Objective This study aimed to compare the effects of music therapy with general recreational day activities in reducing agitation in people with dementia, residing in nursing home facilities. Methods In a randomised controlled design, residents with dementia (n = 94) were allocated to either music therapy or recreational activities. Both music therapy and general activities were offered twice weekly for 4 months. Changes in agitation were measured with a modified Cohen‐Mansfield Agitation Inventory (CMAI) at four intervals on each intervention day. A mixed model analysis was used to evaluate the effectiveness of music therapy, compared with general activities, on CMAI scores at 4 h after the intervention, controlled for CMAI scores at 1 h before the session and session number. Results Data were analysed for 77 residents (43 randomised to music therapy and 34 to general activities). In both groups, the intervention resulted in a decrease in agitated behaviours from 1 h before to 4 h after each session. This decrease was somewhat greater in the music therapy group than in the general activities group, but this difference was statistically not significant (F = 2.885, p = 0.090) and disappeared completely after adjustment for Global Deterioration Scale stage (F = 1.500; p = 0.222). Conclusions Both music therapy and recreational activities lead to a short‐term decrease in agitation, but there was no additional beneficial effect of music therapy over general activities. More research is required to provide insight in the effects of music therapy in reducing agitation in demented older people. Copyright © 2012 John Wiley & Sons, Ltd.