

Pew’s April 2024 survey asked a different question: which kinds of speech should people be allowed to express “even if doing so might offend others.” Seven in ten said speech supporting “Israel’s right to exist as a Jewish state” should be allowed. Smaller majorities said the same of speech opposing a Palestinian state (61%) or opposing Israel’s existence as a Jewish state (58%). Only “one-in-ten adults surveyed say that speech advocating violence against either Jews (10%) or Muslims (11%) is permissible.”

A note for this site

Each case item that cites a protest count should name the dataset and its unit. “3,700 protests” from the CCC and “1,360 demonstrations” from ACLED are not in conflict; they are different measurements. The data template proposed in the editorial plan includes a field for exactly this.

DatasetPeriodUnitHeadline
ACLEDOct 7, 2023–May 3, 2024Event (aggregated)>1,360 student demonstrations; 97% peaceful
CCCOct 7, 2023–late May 2024Protest-day (K–12 included)3,704 days at 525 schools; 3,645 arrests
BDI (ACLED+CCC re-coded)Apr 17–May 12, 2024Encampment event-day~1,150 events; 95% no protester violence reported

Sources

  1. ACLED, “US Student Pro-Palestine Demonstrations Remain Overwhelmingly Peaceful,” May 2024
  2. ACLED, “United States: Scope and coverage in ACLED data.”
  3. Crowd Counting Consortium, “An Empirical Overview of Recent Pro-Palestine Protests at U.S. Schools,” Ash Center, Harvard Kennedy School, May 2024
  4. Princeton Bridging Divides Initiative, “Campus Encampment Protests: Issue Brief,” May 2024
  5. Gallup, “Less Than Half in U.S. Now Sympathetic Toward Israelis,” March 2025
  6. Pew Research Center, “Acceptable speech and the Israel-Hamas war,” April 2, 2024

Further reading

Topics: encampment, violence, data.