Aim: The aim of this study was to determine the prevalence of structural shoulder pathology using magnetic resonance imaging (MRI) in three groups of older people: those with current shoulder pain, those with a previous history of shoulder pain and those with no history of shoulder pain, within a community-based sample. Methods: Thirty subjects (10 within each of the three groups) participated in the study. Subjects were recruited by telephone and underwent a clinical examination of shoulder and neck range of movement (to ensure pain was not referred from the neck). Subjects completed the Shoulder Pain and Disability Index (SPADI) and underwent MRI and X-ray of the relevant shoulder. The X-rays and MRI were read independently by two experienced musculoskeletal radiologists blinded to each participant's symptoms. The MRIs were read using a structured reporting system. Results: The mean range of shoulder movement on both the right and left sides was lower for the current pain group compared to both the no and previous pain groups. On X-ray, there was no significant difference between groups in terms of glenohumeral and/or acromioclavicular degenerative changes. Tendinosis and tears of the rotator cuff were present in the majority of participants in each group. Labral abnormalities were rare among all groups. Conclusion: Shoulder pathology is apparent in both symptomatic and asymptomatic shoulders and clinical symptoms may not match radiological findings. The cost burden of ordering MRI scans is significant and the relevance of the findings are questionable when investigating shoulder pain.