How do FOIA‑obtained ICE datasets define and categorize criminal history (conviction vs. pending charge vs. none)?
Executive summary
FOIA‑obtained ICE datasets classify individuals into three broad criminality categories—convicted criminal, pending criminal charges, and other/no known criminal history—based on entries in ICE’s internal systems of record at the time of the enforcement action (Deportation Data Project documentation and related analyses) [1] [2]. Those classifications reflect administrative labels in ICE databases and come with important caveats: ICE counts people with both convictions and pending charges as “convicted,” pending‑charge labels do not imply guilt, and the raw FOIA tables can miss arrests handled by other agencies or contain incomplete fields [3] [4] [2].
1. What the labels mean inside the datasets: ICE’s three‑category taxonomy
The FOIA releases reproduced by the Deportation Data Project present a three‑category “criminality” field: “Convicted Criminal” for immigration violators who have a criminal conviction recorded in ICE’s system at the time of the enforcement action, “Pending Criminal Charges” for those with active but unresolved charges, and an “Other/No Known Criminal Conviction or Pending Charge” bucket for people without convictions or pending charges on record (the Data Project’s documentation and analyses of ICE definitions make this explicit) [1] [2]. ICE’s own public pages use similar wording—distinguishing individuals with criminal convictions, those with pending charges, and those without known criminal charges—showing that the FOIA categories mirror agency practice rather than an outside taxonomy [4].
2. How overlapping records are handled and why counts can be misleading
ICE’s practice—explicit in agency reporting and independent analysis—is to count someone who has both a prior conviction and a pending charge only in the “conviction” category, which compresses overlaps into a single label and therefore understates the number listed under “pending” if one does not read the documentation (this counting rule is noted in ICE summaries and Pew’s analysis of ICE arrest data) [3] [4]. Journalists and advocates who use the FOIA tables must therefore beware: totals for convictions, pending charges, and “no convictions” are not three mutually exclusive tallies unless a dataset explicitly de‑duplicates by person and documents the coding rules [2] [5].
3. What these labels do—and do not—say about guilt or seriousness
The “pending charges” label in the FOIA datasets denotes only that a charge was recorded in a U.S. criminal system at the time ICE captured the record; it does not mean the person was convicted, nor that charges were serious or sustained (advocacy reports and analysts stress that pending charges are often minor or later dismissed) [6] [7]. Conversely, the “convicted” label records convictions as entered in ICE systems but does not reliably indicate the severity distribution without inspecting offense codes—ICE and independent researchers note many convictions are for traffic, low‑level drug, or immigration offenses rather than violent crimes [4] [8].
4. Data provenance, gaps and the risk of overreach in media interpretations
These FOIA datasets are copies of ICE system records turned over under FOIA and republished by projects like the Deportation Data Project; they do not necessarily include Border Patrol or HSI arrests and can suffer missing or delayed fields such as detainer outcomes and transfer records, meaning apparent “no charge” or “no conviction” values sometimes reflect a lack of update rather than a clean criminal history (Deportation Data Project documentation and journalists’ guides emphasize these limits) [9] [2] [5]. Several recent media narratives that cite high shares of detainees “with no criminal record” rely on these FOIA labels, but critics and data analysts caution that headline figures can be amplified by classification rules, missing data, and differences between administrative arrest reasons and later court outcomes [10] [6].
5. Practical implications for researchers and reporters using FOIA ICE data
Anyone analyzing FOIA‑obtained ICE data must read the accompanying documentation, account for ICE’s rule of assigning people with both convictions and pending charges to the conviction category, validate FOIA tallies against other agency dashboards where possible, and flag missing fields such as detainer lift reasons or absent transfers that could change an individual’s apparent status (best practices recommended by the Deportation Data Project and data‑journalism guides) [2] [5]. Where the FOIA materials are silent—such as the precise timing of system updates or why particular records are absent—researchers should acknowledge those unknowns rather than infer final legal outcomes from the administrative labels alone [2].