Sturmey (2005)argues that the evidence base underlying approaches to intervention based on applied behavioural analysis (ABA) are significantly stronger than that underlying approaches to intervention based on cognitive therapy. He concludes that ‘the ethical imperative of beneficence requires that people, including people with ID, receive known effective treatments. Those effective treatments are based on ABA’ (p. X). In this commentary, I argue that his selection of evidence to support the central argument (the superiority of ABA) involves some highly contestable assumptions and that evidence of the effectiveness of ABA falls far short of that required for evidence-based policy and practice.