How do Deportation Data Project and TRAC classify removals by criminal conviction status?
Executive summary
The Transactional Records Access Clearinghouse (TRAC) classifies removals using case-level ICE records that explicitly include detailed criminal-history fields — most prominently a "Most Serious Criminal Conviction" (MSCC) and other criminal-history indicators — allowing TRAC to label removals as involving convicted individuals, non‑convicted immigration violators, or aggravated felons [1] [2]. The Deportation Data Project (DDP) republishes and documents government enforcement datasets (ICE, CBP, EOIR), leaving classification to the underlying agency fields and to users who typically must merge ICE removal records with EOIR or other prosecution datasets to determine conviction status; DDP also warns that some ICE removal tables have reliability problems and are sometimes withheld [3] [4] [5].
1. How TRAC builds conviction‑based categories: record‑level MSCC and criminal history
TRAC’s defining strength is its FOIA-driven, case‑by‑case reconstruction of ICE removal files that include detailed criminal history fields — TRAC explicitly exposes a "Most Serious Criminal Conviction" variable and other criminal‑history indicators in its removal data tools, enabling analysts to count removals by whether the individual had a conviction and by the severity/type of conviction such as aggravated felonies [1] [2]. TRAC notes it derives findings from the underlying ICE records it obtained through litigation and FOIA, and its public tools present removals broken down by criminal vs. non‑criminal grounds, by administrative removal types, and by apprehension program that can correlate with criminal‑justice involvement [1] [6]. TRAC also cautions users about inconsistent reporting from DHS across summary series and has, in the past, excluded certain DHS aggregate criminal counts when FOIA evidence suggested discrepancies — demonstrating TRAC’s emphasis on record‑level classification rather than blind reliance on agency summary tallies [2].
2. How the Deportation Data Project organizes and signals conviction information
The Deportation Data Project does not invent new conviction labels but republishes anonymized federal enforcement datasets (ICE, CBP, EOIR) and documents the fields available so researchers can classify removals themselves; its guide and data repository describe arrest dispositions, removal types (expedited, reinstatement, Title 42, standard removal) and point users toward EOIR CASE data for court outcomes and charges that can be used to infer conviction or adjudication status [3] [7]. DDP’s ICE documentation and data pages emphasize that their releases are "close‑to‑original" and that linked identifiers across arrests, detainers and detentions can allow merges to trace an individual’s enforcement pathway — a necessary step to determine whether a removal record corresponds to someone with a criminal conviction recorded elsewhere [4] [5]. DDP also acknowledges data quality limits: some ICE removals/encounters tables have been withheld or flagged for reliability problems, meaning conviction classification via DDP may be constrained by missing or suspect fields in those releases [5] [4].
3. Practical differences and consequences for analysis
In practice, TRAC delivers pre‑processed, research‑friendly breakdowns that explicitly tag removals by criminal conviction status (including MSCC and aggravated‑felony labels) because its FOIA litigation secured those granular ICE fields and TRAC has built interfaces and reports to exploit them [1] [2]. By contrast, DDP operates as a curated repository and documentation hub: it preserves agency fields and helps users perform their own merges with EOIR or LIONS prosecution datasets to determine convictions, but it does not centrally recode removals into the same conviction categories TRAC publishes — and it warns where ICE removal files are unreliable or withheld [8] [7] [5]. That difference matters when journalists, advocates, or researchers need quick counts of "removals with criminal convictions" — TRAC provides those ready outputs, DDP provides the raw materials and crosswalks.
4. Limits, caveats and where to look next
Both projects rely on underlying government records and therefore inherit agency labeling quirks: ICE’s internal labels for reasons for removal include non‑criminal grounds (inadmissibility, overstays) and criminal grounds (drug violations, aggravated felonies), but those labels and administrative removal procedures can blur the line between "convicted" and "deported for an offense" — an ambiguity TRAC documents and that makes careful, record‑level work essential [1] [6]. Users seeking defensible counts should consult TRAC’s MSCC fields and reporting for ready conviction classifications and use DDP’s datasets and documentation when they need the raw linkage files, but must heed DDP’s notes about removed or unreliable removal tables [1] [4] [5]. Where sources do not fully describe a specific classification step for DDP, this analysis reports those limits rather than presume them.