The nation's most-cited independent immigration database appears to have swapped a key metric, shifting from tracking detainees with a "criminal record" to counting only those with "criminal convictions", a change that strips out anyone with pending charges and makes federal enforcement look far less focused on public safety threats than the underlying numbers suggest.
The Transactional Records Access Clearinghouse, known as TRAC, is a research project housed at Syracuse University since 1989. Media outlets, fact-checkers, and advocacy groups treat it as the gold standard for federal enforcement data. Just the News reported that since President Donald Trump took office, TRAC changed how it analyzed immigration enforcement data, emphasizing criminal convictions rather than the broader category of criminal records it had previously highlighted.
The practical effect is significant. TRAC's "Immigration Detention Quick Facts" page now states that ICE holds 68,289 individuals in detention and that 73.6% have "no criminal convictions." On its face, that figure suggests nearly three out of four detainees have no brush with the criminal justice system at all. But TRAC's own underlying database tells a different story: a further 26% of those detainees have pending criminal charges that no longer appear in the headline number.
An archived version of the TRAC website from February 2025 shows that at that time, 45.3% of ICE detainees had a "criminal record", a category that included both convictions and pending charges. Under the old methodology, anyone facing criminal proceedings counted.
Under the new framing, only finalized convictions qualify. That single definitional change turned a population where roughly half had criminal entanglements into one where barely a quarter does, not because anyone was released or cleared, but because the label moved.
Combine the conviction and pending-charge categories from TRAC's own current data, and approximately 52% of detainees have a criminal record of some kind. That figure is higher than the 45.3% recorded last February, a period when the detained population was smaller and enforcement operations were still ramping up. Yet the number now featured on TRAC's quick-facts page, 73.6% with "no criminal convictions", paints the opposite picture.
The gap between those two framings is not a rounding error. It is the difference between a narrative in which most detainees are ordinary immigration violators and one in which a majority have been flagged by the criminal justice system.
Sean Kennedy, policy director of the Law Enforcement Legal Defense Fund, described the shift on the Just the News, No Noise TV show:
"They changed their own methodology, TRAC did at Syracuse, which is, they used to say anyone with a criminal record or criminal arrest was in the criminal category, and then they would break it down and say, he's got a criminal conviction, he's under pending charges, or whatever."
Kennedy went further, pointing to how the revised numbers travel from TRAC into mainstream reporting:
"They just took the pending charges number out. So when CNN goes and reports it out, they say, oh, only 30% have convictions. What? What a low bar."
That chain, from TRAC's database to CNN's broadcast desk, matters because the data is not presented with a disclaimer about the methodology change. Viewers hear "only 30% have convictions" and draw the obvious conclusion: the government is rounding up people who have done nothing wrong beyond entering the country illegally.
The pattern extends beyond cable news. FactCheck.org published an article in January that specifically cited the declining proportion of detainees with criminal convictions as evidence of what it called a "transition toward arresting a higher percentage of immigrants with no criminal record." The same piece argued that "the Trump administration insists that it is targeting the 'worst of the worst' in its immigration enforcement" but "has not provided information to substantiate that, and the data that is available suggests the claim has become less accurate over time."
That conclusion rests on data shaped by the very definitional change TRAC made. If pending charges were still included in the headline category, as they were a year ago, the numbers would tell a materially different story.
TRAC is not the only tracker that moved the goalposts. The Berkeley Deportation Data Project, another widely cited source, adopted the same shift in terminology, according to the Just the News report. Two of the most prominent independent databases used to scrutinize federal immigration enforcement now present figures that exclude anyone awaiting trial.
Questions about how government data gets packaged and presented are not new. A recent inspector general report found the Biden administration bypassed procurement rules to steer a $529 million contract to a nonprofit with insider connections, a reminder that the infrastructure around immigration policy has long been shaped by actors with interests that extend well beyond neutral record-keeping.
Kennedy framed the definitional shift as part of a deliberate effort to minimize the criminal profile of the detained population:
"[W]hat we're seeing in these cases is they're defining down what it means to be a criminal alien."
That phrase, "defining down", captures the mechanism precisely. Nobody falsified a number. Nobody invented a detainee who doesn't exist. The data itself may be accurate. But the category that gets the spotlight shrank, and the category that got hidden grew, and the net effect is a public narrative that diverges from the fuller picture sitting one click deeper in the same database.
The distinction between a conviction and a pending charge is real in a courtroom. A defendant is presumed innocent until proven guilty. But in the context of immigration enforcement statistics, the question is different. The public is not being asked to convict anyone. It is being asked whether federal agents are arresting people with no criminal entanglements at all, or people who are actively facing prosecution.
Someone with a pending assault charge, a pending drug trafficking indictment, or a pending DUI case is not the same as someone whose only legal issue is an immigration violation. Lumping the two together under "no criminal convictions" obscures a distinction that matters to every community where these detainees were living before ICE picked them up.
The question of who benefits from muddied enforcement data runs parallel to other accountability gaps in federal contracting and spending. A Vance-led anti-fraud task force recently flagged $6.3 billion in suspect government contracts, a sign that the appetite for oversight is growing, and that the systems producing official numbers deserve the same scrutiny as the systems spending official dollars.
TRAC did not immediately respond to a request for comment from Just the News. No public statement explaining the methodology change has been identified. The timing, after Trump took office and began expanding enforcement, raises an obvious question about whether the shift was analytical or political.
It is fair to note that TRAC may have a defensible statistical reason for narrowing the category. Convictions are a harder, more legally settled data point than pending charges. But if that were the rationale, one would expect a transparent disclosure, a methodological note, a press release, an asterisk on the quick-facts page. None has surfaced.
Meanwhile, the political ecosystem that relies on TRAC data has not paused to ask questions. Fact-checkers built arguments on the new numbers. Cable networks broadcast them. Advocacy groups cited them. The result is a feedback loop in which a quiet definitional change at a university research project shapes the national debate over whether immigration enforcement is lawful, proportionate, and honest.
The financial and political incentives surrounding immigration enforcement data deserve closer examination. Democratic campaign groups have quietly cashed checks from an ICE contractor while publicly blasting federal detention operations, a contradiction that suggests the immigration debate is not always driven by the principles its loudest voices claim.
And the broader landscape of institutional accountability continues to shift. A DHS inspector general probe into billions in no-bid contracts shows that oversight mechanisms are being activated across the federal government, a development that ought to extend to the data pipelines that inform public understanding of enforcement policy.
If TRAC's quick-facts page still used the "criminal record" category it featured last February, the headline number would show that roughly half of ICE detainees have some criminal entanglement, not that three-quarters are clean. The difference is not subtle. It is the difference between a policy that looks indiscriminate and one that looks targeted.
The Trump administration has described its enforcement posture as focused on the "worst of the worst." Critics have used TRAC's revised numbers to argue otherwise. But the critics' case depends on a metric that was quietly narrowed, and the fuller data, available on the same website, tells a story closer to what the administration claims.
None of this means every ICE arrest is justified or every detainee is dangerous. It means the public deserves data presented honestly, with consistent definitions, transparent methodology, and no quiet switches timed to a change in administration.
When the institutions that grade the government's homework start changing the answer key, the grade tells you more about the grader than the student.
