Aggressive challenging behaviour is frequently reported in adults with intellectual disability and it is often treated with antipsychotic drugs. However, no adequate evidence base for this practice exists. We compared flexible doses of haloperidol (a typical, first-generation antipsychotic drug), risperidone (an atypical, second-generation antipsychotic), and placebo, in the treatment of this behaviour. 86 non-psychotic patients presenting with aggressive challenging behaviour from ten centres in England and Wales, and one in Queensland, Australia, were randomly assigned to haloperidol (n=28), risperidone (n=29), or placebo (n=29). Clinical assessments of aggression, aberrant behaviour, quality of life, adverse drug effects, and carer uplift (positive feelings about the care of the disabled person) and burden, together with total costs, were recorded at 4, 12, and 26 weeks. The primary outcome was change in aggression after 4 weeks' treatment, which was recorded with the modified overt aggression scale (MOAS). Analysis was by intention to treat. This study is registered as ISRCTN 11736448. 80 patients had adherence of 80% or more to prescribed drug. Aggression decreased substantially with all three treatments by 4 weeks, with the placebo group showing the greatest change (median decrease in MOAS score after 4 weeks=9 [95% CI 5-14] for placebo, 79% from baseline; 7 [4-14] for risperidone, 58% from baseline; 6.5 [5-14] for haloperidol, 65% from baseline; p=0.06). Furthermore, although no important differences between the treatments were recorded, including adverse effects, patients given placebo showed no evidence at any time points of worse response than did patients assigned to either of the antipsychotic drugs. Antipsychotic drugs should no longer be regarded as an acceptable routine treatment for aggressive challenging behaviour in people with intellectual disability.