As defined in the DSM-IV-TR (1), social anxiety disorder (SAD), or social phobia, is now well recognized as a prevalent and significantly impairing disorder with an onset early in life. Since Liebowitz and colleagues' commentary regarding the neglect of SAD in the contemporary literature (2), there has been considerable research into the nature of SAD, its natural history, its biological and environmental underpinnings, and its treatment. There has also been concern about the pathologizing of normal temperamental variations. This concern is most commonly linked to the use of pharmacologic treatments, particularly in younger patients, and has been exacerbated by the debate about the safety of antidepressant treatment of children and adolescents (3). Epidemiologic studies show that SAD is among the most prevalent of all mental disorders. It is frequently comorbid with other anxiety disorders, mood disorders, and substance use disorders (4). It also remains undiagnosed and untreated for many years after onset. Few psychiatric disorders involve a debate about whether the condition in question should be regarded as a disorder and, if so, whether people with the condition should receive treatment for their distress. In the case of SAD, the debate extends to the question of whether the therapeutic approach with the largest data base regarding efficacy-namely, psychopharmacology-should be employed. Longitudinal investigations of children with behavioural inhibition show that many retain these temperamental and behavioural features as they develop (5) and that they have characteristics of social anxiety and avoidance. Untreated SAD does not usually remit. The educational, interpersonal and vocational pathways of subjects with social anxiety are frequently adversely affected. Early recognition and early intervention are being increasingly advocated in other disorders and deserve consideration for the child and adolescent suffering from anxiety. It has been found that few children identified as meeting criteria for SAD receive any treatment (6). In a sample of 190 parents, only 31% of the children who suffered from current anxiety disorder had received any treatment, compared with 40% of those with depression and 79% of those with attention-deficit hyperactivity disorder. With the exception of specific phobia, the commonest anxiety disorder was social phobia, which had a 1-year prevalence rate of 3.2% (standard error 1.3%) (6). The approach of intervening early begs the question of where the threshold is set for defining a case. Wakefield and colleagues (7) offer a provocative reexamination of social anxiety and its disorders, founded in a critique of the current DSM criteria. In their view, the number of true cases of SAD is markedly lower than epidemiologic studies based on DSM or ICD criteria recognize in community samples. These authors raise the concept of nondisordered social fears that cause suffering as a condition separate from what psychiatrists view as SAD. Should we reexamine the criteria for caseness? Will DSM-V take on that task? It has always appeared unlikely that marked increases in case recognition over brief periods of time reflect significant shifts in the actual prevalence of similarly defined cases. Thus SAD's current high prevalence (often quoted as 13.5%), compared with the fact that it was virtually unrecognized 20 years ago, reflects changes in awareness, in the criteria used for diagnosis, in the methods of screening for cases, and in the thresholds used to define caseness. If we cannot be reasonably sure of how to define cases, do we have sufficient evidence to recommend treatment? Our understanding of the efficacy of interventions in SAD is based on 2 streams of investigation: the common dichotomy of psychosocial and pharmacologic treatments, together with the combination of these treatments in willing subjects diagnosed with SAD according to DSM-III or DSM-IV criteria. …