What percentage of ice arrests in 2025-2026 have been non-criminals

1. The headline numbers: a majority without convictions in many snapshots

Several high‑profile analyses report that a majority of ICE detainees during 2025 and into January 2026 had no criminal conviction on their records: the Migration Policy Institute reported that as of September 2025, 71 percent of ICE detainees had no criminal conviction ^[1], and FactCheck’s summary of Deportation Data Project material noted that in January 2026 roughly 29 percent of detainees had convictions, implying about 71 percent did not ^[2].

2. Conflicting snapshots: lower estimates from other datasets and definitions

Not all reporting produces the same share; Cato’s analysis of leaked ICE data and other public spreadsheets found that the share of detainees without convictions rose to around 40 percent in November 2025 and that 73 percent of those deported had no conviction in another snapshot, while only 5 percent had violent convictions ^[3]. Independent trackers and writers using ICE’s published spreadsheets and single‑day counts produced midrange estimates — for example, one analysis found nearly half of people held lacked a criminal history as of early January 2026 ^{[4] [5]}.

3. Why the numbers disagree: definitions, timing, and data sources

The variance reflects three data problems: first, “non‑criminal” is used interchangeably to mean “no criminal conviction,” “no pending criminal charge,” or “not convicted of violent/property offenses,” which produces different denominators and rates ^{[3] [2]}. Second, timing matters: rapid policy shifts and large increases in detainee populations from spring through winter 2025 changed proportions month to month ^{[6] [7]}. Third, datasets differ — public ICE spreadsheets, Freedom of Information Act releases compiled by the Deportation Data Project, leaked internal Cato data, and single‑day detention tallies each capture different slices of enforcement activity ^{[7] [5] [3]}.

4. What the best synthesis supports: a broad range, but consistent direction

Weighing the sources together, the most defensible conclusion is that a large and growing share of ICE arrests and detentions in 2025–early 2026 involved people without criminal convictions, with credible estimates clustering between roughly 40 percent on the conservative side and about 71 percent in several widely cited analyses for late 2025 and January 2026 ^{[3] [1] [2]}. Multiple independent researchers and advocacy reports concur that the increase in detention was driven overwhelmingly by people without prior convictions, even if exact percentages diverge by dataset ^{[5] [7] [4]}.

5. Caveats and competing interpretations

Advocates and watchdogs emphasize civil‑immigration detentions of non‑convicted people as evidence of a policy shift toward mass interior arrests ^{[6] [7]}, while agencies and some officials have framed operations as targeting dangerous criminals; however, analyses of the underlying data frequently show only a small fraction of detainees had violent felony convictions ^{[3] [2]}. Reporters and analysts caution that gaps in identifiers, shifting reporting categories, and exclusion or lumping of CBP vs ICE arrests make precise, single‑number claims unreliable without careful methodological footnoting ^{[8] [5]}.

6. Bottom line — what percentage?

Given the differing sources and definitions, the most accurate, source‑backed answer is a range: public and independent data indicate that between roughly 40 percent and about 71 percent of people in ICE custody in late 2025–January 2026 had no criminal conviction, with several leading datasets and analyses clustering near the higher end (around 70–71 percent) for September 2025 through January 2026 and others showing lower but still substantial non‑convicted shares depending on how “non‑criminal” is defined and which days are measured ^{[3] [1] [2] [4]}.