With high-stakes midterm elections looming, US President Donald Trump has placed immigration front and center, doubling down on his claims in speeches and campaign ads tying immigrants to crime.
But many researchers argue that immigrants are, in fact, less likely to commit crimes.
Here's a look at some key statistics, where they come from and what people on different sides of the debate say about them.
The number of immigrants ICE arrested in fiscal year 2017 who were convicted criminals. ICE says it's committed to targeting "criminal aliens" for arrest and removal. Advocates accuse ICE of inflating statistics about criminals by including minor charges such as traffic offenses in its tally.
The number of inmates in state and federal prisons who are not US citizens, according to the latest prison population report from the Bureau of Justice Statistics. That's less than 6% of the total prison population. The report says that the total number of inmates who are not US citizens could, in fact, be higher, noting that some states may have undercounted this figure and others did not provide information. A past report from the bureau cautioned that different jurisdictions use different definitions for "noncitizens," at times basing the statistic on where people were born regardless of their citizenship status.
Administration officials have noted that roughly one in five inmates in federal prison are foreign-born, and that the vast majority of those people are in the United States illegally. They argue such statistics show the dangers illegal immigration poses to society. "The illegal immigrant crime rate in this country should be zero," US Attorney General Jeff Sessions said in a statement earlier this year. "Every crime committed by an illegal alien is, by definition, a crime that should have been prevented."
Advocates have slammed statistics released by the administration. "The vast majority of immigrants in federal prison are there for crimes that only immigrants can be charged with -- illegal entry and illegal entry after removal," Tom Jawetz, the vice president for immigration policy at the Center for American Progress, told The New York Times last year. "When you cook the books you shouldn't pretend to be surprised by the results."
The criminal conviction rate per 100,000 immigrants in Texas. This number comes from a study by the Cato Institute, which uses figures from Texas in 2015 as a case study to look at how crime rates compare among immigrant and native-born populations.
"The criminal conviction and arrest rates for immigrants were well below those of native-born Americans," Cato's Alex Nowrasteh wrote.
According to the libertarian think tank's analysis, the rate per 100,000 residents in each subpopulation was 899 for undocumented immigrants, 611 for legal immigrants and 1,797 for native-born Americans.
This is how much more likely undocumented immigrants are to be convicted of a crime than other Arizonans, according to a study of data in that state by economist John Lott of the Crime Prevention Research Center.
This study has been pointed to by the Justice Department and by a number of politicians and organizations who favor restricting immigration. It's also drawn sharp criticism from the Cato Institute and other researchers, who argue that the study misinterpreted data. Lott has stood by his research.
An analysis of the study and critiques by The Washington Post's Fact Checker was inconclusive.
The number of years of crime and immigration data researchers looked at in a recent study published in the journal Criminology. "Increases in the undocumented immigrant population within states are associated with significant decreases in the prevalence of violence," authors Michael Light and Ty Miller wrote. An analysis of data from 1990-2014, they said, suggests "undocumented immigration over this period is generally associated with decreasing violent crime."
The authors noted that more research needs to be done, and that it's possible the data could be interpreted differently.
"Because we are using official crime statistics," they wrote, "there is a plausible alternative interpretation: Increased unauthorized immigration results in fewer crimes reported to the police."
In a release from the University of Wisconsin-Madison announcing the study's findings, Light, a sociology professor, said it's important to look at data as the immigration debate continues.
"I think public debate divorced entirely from data is problematic," he said. "There are reasons to think undocumented immigrants would increase crime, but the data doesn't support those arguments. It's telling us that has not been the case."