The aim of this study was to assess the value of rigid endoscopy in patients presenting with globus symptoms. We conducted a retrospective analysis of 250 patients who underwent rigid endoscopy for globus symptoms over a 12-month period. In 217 patients (86.8 per cent) the examination of the larynx, pharynx and upper oesophagus was entirely normal. Abnormal findings included cricopharyngeal spasm in 12 patients (4.8 per cent), reflux in 11 (4.4 per cent), pharyngitis in two (0.8 per cent), web in two (0.8 per cent), and retention cyst in three (1.2 per cent). The 95 per cent confidence interval (CI) for the mean number of persons with malignancy based on the Poisson distribution is 0 and 3.7 (0 and 14.8 as rates per 1000). The relationship between the clinical diagnosis and endoscopic findings was examined using the chi-square test, with a p value of 0.0001. These results suggest that patients presenting with globus sensation are unlikely to harbour neoplastic lesions and therefore rigid endoscopy may well be an inappropriate investigation in this group. The risks, costs and discomfort associated with this intervention can often be avoided.