Background and purpose: Reports on shoulder function after non-surgical treatment are not available. In the present study shoulder morbidity after surgical and non-surgical treatment of the neck is determined and compared. Materials and methods: In 100 head and neck cancer patients 174 neck sides were treated by surgery (n = 51) or (chemo)radiation (n = 123). Abduction, anteflexion, endorotation and exorotation were assessed. Subjective measurements were performed using the Visual Analogue Scale for pain, the Shoulder Disability Questionnaire (SDQ) and stiffness reporting. Results: Predictive factors for SDQ-score > 0 (n = 54) were VAS pain score, stiffness, abduction, anteflexion, physiotherapy, low shoulder position and surgical treatment. The SDQ, stiffness and pain scores were significantly higher in the surgically treated group than in the non-surgical group (p < 0.01). Anteflexion, abduction and exorotation were less impaired in the non-surgically treated group than in the surgically treated group (p < 0.01). No differences between neck dissection and neck dissection with post-operative radiotherapy, and radiotherapy and chemoradiation were found for these movements. Conclusions: Shoulder morbidity is often present after non-surgical treatment of the neck, but to a lesser extent compared to surgical treatment. Radiotherapy adds no morbidity to neck dissection and chemotherapy does not add extra morbidity to primary radiation. © 2008 Elsevier Ireland Ltd. All rights reserved.