Neighborhood Characteristics by Census Tract, cross-walked to 2000 FIPS codes. Original data from “The Opportunity Atlas: Mapping the Childhood Roots of Social Mobility”, available at: https://opportunityinsights.org/data/.

County and State Covariates Related to Opioid Mortality. This dataset contains four county- and state-level covariates used in causal research on the opioid epidemic, covering 3,135 U.S. counties: (i) triplicate — Indicator for whether the state operated a triplicate prescription program at OxyContin's 1996 launch. Physicians in these states were far less likely to prescribe OxyContin, making this a standard instrument for early opioid exposure (Alpert, Evans, Lieber, and Powell 2022); (ii) cancer_rate_1996 — County cancer mortality rate in 1996. Purdue Pharma initially targeted the cancer pain market, so pre-launch cancer mortality predicts subsequent OxyContin distribution and is used as an instrument for local prescription opioid exposure (Arteaga and Barone 2025); (iii) ssdi_rate_1990 — County SSDI participation rate in 1990. Captures pre-existing chronic pain burden before OxyContin's launch and is a strong predictor of subsequent drug mortality growth (Cutler and Glaeser 2021; Evans and Lieber 2025); and (iv) ln_imports_per_capita_08 — Log state imports per capita in 2008, assigned to counties within each state. States with greater import throughput faced higher illicit fentanyl exposure through legal trade flows; the 2008 baseline predates the fentanyl era to avoid endogeneity (Moore, Olney, and Hansen 2023).

NHGIS Neighborhood Characteristics by Census Tract, 1970-2010.This dataset contains employment rates, poverty rates, and median income by census tract from 1970-2010.