ICE Arrests of children

ICE arrested more than twice as many children during the first half of 2026 compared to the same time period the previous year. Starting with Trump’s return to office, child arrests climbed steadily through July, the latest full month with available data. That month, more than 1,100 children were arrested.

Most immigration violations are civil, not criminal, and the vast majority of children counted in ICE’s arrest data have no criminal conviction or pending charges beyond the immigration violation.

ICE Detentions of children

The daily population of children in ICE detention has risen sharply since family detention was reinstated in early 2025. Since Trump’s return to office, ICE holds a daily average of roughly 200 children in custody compared to just 23 the previous year — an eightfold rise.

Consequences for children following release

Under previous presidential administrations, children in ICE detention were most likely to be granted release into the U.S. through orders of recognizance, parole or other discretionary forms of relief. In Trump’s second term, however, most were either deported or left the country as a “voluntary” departure.

Deportations of children

Over 22,000 children have been deported since Trump returned to office. These deportations most frequently happen shortly after the child has crossed the border into the U.S. But children can also be deported after ICE agents or other immigration enforcement officers take them from the communities where they may have been living, often for years, with their families. These types of deportations, which are classified as being from the country’s interior, have risen dramatically, far surpassing levels seen over the past several years, including Trump’s first presidential term.

Methodology

Our analysis is based on official government data obtained by the Deportation Data Project through public information requests to ICE.

All large government datasets, especially those aggregated from disparate sources, inevitably have some errors in individual records. However, based on its internal verification processes, the Deportation Data Project characterizes the data as painting a broadly reliable picture of the government’s understanding of its immigration enforcement system.

Children, in our analysis, are defined as under the age of 18. Babies and toddlers are defined as under the age of 4. The Deportation Data Project’s records indicate someone’s birth year, but not the exact date of birth. We made conservative estimates using the time between a given date of arrest, detention or removal and Jan. 1 of the person’s birth year, representing the oldest this person could be on a particular date. As a result, our assessments for the number of children who were arrested, detained or deported are almost certainly undercounts.

Our counts of the total number of children arrested represent the number of individual children who were arrested. If the same child was arrested twice, they would count as one child in the total number of children arrested. Conversely, the number of child arrests over time represents the number of times a child was arrested. If the same child is arrested twice, they are counted twice in counts of arrests.

Arrests are attributed to the second Trump administration when the apprehension date is Jan. 20, 2025, or later.

Detentions are characterized in terms of stays and stints. A stint is a period of time a person spends in a single detention facility. Since people are often transferred between different facilities as part of a continuous term of confinement, each associated stint is also grouped into a larger stay, which represents the overall period of detention.

The total number shown of children detained under the second Trump administration includes those with stay book-in dates starting Jan. 20, 2025, or later, and represents the number of individuals who were detained. If the same child has more than one stay in this timeframe, they are only counted once.

The daily number of child detentions represents the number of individual children in custody on a given day. Each child is counted each day they are in custody. If the same child is transferred among multiple facilities on the same day, that child is just counted once for that day.

The total number of child arrests and detentions are not necessarily equal. It is possible for a child to be detained by ICE without being arrested by the agency, or for a child to be arrested without being detained.

Releases from detention are categorized into broad types based on the reason for the stay release. Each stay release is counted once on the final book-out day of the stay. If a child has more than one stay in ICE detention, each release will be counted separately. Stint book-ins and book-outs between facilities are not counted as releases.

The total number of child deportations includes every time a child was removed from the country. If the same child was deported, reentered the country, and then deported again, they will be counted twice.

Removals and Title 42 returns are categorized as deportations. “Voluntary departures” and “voluntary returns” are both categorized as leaving the country voluntarily. Orders of recognizance, parole and other discretionary forms of relief are categorized as releases into the U.S.

We classify deportations as either originating from the border or from the interior of the country. Most deportations are in the former category, taking place during border crossing encounters or when someone crosses the border and willingly identifies themselves to U.S. officials. Deportations can also originate from the interior of the country, after immigration officers take people in from the communities where they may have been living.

We infer the type of each deportation using the data’s “final program” field. When the final program is labeled as Border Patrol, or as land, sea or air inspections, we assume the deportation originates at the border. If the final program mentions interior programs designed to apprehend people who have not just presented at the border — like Enforcement and Removal Operations, Homeland Security Investigations, the 287(g) Program, or other similar arresting agency programs — we assume the deportation originated from the interior of the country.

The government doesn’t provide detailed documentation of all the paths that can result in a deportation. As such, these numbers represent our best estimates of how many interior and border deportations are taking place. The final program field in our data was left empty in less than 0.2% of child removals during the time period we examined.

A deportation is attributed to the second Trump administration when the date of departure is Jan. 20, 2025, or later.