Correspondence experiments are the gold standard for measuring hiring discrimination. However, two concerns persist. Callbacks measured in correspondence studies only track the initial interest of employers and not hiring itself, and the distinctive names that signal race in these experiments may signal socioeconomic status as well as race. We address both concerns by matching the firms in Kline, Rose, and Walters (2022)’s experiment to LinkedIn employment records of 9 million employees. We find that firms with larger callback gaps employ fewer Black workers. The employment gap is three times what callback gaps would predict, consistent with discrimination continuing beyond callbacks or with Black applicants sorting away from discriminatory firms. The relationship is strongest among the most discriminatory firms.. About 3% of the Black workers we identify carry the experiment’s distinctive names, yet callback gaps predict employment differences for the broader Black workforce. We release a firm-by-year panel of employment shares by race and gender for the audited firms and all Fortune 1000 firms from 2010 to 2024, together with our race-classification software, so that future correspondence experiments can be evaluated against hiring, not only callbacks.