Cognitive behavior group therapy (CBGT) is effective for social anxiety disorder (SAD), but a substantial proportion of patients do not typically achieve normative functioning. Cognitive behavioral models of SAD emphasize negative self-imagery as an important maintaining factor, and evidence suggests that imagery is a powerful cognitive mode for facilitating affective change. This study will compare two group CBGT interventions, one that predominantly uses verbally-based strategies (VB-CBGT) and another that predominantly uses imagery-enhanced strategies (IE-CBGT), in terms of (a) efficacy, (b) mechanisms of change, and (c) cost-effectiveness. This study is a parallel groups (two-arm) single-blind randomized controlled trial. A minimum of 96 patients with SAD will be recruited within a public outpatient community mental health clinic in Perth, Australia. The primary outcomes will be self-reported symptom severity, caseness (SAD present: yes/no) based on a structured diagnostic interview, and clinician-rated severity and life impact. Secondary outcomes and mechanism measures include blind observer-rated use of safety behaviors, physiological activity (heart rate variability and skin conductance level) during a standardized speech task, negative self-beliefs, imagery suppression, fear of negative and positive evaluation, repetitive negative thinking, anxiety, depression, self-consciousness, use of safety behaviors, and the EQ-5D-5L and TiC-P for the health economic analysis. Homework completion, group cohesion, and working alliance will also be monitored. The outcomes of this trial will inform clinicians as to whether integrating imagery-based strategies in cognitive behavior therapy for SAD is likely to improve outcomes. Common and distinct mechanisms of change might be identified, along with relative cost-effectiveness of each intervention.