Background Secondary prevention of ischemic stroke (IS) requires adequate diagnostic evaluation to identify the likely etiologic subtype. We describe hospital‐level variability in diagnostic testing and IS subtyping in a large nationwide registry. Methods and Results We used the GWTG‐Stroke (Get With The Guidelines–Stroke) registry to identify patients hospitalized with a diagnosis of acute IS at 1906 hospitals between January 1, 2016, and September 30, 2017. We compared the documentation rates and presence of risk factors, diagnostic testing, achievement/quality measures, and outcomes between patients with and without reported IS subtype. Recording of diagnostic evaluation was optional in all IS subtypes except cryptogenic, where it was required. Of 607 563 patients with IS, etiologic IS subtype was documented in 57.4% and missing in 42.6%. Both the rate of missing stroke pathogenesis and the proportion of cryptogenic strokes were highly variable across hospitals. Patients missing stroke pathogenesis less frequently had documentation of risk factors, evidence‐based interventions, or discharge to home. The reported rates of major diagnostic testing, including echocardiography, carotid and intracranial vascular imaging, and short‐term cardiac monitoring were <50% in patients with documented IS pathogenesis, although these variables were missing in >40% of patients. Long‐term cardiac rhythm monitoring was rarely reported, even in cryptogenic stroke. Conclusions Reporting of IS etiologic subtype and supporting diagnostic testing was low overall, with high rates of missing optional data. Improvement in the capture of these data elements is needed to identify opportunities for quality improvement in the diagnostic evaluation and secondary prevention of stroke.