How do researchers estimate crime rates among undocumented immigrants versus native‑born Americans?
Executive summary
Researchers estimate crime rates among undocumented immigrants versus native‑born Americans by combining administrative criminal records, population estimates and statistical adjustments to account for undercounting and selection; multiple recent studies find undocumented immigrants have lower arrest and incarceration rates than the U.S.‑born in the datasets examined [1] [2]. The strongest direct evidence comes from Texas arrest records that record immigration status and large national studies and reviews that use modeling or historical incarceration comparisons to reach similar conclusions [3] [4] [5].
1. How the raw data are built: arrests, convictions and incarceration records
Analysts start with criminal justice administrative records — arrest histories, conviction and incarceration data — because they are concrete events that can be counted, and the Texas Department of Public Safety data used in the PNAS study is an unusually complete example because it records immigration status for every arrestee [3] [6]. Those records let researchers calculate offense‑specific arrest rates per 100,000 people for U.S.‑born citizens, legal immigrants and those classified as undocumented in the data, and studies show robust differences across these outcomes when using alternative measures like misdemeanors, felonies and convictions [1] [6].
2. How population denominators are estimated and why that matters
To turn counts into rates, researchers need reliable population denominators for undocumented people, which are not directly measured in the census and must be estimated using survey‑based methods, undercount adjustments, or external estimates from groups like Pew or Center for Migration Studies; sensitivity to those assumptions is tested in robust studies and the Texas work reports that its core findings hold across alternative population estimates [1] [7]. Where national direct counts are unavailable, scholars use statistical modeling or extrapolate from smaller samples — a limitation highlighted by fact‑checking outlets and meta‑analyses [7].
3. Statistical adjustments and strategies to reduce bias
Studies address selection and measurement bias through several strategies: adjusting for underenumeration (assuming recent entrants are more likely undercounted), reclassifying ambiguous immigration status at arrest, substituting convictions for arrests, and testing alternative estimates of the undocumented population; the Texas study and related NIJ summaries report that the lower offending rates for undocumented immigrants are robust to these checks [1] [6] [8]. Broader research also uses regression models across states and cities to check whether changes in immigrant shares correlate with aggregate crime trends; these analyses commonly find no positive relationship between immigration levels and crime rates [9] [2].
4. What the best current empirical findings show
Using Texas data from 2012–2018, researchers reported felony arrest rates around 1,000 per 100,000 for U.S.‑born citizens, ~800 per 100,000 for legal immigrants, and ~400 per 100,000 for undocumented immigrants, and found undocumented arrest and conviction rates lower across violent, property and drug offenses [1] [3]. NIJ summaries and national studies echo that undocumented immigrants are arrested at substantially lower rates for violent and drug crimes and much lower for property crimes compared with the U.S.‑born [8] [10]. Long‑run research likewise finds immigrants have been less likely to be incarcerated than the U.S.‑born for over a century, with recent work concluding immigrants are substantially less likely to be incarcerated today [2] [11].
5. Limits, alternative findings and sources of uncertainty
Despite consistent findings in many studies, important caveats remain: nationwide direct measures of undocumented offending are scarce, so national estimates sometimes rely on modeling or extrapolation; some European studies and select U.S. subanalyses find different patterns under particular conditions, and a small body of work reports associations between undocumented populations and drug‑related arrests in specific contexts [7] [5]. Researchers explicitly note they cannot always explain causal mechanisms — hypotheses include selection (migrants self‑select for lower criminal propensity), deterrence (fear of deportation), and assimilation effects — and they urge further study to unpack these drivers [1] [5].
6. How policymakers and the public should read the evidence
The most direct, high‑quality data—such as Texas’ immigration‑tagged arrest records—consistently show lower arrest and incarceration rates among undocumented immigrants compared with the U.S.‑born, and national reviews and regression analyses generally find no evidence that larger immigrant shares raise crime rates; yet recognition of methodological limits and context‑specific exceptions is essential before generalizing to every jurisdiction or policy question [3] [9] [2] [8].