The diagnosis of a stroke in the emergency department (ED) can be difficult when a patient presents primarily with dizziness. Approximately 1/3 of such patients are misdiagnosed with a benign ear condition. While MRI may detect a stroke in such patients, it may not be readily available in many ED. Therefore, there is a need to improve clinical recognition of strokes among acutely dizzy patients. The “bucket test” (BT) involves the use of a bucket that has inside of it a series of images arranged in a straight line that can be aligned with earth’s vertical. In research lab settings, more than 90% of patients with posterior circulation strokes misalign the images off the vertical. However, the utility of a portable version of BT in diagnosing stroke among dizzy patients in ED setting has not been evaluated. We hypothesized that BT can be used by ED physicians as a screening tool for stroke in dizzy patients. We tested a portable version of BT on dizzy patients in ED of a tertiary center. We prospectively enrolled 81 acutely dizzy patients presenting to ED between February 2012 and February 2015. Patients also had brain imaging (CT in 5, MRI in 22). Using MRI/stroke specialist as reference, a total of 7 patients had stroke/TIA. All had clearly abnormal BT and no patient with a stroke/TIA had a normal/equivocal BT (BT sensitivity 100%; 95% CI 59 to 100%). Meanwhile, MRI demonstrated stroke in 5/7 of stroke/TIA patients. One patient with negative initial MRI was diagnosed with AICA stroke later by blinded stroke specialist. And one patient had TIA, with initial negative CT, only to return to ED 12 days later with more dizziness and brain MRI then showed an acute stroke. Therefore, the sensitivity of brain MRI was 83%, consistent with prior literature. In sum, the bucket test is a simple bedside screening test that may improve the diagnosis of stroke in the emergently dizzy patient.