What public datasets break down ICE detainees by immigration legal status and criminal history?

1. ICE’s public statistics and bi‑weekly detention reports — the primary official source

ICE publishes Enforcement and Removal Operations statistics and bi‑weekly detention counts that explicitly categorize detained individuals as “with criminal convictions,” “with pending criminal charges,” or “no convictions or pending charges” and also provides population breakdowns by country of citizenship and other administrative fields; these official pages are the starting point for direct counts and trend analysis ^{[1] [7]}.

2. SSOR / ICE Detentions extracts (EID/IIDS → OHSS) — administratively detailed but technical

The Office of Homeland Security Statistics constructs SSOR files from ICE’s Enforcement Integrated Database extracts (via the IIDS Data Mart) and includes a Yes/No variable for prior criminal convictions as well as book‑in/book‑out events and removal outcomes; researchers use these files to generate detention‑stint level views and linkages to arrests and removals ^[2].

3. Deportation Data Project — FOIA originals and linked, person‑level tables

The Deportation Data Project has posted near‑original individual‑level ICE datasets obtained via FOIA — arrests, detainers, detentions, encounters and removals — including linked identifiers that allow tracing an individual’s pathway through enforcement and thus disaggregation by legal status and criminal history when merged properly ^{[3] [4]}.

4. Independent analysts and aggregators — TRAC, Vera, media and think tanks

Non‑governmental groups process and interpret ICE outputs: TRAC and Vera host processed detention tables and facility maps and have highlighted how many detainees lack criminal convictions; media outlets and think tanks (e.g., The Guardian, KPBS, Cato) have used the ICE and FOIA datasets to report shares such as large percentages of detainees without criminal convictions, but these analyses reflect choices about which ICE tables and date snapshots to use ^{[8] [6] [9] [10] [11] [5]}.

5. Strengths, limitations and what the datasets do not fully reveal

While ICE’s published categories let users quantify convictions vs pending charges vs no convictions, the datasets differ in coverage (facility lists, border detention vs interior), contain processing errors and oddities noted by Vera (e.g., inconsistent facility reporting, zero‑minute stays, corrected fields), and often lack standardized offense coding that makes fine‑grained criminal‑history severity analysis difficult without additional matching or cleaning ^[6].

6. Institutional framing and competing narratives to account for

ICE’s public pages and FAQs stress that immigration violations alone can merit detention and emphasize threat‑level classifications for “criminal aliens,” which frames interpretation of the same data differently than advocacy groups highlighting the rising share of detainees with no convictions; third‑party analysts therefore must disclose methodological choices because agency language and selective snapshots can support divergent policy arguments ^{[12] [7] [9] [13]}.

7. Practical guidance for researchers seeking these breakdowns

To produce a defensible breakdown, start with ICE’s detention snapshot and the SSOR variables for conviction status (official counts), then supplement with the Deportation Data Project’s person‑level FOIA releases for linkage and cross‑validation, and finally consult TRAC/Vera processed tables for cleaned trend context while explicitly documenting dataset versions and known anomalies described by Vera ^{[1] [2] [3] [4] [6]}.