State-Level Predictors of ICE Arrest Rates
Question
Is a state's 2024 presidential vote, racial composition, unauthorized immigrant share, or sanctuary status associated with its rate of ICE arrests?
Data and method
The unit of analysis is the state (n = 50). The outcome is ICE arrests per 100,000 residents from January 20 to May 20, 2025. Predictors are whether Trump won a plurality in 2024, the share of residents who are non-Hispanic White, the unauthorized immigrant share of the population, and whether the Department of Justice lists the state as a sanctuary jurisdiction. I used descriptive statistics, t-tests, correlations, and an ordinary least squares regression. Sources are in the codebook.
Findings
Each predictor was significantly associated with arrest rates on its own. In the combined regression (R² = 0.36), only the share of White residents remained significant at the 0.05 level, with a negative coefficient. Trump plurality was marginally significant (p = 0.08); unauthorized immigrant share and sanctuary status were not.
Limitations
With 50 observations, the model has limited statistical power. The data describe states, not individuals, and the analysis shows associations, not causes. The sanctuary list postdates the arrest period, and I did not test for outliers such as Texas or run other diagnostics.
Materials
- Statistical analysis (PDF): descriptive and bivariate results
- Regression slides (PDF)
- Codebook and data sources (PDF)
- Data (CSV)