What do ICE case‑level data show about the composition (criminal vs. noncriminal) of arrests under Trump vs. Obama?

1. What the raw counts and public tables show: more non‑convicted people held under Trump-era tactics

Multiple case-level analyses and ICE releases indicate the proportion of people in ICE detention without criminal convictions rose sharply after enforcement shifted from jails to community "at-large" arrests under the Trump administration; leaked ICE data summarized by the Cato Institute showed roughly two‑thirds of arrests were of people without convictions by late July of a reporting year ^[4], and Syracuse/TRAC snapshots documented that in certain 2025 counts about 72% of people in ICE detention lacked criminal convictions ^[5]. Independent outlets and advocacy groups corroborate a pattern of big increases in arrests of people without criminal records tied to new tactics such as roving patrols, worksite actions and re‑arrests at routine check‑ins ^{[3] [6]}.

2. Why headline percentages differ: definitions, sampling, and what ICE publishes

Comparisons are complicated because sources use different denominators and definitions: ICE and some administration statements count people with pending charges as "criminal arrests," whereas case‑level datasets and research groups often separate convictions, pending charges and immigration‑only violations ^{[4] [7]}. ICE’s public "Initial Book‑Ins" and detention management tables only record arrests that result in entry to ICE detention, likely undercounting some enforcement activity and skewing interpretation if one assumes they represent all arrests ^[8]. TRAC has repeatedly cautioned that interior administrative arrests rose relative to the late Obama years but were still below peak Secure Communities-era levels earlier in the prior decade, adding context to raw year‑to‑year comparisons ^{[2] [9]}.

3. Context from Obama-era enforcement: more jail‑based, conviction‑centered removals but large totals

Obama administrations built extensive capacity for identifying and removing unauthorized migrants, and programmatic priorities led to many arrests of people with criminal convictions held in local jails; TRAC counted roughly 65,332 detained and deported by ICE in FY2016 as a reference point for comparison ^[1]. That jail‑centric model meant a higher fraction of ICE arrests were of people already convicted or charged in criminal court, even as the agency also conducted immigration‑only operations; policy memos and enforcement priorities under successive administrations shaped those patterns ^{[1] [10]}.

4. Alternative interpretation and limits of the evidence

Some reporting and internal ICE tallies emphasize that the agency has identified hundreds of thousands of noncitizens with criminal convictions and that high‑priority criminal removals have continued, noting arrests of people convicted of murder or sexual assault among the broader enforcement effort ^[11]. At the same time, fact‑checks and independent researchers stress that headline percentages of "criminal" arrests can be misleading without distinguishing convictions from charges and immigration violations; available datasets do not always provide consistent, case‑level coding across years, constraining definitive, apples‑to‑apples conclusions ^{[11] [5] [8]}.

5. Bottom line for readers following the numbers

Case‑level data and investigative reporting converge on a firm, qualified finding: enforcement tactics under Trump produced a substantial rise in interior, at‑large arrests and a much larger share — and number — of people in ICE custody without criminal convictions compared with the later Obama years, but exact percentages hinge on definitional choices and which ICE tables are counted; researchers caution that ICE’s published tables and internal counts must be read carefully to avoid overstating either continuity or change ^{[3] [4] [2] [5]}.