Thus, for example, once you get past the paywall, you will find that Piquero and Bersani's joint study, "Comparing Patterns and Predictors of Immigrant Offending Among a Sample of Adjudicated Youth," used as its base group "adolescents who were found guilty of a serious offense."

THAT'S NOT A REPRESENTATIVE SAMPLE OF AMERICANS! It's a representative sample of teenagers who are convicted criminals.

Similarly, professor Bersani's oft-cited, but never-read study, "An Examination of First and Second Generation Immigrant Offending Trajectories," looked at a population group that included "an over-sample of Hispanic and African-American youth."
Instead of immigrants who are less crime-prone than our native blacks and Hispanics, we were hoping for immigrants less criminal than our Norwegians.

True, as Bersani explains, "because many immigrants initially settle in disadvantaged environments and are exposed to a number of crime-inducing risk factors, their experiences may be similar to many native-born minorities—particularly the African-American population."
But here's an idea: How about NOT taking in immigrants who are poor, uneducated, come from dysfunctional families and settle in disadvantaged environments?