Drumming may have therapeutic and learning benefits but there exists little causal evidence regarding the benefits for children with emotional and behavioural difficulties (EBD) such as Autistic Spectrum Disorder. Six EBD pupils (EBD Drum) and 6 peers (Peer Drum) were given 2, 30 minute rock drumming lessons per week, over 5 weeks. Six matched individuals received no drumming instruction (3=EBD Control; 3=Peer Control). An exploratory, mixed-methods analysis was used to explore quantitative changes in skills and qualitative perspectives of the teaching staff. All pupils were tested two times (pretest and posttest) on drumming ability and Motor skills (Movement Assessment Battery for Children, version 2). Teacher’s rating of social behaviour (Strength and Difficulties Questionnaire; SDQ) was tested two times (pretest and retention).Significant differences in total SDQ difficulties between the four groups (χ2(3) = 8.210, p = 0.042) and the hyperactivity subscale (χ2(3) = 10.641, p = 0.014) were observed. The EBD Drum group had greater reductions in total difficulties compared to the Peer Drum (p = 0.009) group and specifically greater reductions in hyperactivity compared to Peer Drum (p = 0.046) and the EBD Control (p = 0.006) group. In follow-up interviews, staff spoke positively about changes in pupil’s attitudes towards learning and social confidence. The positive changes to social and behavioural skills reported in this pilot study are similar to those recorded for other music modalities.