Although cognitive behavior therapy (CBT) has been shown to be generally effective in the treatment of social anxiety disorder (SAD), not all individuals respond to treatment, and among those who do respond the degree of improvement is sometimes far from optimal. Little research has examined the impact of variations in the format of treatment delivery in this area. Participants were randomly assigned to either a standard, 12-session CBT program for generalized SAD in which treatment was delivered in successive weekly sessions (standard treatment) or a similar program in which the 12 sessions were delivered over 18 weeks (extended treatment). Intent-to-treat analyses revealed that the standard treatment program resulted in superior outcome in terms of self-rated symptom and impairment levels, categorical ratings of responder status, and lower dropout rates. Analyses of treatment completers only revealed comparable gains between the two conditions by post-treatment. However, the standard treatment condition revealed a more rapid improvement in magnitude initially. These findings suggest no benefit from extending the course of CBT treatment over a greater length of time, and suggest that such extension may in fact substantially increase the likelihood of premature termination.