“Simply incorrect” overstates what your CIS link shows.
Yes, status isn’t always known at arrest, and time-lag/unknown-status classification is a real measurement issue. But that’s not a demonstration that the cited studies are false; it's a methodological dispute about how Texas data should be interpreted.
Even CIS effectively concedes the key limitation: “any crime” conviction rates aren’t meaningful under their own description because identification is biased toward longer prison terms/serious offenses. That means their approach can’t legitimately be used as a general claim that “undocumented commit more crime.”
Yes, status isn’t always known at arrest, and time-lag/unknown-status classification is a real measurement issue. But that’s not a demonstration that the cited studies are false; it's a methodological dispute about how Texas data should be interpreted.
Even CIS effectively concedes the key limitation: “any crime” conviction rates aren’t meaningful under their own description because identification is biased toward longer prison terms/serious offenses. That means their approach can’t legitimately be used as a general claim that “undocumented commit more crime.”
Also, Texas is one of the few places where researchers do try to reconcile arrest/ID systems (e.g., Light et al., PNAS 2020): https://www.pnas.org/doi/10.1073/pnas.2014704117
And there are direct responses to CIS’s Texas framing (e.g., Cato 2024): https://www.cato.org/policy-analysis/illegal-immigrant-murde...
So: criticize uncertainty, sure, but “therefore the low-crime finding is simply incorrect” doesn’t follow.