What proportion of 2025 ICE arrests involved individuals with criminal convictions versus civil immigration violations?
Executive summary
Analysis of multiple datasets and reporting shows that during 2025 the share of people arrested by ICE who had prior U.S. criminal convictions fell sharply while the share arrested solely for civil immigration violations rose — by mid-to-late 2025 roughly a third to nearly half of ICE arrests were of people without criminal convictions, though exact percentages vary by dataset and month [1] [2] [3].
1. The core finding: a rapid shift from criminal to immigration-only arrests
Across public and independent analyses, the headline is consistent: as ICE’s total arrests surged in 2025, the portion of those arrested who had U.S. criminal convictions declined, while arrests for immigration violations alone increased; Stateline’s compilation finds immigration violations accounted for about 20% of arrests in April and grew to roughly 44% by October 2025 [2], and the Deportation Data Project documents a steep rise in street arrests and transfers that included many people without criminal convictions [3].
2. How big a share had no criminal convictions?
Estimates differ by source and month, but several independent analyses converge on the conclusion that by late 2025 into early 2026 a substantial minority or plurality of ICE arrestees lacked criminal convictions: the FactCheck analysis reports that the share of detainees with no criminal convictions or pending charges rose from about 14.7% in February 2025 to 42.7% in January 2026 [1], and Stateline’s October estimate of immigration-only arrests near 44% aligns with that trend [2]; other trackers put the proportion of people in ICE custody without convictions even higher for some snapshots (TRAC reports 74.2% of those held in ICE detention had no criminal conviction as of Jan. 25, 2026, though that figure describes detention population not only arrests) [4].
3. What counts as “criminal conviction” versus “civil immigration violation”?
Part of the variation in reported proportions stems from differing category definitions: ICE’s own public stat categories separate people with U.S. criminal convictions, those with pending criminal charges, and those with neither convictions nor pending charges but with immigration offenses such as visa overstays or reentry [5]. Many reports combine convictions and pending charges when measuring “criminal history,” while others isolate convictions only; PolitiFact’s read of Deportation Data Project material shows roughly 64–66% of detained/arrested people had either a conviction or pending charge in the Jan–Oct 2025 window [6].
4. Where analysts disagree and why the debate is heated
Analysts and commentators dispute how to interpret short-term snapshots versus trends and whether ICE’s stated focus on “the worst of the worst” matches practice: some researchers highlight that while the absolute number of arrestees with prior convictions remained higher than under 2024, the relative share fell as enforcement expanded, meaning growth was driven mainly by nonconvicted immigration violators [7] [3]; administration and some pro-enforcement voices emphasize that many arrested do have criminal histories when pending charges are included, citing ICE and aggregated datasets [6] [8]. This methodological divergence — convictions-only versus convictions-or-pending-charges — explains much of the disagreement.
5. Limitations and what the numbers cannot show
Public reporting and FOIA-derived datasets are powerful but imperfect: ICE’s public statistics use categorical groupings that mix reentry/immigration-fugitive issues with nonconvicted status [5], several independent projects cover different time windows (Jan–Oct 2025, through Jan 2026, or single-day spikes) and sometimes different denominators (daily book-ins, monthly arrests, or detention population) [1] [3] [4], and leaked or nonpublic ICE slices cited by think tanks may not be fully described in public analyses [9]. Where sources do not resolve exact year-end totals for all of 2025, this account does not invent a single definitive percentage.
6. Bottom line answer
Using the best-available public analyses for 2025, the proportion of ICE arrests involving people with criminal convictions declined across the year and by late 2025 roughly one-third to nearly one-half of arrests were of people without criminal convictions (fact-based estimates put immigration-only arrests near ~40–44% in autumn/winter 2025 and the share with no conviction or pending charges at ~42.7% in January 2026 per FactCheck and Stateline) while the remainder had convictions or pending charges depending on the dataset’s definitions [2] [1] [6]. Exact proportions depend on whether pending charges are counted as “criminal,” which dataset and month are used, and whether the metric is arrests versus detained population [5] [4].