Which public reports or datasets provide breakdowns of ICE arrests and conviction status for 2025?

1. ICE’s official ERO statistics — the baseline government dataset

ICE publishes Enforcement and Removal Operations statistics that include arrests by country of citizenship and categories of criminal history, and the agency frames these as the canonical public record for arrests, removals, detention, and alternatives to detention; those pages are intended to be the starting point for anyone seeking official breakdowns of arrests and conviction status for 2025 ^[1].

2. Deportation Data Project — FOIA-assembled, researcher-ready microdata

The Deportation Data Project has posted detailed ICE data releases covering arrests, detainer requests, and detentions through October 15, 2025, with linked identifiers in some releases that let analysts trace individuals through enforcement pipelines; the project also supplies a codebook and guidance for journalists and advocates using ICE-provided FOIA data ^[2].

3. TRAC (Transactional Records Access Clearinghouse) — independent, law‑school‑style counts and trends

TRAC produces periodic national and field-office breakdowns of ICE arrests and detention populations (including conviction-status tallies) and was cited repeatedly in 2025 reporting for showing shifts in the share of people arrested without criminal convictions; TRAC’s analyses feed much local and national coverage about how many detainees lack convictions ^{[3] [5] [6]}.

4. Media data trackers and newsrooms that synthesize multiple sources

Large newsrooms and data teams — for example NBC News’ immigration tracker — combine ICE, CBP, and FOIA-derived datasets to report counts of arrests by offense type and to track convictions such as murder and sexual assault among those arrested; these aggregations are valuable for narrative context and month-to-month changes but depend on underlying public releases and FOIA dumps ^[7].

5. Independent analyses and leaked datasets — deeper but sometimes nonpublic

Research groups and think tanks have published analyses based on leaked ICE extracts that are not fully public; for example, the Cato Institute published findings from nonpublic ICE data showing a large share of detainees and deportees without criminal convictions in late 2025, a claim that relies on data not released directly via ICE’s public portal ^[4]. Such sources can reveal granular conviction-status shifts but require caution about provenance and completeness.

6. Advocacy and watchdog summaries that reprocess government numbers

Organizations like Prison Policy and other advocacy projects have used ICE-supplied data and Deportation Data Project tables to produce state-by-state arrest-rate and conviction-status breakdowns for 2025, emphasizing patterns such as concentration of arrests in collaborating states and the proportion of arrests originating in local jails ^[8]. These analyses are transparent about using ICE data but also carry advocacy frames that readers should recognize.

7. Limitations, contested claims, and political framing to watch for

Several reputable sources show divergent emphases: DHS press releases highlight arrests of “the worst of the worst,” citing specific violent convictions to justify enforcement ^{[9] [10]}, while multiple data analytic groups and journalists report that a large and growing share of people arrested or detained in 2025 had no criminal convictions or only minor offenses ^{[4] [5] [11]}. The chief methodological limits are timing (many public releases cover through mid‑October 2025), differing definitions of “criminal conviction” vs. “pending charge,” and the existence of leaked versus officially posted records that produce different headline percentages ^{[2] [1] [4]}.