An Entity-Resolved Crosswalk Between U.S. Department of Labor H-1B Labor Condition Applications and USCIS Petition Outcomes, FY2019-FY2026
收藏资源简介:
Employer-level crosswalk linking the Department of Labor's Labor ConditionApplication disclosure files (Form ETA-9035/9035E) to the USCIS H-1B EmployerData Hub. The two federal sources record different stages of the H-1Bsponsorship process, certified intent and adjudicated outcome respectively, andshare no identifier usable to join them. Linkage is name-based and evaluated against a machine-verified referencestandard built from agency-assigned Employer Identification Numbers. Held outfrom the EIN feature, which is the configuration applying to the fiscal yearswhere DOL publishes no FEIN, linkage achieves F1 = 0.9446 (FY2025) and 0.9473(FY2026). Files are released at match threshold 0.55, where held-out precisionis 0.9521, admitting an edge only when state or EIN last-four agrees. FY2025 contains 46,775 links over 44,263 distinct DOL employers. FY2026contains 14,621 links. Each file ships with a manifest recording row counts,operating point, held-out precision and a SHA-256 digest. See DATASHEET.md forcomposition, collection, preprocessing and intended-use notes.



