After fleeing Central America, child migrants now face a very uneven brand of justice in U.S. immigration courts, where access to a lawyer and the location of the court itself are often the deciding variables in who stays and who doesn’t.

A POLITICO analysis of government data shows that fully 88 percent of the removal orders issued since July have gone to children without an attorney. What’s more, a juvenile assigned to judges in Texas, North Carolina and Georgia in the past 16 months was at least three times more likely to receive such an order than a child in California, Florida or New York.

The same humanitarian crisis that gripped the White House and Congress last year is now a less visible one of American justice. And it still poses a major test for the Obama administration, which promised compassion and fairness in the treatment of the child migrants but also contributed to the current problem by expediting deportation proceedings.

The analysis shows that instead of a relatively uniform system of federally administered justice, what emerges is more of a patchwork quilt.

“I’m baffled by why the government can’t speak with one voice. It’s the luck of the draw,” said Rebeca Salmon, an immigrant defense attorney in Atlanta. “The worst part is that you land in Georgia vs. landing in California and you’re going to be 90 percent more likely to lose all your due process protections and be deported. That is a huge problem at the federal level.”

Since last summer, substantial progress has been made in finding attorneys for the children, and the cases of the great majority are still pending. But thousands are still without counsel. And thousands more, according to the data, have already been caught in the undertow, forced to live now under the threat of removal orders that will kick in with true force once they turn 18.

President Barack Obama’s decision last July to greatly expedite deportation proceedings — before he had secured money for counsel — has had a lasting impact.

The stated goal was to send a signal to discourage more children from coming. But critics say the speed of these so-called rocket dockets overwhelmed the small community of defense attorneys, and individual judges seized the chance to quickly issue removal orders or pressure the children to accept voluntary departure.

How much this “get tough” stance by Obama truly affected the flow of future migrants is still a matter of debate. The pace of border crossings is down from last year. But the credit may go as much to Mexico, which has greatly increased its own deportations of Central Americans attempting to cross Mexico to reach the United States.

There’s no doubt, however, that Obama’s decision has had a dramatic effect on the immigration courts — and the children already in government hands.

Cases were moved to the front of the line so that master calendar hearings or arraignments could be held within 21 days of a child being turned over to the courts. The pace was exceptional for a system that was already underfunded and struggling with a major backlog of prior cases. From mid-July to Christmas, there were 24,333 arraignment hearings for unaccompanied children, more than 1,000 per week.

This record is central to a federal lawsuit coming to a head before U.S. District Court Judge Thomas Zilly in Seattle. Closing arguments are slated for Friday.

Brought on behalf of 11 child migrants, ages 3 to 17, the suit names Attorney General Eric Holder as the chief defendant. It argues that “to fulfill its statutory and constitutional obligations, the Government must ensure that no child faces the life-altering prospect of deportation without legal representation.”

Holder himself told Congress in 2013 that for a child to be without counsel in a deportation proceeding is “simply not who we are as a nation.” And in oral arguments on Sept. 3, Justice Department attorneys insisted there was never any intent to rush the children through the immigration courts without allowing time for them to try to obtain legal counsel.

But the results on the ground have often been quite the opposite, according to the data. Despite millions of dollars in new funding for more lawyers, many children are still being asked to make difficult legal decisions without counsel to advise them.

“I know the attorney general’s office is saying there’s no pressure, they can continue [the cases] as long as they want, but I don’t think anybody’s seeing that on the ground,” said Mark Bowers, a staff attorney with Legal Services of Southern Piedmont in Charlotte, North Carolina. “Things have gotten better. … But at some point you feel like you’re a triage surgeon in the battlefield. There’s only so much you can do, and there are so many of them.”

To measure this, POLITICO examined data provided directly by the Executive Office of Immigration Review, which oversees the courts, as well as a broader set of numbers also collected from EOIR by the Transactional Records Access Clearinghouse, a nonprofit based at Syracuse University in New York.

Any such analysis is complicated by the fact that EOIR began a new coding system for the unaccompanied children in the middle of the crisis last summer. But taken together, the sample — cross-checked against other reports by the Department of Homeland Security — offers a fair benchmark for the cases over the past 16 months.

TRAC, established in 1989, has a respected record of collecting public government data from law enforcement and domestic security agencies and making it available under the Freedom of Information Act. In the ongoing immigration debate, TRAC’s public website includes tools that allow a visitor to sort through its data from EOIR and assess how much variation there has been from one state to the next.

Working with these numbers, POLITICO looked at 26 states in which the immigration courts handled the vast bulk of the more than 62,300 juvenile cases included in TRAC’s data from EOIR for all of fiscal 2014 and the first months of fiscal 2015.

Ninety-three percent of the children came from Central America, with Honduras, El Salvador and Guatemala leading the way. These totals closely track DHS data on the number of unaccompanied children encountered in the same period.

The sample of court cases inside the U.S. covers a far larger area than just those 26 states, since the jurisdiction of immigration courts in one state can cross into others. For example, the Dallas immigration court handles migrant cases not just in northern Texas but also Oklahoma. The Atlanta court reaches into Alabama. Charlotte extends into South Carolina.

Taken as a whole, the state-by-state breakdown provides a snapshot of how different the outcomes can be for children across the nation. The names of the immigration judges, like those of the children, are protected.

POLITICO’s focus was on two measures: the level of legal counsel provided to the children and what share of the cases ended with removal orders or voluntary departure.

Having legal counsel emerges as a huge asset for the children. Many have difficulty with English, and some from Guatemala speak a Mayan dialect, rather than Spanish. The government always has a lawyer on its side; the child without counsel is at an immense disadvantage.

While most of the 2014-15 cases are still pending in the courts, about 20 percent have been completed. The outcomes are stark: Nine out of 10 juveniles subjected to a removal order or voluntary departure had no lawyer. Seven out of 10 who did have a lawyer won decisions allowing them to stay longer in the U.S.

In Massachusetts, where about half the children have had legal counsel, 6 percent of the cases have ended thus far with orders of removal or voluntary departure. In Illinois, where more than 80 percent of the children lacked counsel, the data show that about 21 percent of the cases ended with these tougher results.

The legal risks of not having a lawyer are aggravated by the very different approaches children face from one court to the next.

North Carolina tops the list with the toughest outcomes: An estimated 34 percent of the juvenile defendants ended up either subject to an order of removal or agreeing to voluntary departure. Georgia and Texas trailed closely at 30 percent and 29 percent, respectively.

That’s far higher than the average national rate of 15 percent for such outcomes. Courts in a diverse mix of states — Maryland, Pennsylvania, Missouri, Oregon, Arizona, Nevada and Tennessee — all come close to that national average. And there is a spread on each side: Minnesota, like Illinois, is higher; Virginia and New Jersey stand out for being much lower.

In response to questions from POLITICO, EOIR said it takes “seriously the need to fairly and accurately apply the law to the facts of each case before the immigration courts.”

But the agency did not address the specific evidence that some courts have been much quicker than others to issue removal orders. EOIR said it “takes steps to evaluate disparities in immigration adjudications” but cautioned that “perceived or actual disparities often result from variations in the law applicable to the type of case, the facts presented, and a host of other factors.”

But even allowing for some margin of error in the data, child defendants in Texas, Georgia and North Carolina are far more likely to receive a removal order or agree to voluntary departure.

From a national perspective, immigration judges in these three states handled about 26 percent of all juvenile cases in the past 16 months. But their share of all the removal or voluntary departure orders in the same period was double that, about 53 percent.

Indeed, if the combined totals for Texas and North Carolina are paired against those for California and Florida — two other big migrant states — it can seem like two worlds.