What were official counts of children in ICE or family detention each year from 2016–2025?
Executive summary
The sources provided do not contain a clean, official year-by-year table of how many children were held in ICE custody or in family detention for each year 2016–2025; ICE publishes bi‑monthly detention statistics and related datasets, but the publicly posted summaries and third‑party analyses supplied here give only partial snapshots and facility‑level counts rather than a single annual child count series [1] [2] [3]. What follows is a clear accounting of what the reporting can and cannot confirm, plus specific datapoints available in the record and instructions for where to get the official series if the reader seeks the raw numbers (ICE’s statistics pages and datasets obtained by data projects) [4] [1] [3].
1. What the official public sources publish and their limits
ICE’s public statistics program posts bi‑monthly detention reports, alternatives‑to‑detention tallies, and facility statistics, and describes changes in program definitions (for example converting Family Residential Centers to Family Staging Centers in 2021 and shifting resources to Alternatives to Detention in FY2022), but these pages do not present a single, consolidated annual count of children in ICE or family detention for every year 2016–2025 in the materials supplied here; ICE itself notes that its datasets fluctuate until fiscal year “lock” and that it posts multiple report types rather than a simple calendar‑year child‑count table [4] [1] [2].
2. Concrete datapoints the sources do provide
Independent researchers and watchdogs have extracted and published specific figures from ICE releases: Vera’s analysis and reporting documents ICE facility use and daily populations through 2025 and cites facility‑level family detention activity (including Karnes and Dilley family centers reopening and holding large numbers in mid‑2025) [5] [6]. For example, Vera reported Karnes held 1,187 people and Dilley 575 on June 10, 2025—numbers that include family units rather than isolating children alone [6]. ICE’s public site and composite datasets (made available and re‑published by projects like the Deportation Data Project) are the primary sources for extracting annual totals if one compiles the bi‑monthly and facility reports into a year‑end figure [1] [3].
3. What investigative and advocacy reporting adds — and its caveats
Investigations have shown policy shifts and spikes in family detention during certain administrations: journalists and researchers have documented a revival of family detention and clusters of releases near court‑established benchmarks for holding children (the Marshall Project’s analysis found release spikes around the 20‑day Flores benchmark), but these pieces generally use ICE data or FOIA releases to calculate counts rather than supplying an uninterrupted official annual time series from 2016 through 2025 in the provided documents [7]. Such work is valuable but depends on how researchers define “children in ICE or family detention” (e.g., children counted inside Family Residential Centers, children held with detained parents, or children reported in staging centers or elsewhere).
4. Why a single authoritative annual series is hard to extract from the supplied reporting
The difficulty lies in shifting definitions, program changes, facility reclassifications, and ICE’s reporting cadence: ICE reclassified family facilities in 2021–2022, used staging centers and ATD enrollments variably, and posts bi‑monthly snapshots that require aggregation to assemble an annual total; third‑party repositories (Deportation Data Project, Vera) and NGO summaries exist and are the most reliable paths to construct a year‑by‑year series, but the specific year‑end child counts for 2016–2025 are not directly printed as a single table in the materials provided [4] [1] [5] [3].
5. Practical next steps to obtain the official annual counts
To produce precise annual counts for 2016–2025, compile ICE’s bi‑monthly detention facility reports and family‑center population files from ICE’s Detention Management/Statistics pages and the FOIA releases republished by the Deportation Data Project, then normalize changes in definitions (FRC → FSC, staging centers, ATD enrollments) as documented by ICE and third‑party technical notes; the relevant ICE pages and the Deportation Data Project are the primary sources to extract raw numbers, while Vera and other analyses provide helpful cleaned datasets and methodology notes [4] [1] [3] [5].