The Comprehensive Courts of Appeals Database (CCAD) is available on Harvard Dataverse. Individual files for download are below.
Data Files
In between Dataverse updates, individual file links on this page point to temporary Dropbox copies.
- Stata (.dta.gz) — 949.9 MB
- CSV (.csv.gz) — 777.9 MB
- Apache Parquet (.tar.gz) — 878.5 MB
- R (.RData.gz) — 627.2 MB
Crosswalk Files
Coverage
The database currently covers all published opinions issued by the U.S. Courts of Appeals from 1892 through 2025. Variables are drawn from two sources. Case metadata — including case identifiers, opinion types, and panel composition — are derived directly from digitized copies of the Federal Reporter (via the Caselaw Access Project and CourtListener) and the FJC Biographical Directory of Article III Judges. Substantive annotations — including issue area, disposition, litigant types, and ideological direction codings — are calibrated to the conventions of the Songer U.S. Courts of Appeals Database and the Supreme Court Database (SCDB) and were produced using a large language model pipeline validated against both ground-truth datasets.
Updates
The database will be updated on an annual basis to incorporate newly issued opinions. Version history will be maintained at the Harvard Dataverse deposit.