Only a minority of patients with social anxiety disorder (SAD) has a robust therapeutic response to evidence-based serotonin reuptake inhibitor (SSRI) treatment. To help improve the personalized medicine approach to psychiatric care, we evaluated several candidate genetic predictors of SSRI response in SAD. At the start of a randomized controlled trial (NCT00282828), 346 patients with SAD at three sites received protocol-driven, open-label treatment with sertraline, up to 200. mg/d over 10 weeks. Efficacy was determined using a continuous measure of outcome (Liebowitz Social Anxiety Scale (LSAS)) and dichotomous indicators of response (LSAS ≤ 50) and remission (LSAS ≤ 30). Predictors of efficacy were examined in multivariate regression models that included eight polymorphic variants in four candidate genes (four in RGS2, two in HTR2A, one in SLC6A2, and one in SLC6A4). Adjusting for genetic ancestral cluster and non-genetic predictors of response, all four single-nucleotide polymorphisms (SNPs) in RGS2 predicted change in LSAS over time, at study-wise significance (p=0.00833), with the minor allele associated with less improvement over time. After adjusting for genetic ancestral cluster and non-genetic predictors of remission, two of the four RGS2 SNPs predicted likelihood of remission at or just below study-wise significance (p=0.025): rs4606 (AOR=0.49 (95% CI=0.27-0.90), p=0.022) and rs1819741 (AOR=0.50 (95% CI=0.28-0.92), p=0.027). Variation in RGS2, a gene previously shown to be associated with social anxiety phenotypes and serotonergic neurotransmission, may be a biomarker of the likelihood of substantially benefiting from sertraline among patients with SAD.