How many lawful permanent residents (green card holders) are recorded in ICE arrest datasets and where can those fields be accessed?
Executive summary
The public sources reviewed do not publish a single authoritative aggregate figure for how many lawful permanent residents (LPRs, “green card” holders) appear in ICE arrest datasets; instead, researchers must download person‑level arrest records and compute counts themselves because the datasets and documentation that have been released include fields for citizenship and immigration status but provide raw, event‑level rows rather than precomputed LPR totals [1] [2] [3].
1. What the user is really asking and what the records contain
The question seeks a simple numeric answer and the data locations for verification, but the available FOIA‑released ICE datasets are structured as event‑level tables (one row per arrest/encounter) with variables that allow analysts to classify an individual as an LPR or by country of citizenship; the Deportation Data Project hosts the original individual‑level ICE arrest files and accompanying documentation explaining the tables and fields [2] [1], and ICE and DHS statistical portals publish related summary dashboards and monthly tables that break out arrests by citizenship and other status categories [3] [4].
2. Why there is no single pre‑published “how many LPRs” number in these sources
The Deportation Data Project posts the raw individual records it obtained through FOIA rather than canned summary statistics, and its documentation notes that it does not produce counts or reports itself; researchers are expected to merge and filter records (for example, by an immigration‑status variable or citizenship field) to derive totals of interest [1] [2]. ICE’s public ERO dashboards and DHS/OHSS monthly tables focus on citizenship and criminality breakdowns but do not present a single nationwide LPR arrest total in the downloadable arrest dataset documentation that the Deportation Data Project republishes [3] [4].
3. Where to access the relevant files and which fields to use
To produce an LPR count, download the ICE arrest/apprehension datasets and their field documentation from the Deportation Data Project data pages (the arrest datasets and the documentation pages are explicitly posted there) and from the Deportation Data Project’s general data repository [2] [5]. Complementary official sources include ICE’s Enforcement and Removal Operations statistics dashboards (which provide arrest breakdowns by citizenship) and DHS/OHSS monthly tables and immigration statistical tables where “lawful permanent residents (LPRs)” are defined and used in reporting [3] [4] [6]. The documentation pages on the Deportation Data Project describe the available variables so analysts can identify status/citizenship fields to filter for LPRs; because the project posts both raw and processed versions, those codebooks are the authoritative starting point [1] [2].
4. Practical next steps and caveats for anyone calculating the count
Researchers should download the arrest dataset (event‑level rows) from Deportation Data Project, open the dataset’s codebook to identify the exact field name that record immigration status or class (the project’s documentation explains the tables and fields), and then filter or aggregate records matching the LPR‑status code to compute totals; the sources emphasize that records may be duplicated across table types and must be de‑duplicated or linked using provided identifiers when tracing an individual across arrests, detainers, and detentions [1] [2]. Users should also check ICE’s and OHSS’s official dashboards and monthly tables for complementary breakdowns by citizenship and criminality, but note that those dashboards focus on citizenship and event counts rather than an explicit precomputed LPR arrest total in the public releases [3] [4]. Finally, the public reporting and third‑party visualizations (Vera, TRAC, news outlets) rely on these same underlying files and may differ in methodology, so any published LPR count should document the dataset version, filters, deduplication steps, and field names used [7] [8] [9].