Cognitive Behavioral Therapy for Social Anxiety: What Trials Show

Cognitive behavioral therapy for social anxiety teaches people to test their predictions about being judged and to stop avoiding the situations they fear. The National Institute of Mental Health (NIMH) calls CBT the "gold standard" psychotherapy and says it is commonly used to treat social anxiety disorder^[6].

A 2014 network meta-analysis in The Lancet Psychiatry, covering 101 trials and 13,164 adults, concluded that individual CBT "should be regarded as the best intervention for the initial treatment of social anxiety disorder"^[1]. How common the disorder is, by age, state and population, is tracked on our social anxiety disorder pillar page; this guide covers what the therapy involves and what the trials found.

CBT is a family of techniques rather than one script, all aimed at the thoughts and behaviors that maintain anxiety over time^[7]. For social anxiety, the core ingredients are exposure to feared social situations and cognitive restructuring, and many programs add social skills practice^[7][6].

Most CBT protocols for social anxiety rest on a cognitive model in which inward-focused attention, negative beliefs about how one comes across and "safety behaviors" keep the fear from being disproven. Cognitive therapy based on the Clark and Wells model targets those processes directly, and in a 2022 trial changes in them accounted for around 60% of the improvement^[4]. That protocol had the largest effect of any psychotherapy in a 2025 network meta-analysis^[8].

Exposure and behavioral experiments

Exposure to feared social situations is the primary treatment approach for social anxiety disorder^[7]. NIMH describes exposure therapy as progressively confronting the fears underlying the disorder so a person can return to activities they have been avoiding^[6].

Exposures typically come in two stages^[7]. The first tests the overestimate that something bad will happen, for example by recording a 2-minute conversation and listening back to check whether the feared verbal slips occurred^[7]. The second, called social cost exposure, has a patient embarrass themselves on purpose, for example by singing in a crowded street, to learn whether the outcome is as intolerable as predicted^[7].

Cognitive restructuring

Cognitive restructuring means identifying anxious predictions and questioning them against the evidence. In social anxiety treatment it is used alongside exposure to reinforce the new learning that exposure produces^[7].

Self-focused attention, safety behaviors and video feedback

Face-to-face cognitive therapy uses an in-session experiment in which a patient does a social task twice: once focused on themselves and using safety behaviors, then focused outward with key safety behaviors dropped^[9]. Video feedback, in which patients watch recordings of themselves, is used extensively to correct negative self-perceptions^[9].

Behavioral experiments are used in most sessions, and attention training is another key feature of the treatment^[9].

Social skills practice

NIMH notes that CBT can help a person learn and practice social skills, which it calls very important for treating social anxiety disorder^[6]. It adds that CBT delivered in a group format can offer unique benefits for this condition^[6].

On the strongest available evidence, it does. In the 2014 Lancet Psychiatry network meta-analysis, individual CBT had the largest effect of any psychological treatment class compared with a waitlist (SMD 1.19), ahead of group CBT (0.92) and psychodynamic psychotherapy (0.62)^[1].

Placebo comparisons are the stricter test. Individual CBT outperformed psychological placebo (SMD 0.56), and SSRIs and SNRIs outperformed pill placebo (SMD 0.44); these were the only two classes that beat an appropriate placebo^[1]. Individual CBT also outperformed psychodynamic psychotherapy (SMD 0.56) and a combined class of interpersonal psychotherapy, mindfulness and supportive therapy (SMD 0.82)^[1].

A 2025 Bayesian network meta-analysis of 92 studies published through June 2024 reached the same ranking^[8]. CBT was generally the most effective psychotherapy, and the Clark and Wells protocol had the highest efficacy against a waitlist (SMD 1.42)^[8].

Individual trials turn those effect sizes into rates. In the 495-patient SOPHO-NET trial, 60% of patients responded to CBT and 36% reached remission, compared with 15% and 9% on a waiting list^[2]. CBT was significantly better than psychodynamic therapy for remission but not for response, and the between-group effect sizes were small^[2].

Standard individual cognitive therapy for social anxiety disorder involves up to 14 face-to-face sessions of 90 minutes, some of them held outside the office for behavioral experiments^[4]. In the 2022 Psychological Medicine trial, treatment lasted 14 weeks^[4].

An earlier trial put the average therapist time in standard cognitive therapy at about 19 hours per patient^[9]. Shorter versions using 55- to 60-minute sessions have also been evaluated in trials^[9].

Much of the work happens between sessions. In a development series of 11 patients using an internet version, participants completed an average of 17.9 behavioral experiments and spent about 35 hours on the program^[9].

Follow-up data are encouraging. Two years after the SOPHO-NET trial ended, response rates were about 70% for both CBT and psychodynamic therapy, remission rates were nearly 40%, and rates were stable or rising across the 24 months^[3].

In the 2022 cognitive therapy trial, 84% of face-to-face patients no longer met criteria for social anxiety disorder at the end of treatment and 87% at 12 months; for internet cognitive therapy the figures were 72% and 84%^[4]. Adolescent results in the SOPHO-NET youth trial were stable at 6- and 12-month follow-ups^[5].

CBT's short-term edge over psychodynamic therapy did not persist: the adult follow-up found no significant differences between the two after 6 months^[3].

NIMH describes treatment for social anxiety disorder as psychotherapy, medication or both, chosen according to a person's needs, preferences and medical situation^[6]. Medication options include SSRIs and SNRIs, beta-blockers and benzodiazepines^[6].

In the 2014 network meta-analysis, SSRIs and SNRIs had a large effect against a waitlist (SMD 0.91) and a smaller one against pill placebo (SMD 0.44)^[1]. The authors concluded that for people who decline psychological treatment, SSRIs show the most consistent evidence of benefit, and noted that other studies have found individual CBT carries a lower risk of side effects than medication^[1].

Beta-blockers are commonly the medication of choice for the performance-only type of social anxiety because they can control rapid heart rate, sweating and tremor^[6]. Benzodiazepines work quickly, but some people build up tolerance or become dependent, so providers may prescribe them only for brief periods^[6]. Drug classes and their trial results are covered in our guide to anxiety medication.

Among other talk therapies, psychodynamic therapy has the most support after CBT. It was the most effective non-CBT psychotherapy in the 2025 network meta-analysis (SMD 0.98 against a waitlist)^[8] and produced a 52% response rate in SOPHO-NET^[2]. NIMH also lists acceptance and commitment therapy (ACT), which uses mindfulness and goal setting, as an option with a growing evidence base^[6].

How CBT performs across other anxiety disorders is covered in therapy for anxiety, and the same method applied to low mood in cognitive behavioral therapy for depression.

Therapist-guided internet CBT now has head-to-head trial data. In a 2022 trial, 102 patients were randomized to internet cognitive therapy, face-to-face cognitive therapy or a waitlist for 14 weeks^[4]. The internet version did not differ significantly from face-to-face therapy on the primary outcome, and the authors could rule out its being more than 25% less effective^[4].

The difference was therapist time. Internet patients reached their full symptom change after 6.45 hours of therapist contact, while face-to-face patients needed 15.8 hours for the same drop, or 2.45 times more change per therapist hour for the internet format^[4]. A behavior test held in the clinic did show a modest advantage for face-to-face therapy (d = 0.47)^[4].

Across studies, the 2025 network meta-analysis found internet-based treatments comparable to face-to-face therapy and guided therapies more effective than unguided ones^[8]. In the 2014 review, self-help with support (SMD 0.86) and without support (SMD 0.75) both beat a waitlist but ranked below individual CBT^[1]. NIMH notes that psychotherapy can be effective in person or by telehealth^[6].

Social anxiety disorder usually starts in childhood or adolescence, and in young people it can appear as school avoidance, difficulty making friends or refusing to speak in social situations^[6]. Mental health data for this age group are on our adolescents page.

A 2024 network meta-analysis of 30 randomized trials with 1,547 children and adolescents found that internet-delivered CBT, group CBT and individual CBT all significantly reduced social anxiety symptoms^[10]. Internet CBT ranked highest for reducing social anxiety and depression symptoms, and group CBT ranked highest for improving functioning; the authors cautioned that the number of studies was limited^[10].

In the SOPHO-NET youth trial of 107 patients aged 14 to 20, 66% responded to CBT and 47% reached remission, compared with 20% and 6% on a waiting list^[5].

Most comparisons are against waitlists, which produce larger effects than placebo comparisons. Individual CBT's advantage fell from SMD 1.19 against a waitlist to 0.56 against psychological placebo^[1].

Remission is the harder bar. In adults, 36% of CBT patients were in remission at the end of treatment and nearly 40% two years later, so most still scored above the remission cutoff^[2][3].

Co-occurring conditions are common, including depression, other anxiety disorders and alcohol or substance misuse^[6]. In the 2022 trial, 72% of patients also met criteria for avoidant personality disorder at the start, and 82% of those reassessed a year later no longer did^[4].

Reaching treatment is its own gap. NIMH notes that people with social anxiety disorder are less likely to seek help, and that providers often identify the disorder while treating a co-existing condition^[6].