Objective: In this randomized controlled trial, cognitive–behavioral group therapy (CBGT) for social anxiety disorder (SAD) was compared to group psychotherapy (GPT), a credible, structurally equivalent control condition that included only nonspecific factors of group treatment (such as group dynamics). Methods: Participants were 45 college students at the University of Colorado with a primary diagnosis of SAD. Each treatment condition comprised eight group sessions lasting 2 hr each. Independent assessors (blind to treatment assignment) assessed participants at baseline and posttreatment with the Clinical Global Impression Scale (CGI) and the Liebowitz Social Anxiety Scale (LSAS). Results: Both treatments were found to be equally credible. There were five noncompleters in the CBGT condition (21.7%) and only one in the GPT condition (4.3%). There were no statistically significant differences posttreatment (controlling for pretreatment scores) between the two treatment conditions, and both treatments were found to be efficacious. Effect sizes for CBGT were similar to earlier studies, and adherence ratings revealed excellent adherence. Conclusions: Treatment of SAD appears to be moving toward individual CBT, partly because of high attrition rates and underutilization of group dynamics in group CBT. However, group therapy has unique therapeutic ingredients, and it may be too early to give up on group treatment altogether. Discussion of these findings included future directions with this treatment modality, especially whether these two types of group treatment could be combined and whether such combination might serve to decrease attrition, enhance efficacy, and facilitate dissemination. Depression and Anxiety, 2011. © 2011 Wiley Periodicals, Inc.