A mother migrating from Honduras holds her 1-year-old as she surrenders to U.S. Border Patrol agents in Texas. (David J. Phillip/AP)

Central to President Trump’s 2016 campaign platform was changing how the country handled immigration. No longer, his campaign website pledged, would those entering the country illegally be released from custody before hearings to determine whether they should be deported. This was the impetus for the administration’s “zero tolerance” policy announced earlier this year, that the past practice of monitored release of immigrants crossing into the United States was coming to an end.

“You release him, and he never shows up again,” Trump said when talking to reporters last month. “He goes into our society, and then we end up getting him in a different way, oftentimes after he’s killed somebody.” In a speech a few days later, Trump estimated that only 3 percent of those released ended up appearing at the hearings meant to evaluate whether they should be deported.

On Tuesday, the New York Times reported that, in the wake of a legal decision limiting how long children could be held in detention, the government would again begin releasing families apprehended crossing the border illegally before their immigration hearings. Those released would be tracked with ankle bracelets, a process referred to by Immigration and Customs Enforcement as “alternative to detention,” or ATD.

It raises the question: How likely is it that those families who are released actually return for their hearings? Concrete data to answer that question is hard to come by.

According to the Department of Homeland Security, there was an “abscond rate” of 24 percent for alternative-to-detention participants from October 2017 through May 2018, meaning that they failed to appear when mandated. That apparently includes all participants in the program, though, including those awaiting deportation — not just those waiting for immigration hearings. As a tool for deporting people, the ATD program is neither effective or inexpensive, according to DHS, because only 1 percent of participants end up being removed from the country and because, while the daily cost is significantly lower than detention, removing ATD participants takes far longer than removing those in detention.

As a tool for ensuring that they attend court hearings, though? DHS says it’s effective.

How effective? When an immigrant fails to show up for a hearing, his or her case is evaluated in absentia. According to 2016 data from the Justice Department, 39 percent of immigration court decisions related to immigrants who had never been detained were determined in absentia.

Syracuse University’s Transactional Records Access Clearinghouse (TRAC) compiles data based on individual immigration court cases. TRAC’s data shows a smaller number: 17 percent of those who posted bond and about 20 percent of all cases were completed in absentia — meaning that 83 percent and 80 percent of immigrants, respectively, appeared as ordered.

That’s all cases, but TRAC notes two ways in which its methodology differs from the government’s in a footnote. First, the government doesn’t count all hearings in its calculations, excluding some based on the judge’s decision. Second, its data includes only the first hearing an immigrant is supposed to attend — so if another hearing is scheduled and the immigrant attends, the figure changes.

One argument made to suggest that immigrants have a disincentive to appear for a hearing once released is the low rate at which asylum is granted. (Most of those entering the United States make an asylum claim on entry.) If you hope to stay in the United States and recognize that the likelihood you’ll receive asylum is low — meaning you’ll be slated for removal — why bother starting the asylum process?

It’s true that most asylum seekers are rejected — particularly those lacking legal representation. TRAC’s data suggests that 62 percent of claims were rejected last year, including 90 percent of those who weren’t represented by an attorney.

Government data, though, has a higher rate of rejection, indicating that only 22 percent of claims made from October 2017 to the end of March were granted in hearings. (Those are not new entries during that period; there’s a massive 700,000-case backlog of immigration hearings that’s being worked through.) That percentage, though, includes cases resulting in something other than a granting or denial of asylum. Considering only cases in which one of those outcomes occurred, 35 percent of claims were granted — similar to TRAC’s numbers.

Again, though, those figures lack nuance. Attorney General Jeff Sessions announced earlier this year that the government was scaling back the conditions under which asylum could be sought, including denying claims based on fear of gang violence. Many of those arriving at the border with Mexico are fleeing precisely such violence in Central America. Their claims may be much less likely to be granted now.

It’s safe to say, though, that the president’s rhetoric about releasing immigrants before hearings is inaccurate. Most of those released appear as ordered, even if the results are likely to be that they are eventually ordered to be deported.

It’s also not the case — at all — that those who are arrested by ICE after absconding are caught after committing a murder. Data from ICE indicates that the most common crime committed by those arrested by its agents is drunken driving, followed by drug crimes and, unsurprisingly, immigration offenses. Homicide is one of the least common crimes.

On this, too, Trump’s repeated rhetoric far outpaces reality.