The Clark and Wells cognitive model for social anxiety disorder is the most effective psychological treatment for SAD in the published research, with effect sizes around d = 1.5 and superiority over both group CBT and selective serotonin reuptake inhibitors in head-to-head trials. The 14-16 session protocol differs meaningfully from general CBT for anxiety — it places almost no emphasis on relaxation or generic exposure and substantial emphasis on attention training, video feedback, and behavioral experiments that drop safety behaviors. This guide covers the model, the protocol, and the implementation details that distinguish effective from ineffective delivery.
The Clark and Wells model: why social anxiety persists
The Clark and Wells model identifies three maintaining mechanisms that together explain why social anxiety persists despite frequent social interaction (which one might expect to reduce it).
Self-focused attention. In a feared social situation the client's attention shifts inward to their own felt sense of how they appear — shaky voice, blushing, sweating, mind going blank. This internal data is then used as evidence for how they actually appear to others, which is typically much worse than the external reality.
Safety behaviors. Subtle behaviors deployed to prevent the feared outcome — gripping a glass to hide a trembling hand, rehearsing sentences before speaking, avoiding eye contact, agreeing with everything to avoid disagreement. These prevent new learning by ensuring the client never tests whether the feared outcome would have occurred without the safety behavior.
Post-event processing. After a social situation, the client mentally rehearses what went wrong, what people must have thought, evidence of failure. This consolidates the negative self-image and primes anticipatory anxiety for the next situation.
The 14-16 session protocol
Sessions 1-2: Assessment and case formulation. Administer SPIN and LSAS. Build idiosyncratic version of the Clark and Wells model with this client's specific safety behaviors, self-focused attention patterns, and feared outcomes.
Sessions 3-4: Attention training. Teach external focus through in-session practice. Client deliberately shifts attention from internal sensations to specific external cues (the actual content of the other person's words, three details about their appearance, the colors in the room).
Sessions 5-7: Video feedback. The hallmark intervention. Record the client during a feared social situation (in-session role play first, then real-world). Have them predict in detail how they will appear. Watch the recording together. The discrepancy between catastrophic prediction and actual appearance produces immediate cognitive change.
Sessions 8-11: Behavioral experiments dropping safety behaviors. Client deliberately does the feared behavior (asking a question, expressing an unpopular opinion, sustaining eye contact, allowing pauses) without safety behaviors. Compare predicted vs actual outcome.
Sessions 12-14: Schema-level work. Address core beliefs that emerged through the earlier work — typically around being judged, being defective, being unlikeable.
Sessions 15-16: Relapse prevention.
Video feedback: the protocol-defining intervention
Video feedback is unique to the Clark and Wells protocol and is the most powerful single intervention for SAD. Implementation details that matter:
Predict in detail before watching. The client must commit to specific predictions: "My face will be visibly red." "My voice will tremble noticeably." "I will look unconfident." Without specific predictions, the post-watching cognitive change is muted.
Watch as a stranger would. Instruct the client to watch the recording as though they had never met this person, were watching with the sound off, were watching from across the room. This breaks the fused self-as-seen-from-inside perspective.
Do not minimize visible anxiety. If the client did appear visibly anxious, do not pretend they did not. The goal is calibration with reality, not reassurance. Most of the time the client looks far less anxious than they predicted; on the rare occasion they look as anxious as predicted, the work shifts to "and what does that mean?"
Video feedback typically produces a 30-50 point drop in the client's estimate of how visibly anxious they appeared, immediately.
Why generic exposure therapy under-performs for social anxiety
Many clinicians attempt to treat social anxiety with generic exposure therapy (graded exposure to social situations) and find it produces modest gains. The Clark and Wells model explains why: graded exposure without the three core interventions does not address the maintaining mechanisms.
The client doing generic graded exposure typically: (1) maintains self-focused attention throughout the exposure, (2) deploys safety behaviors that prevent new learning, (3) engages in post-event processing that consolidates the negative self-image rather than new evidence.
The result is repeated exposure with limited cognitive change. The Clark and Wells protocol works because it specifically targets each of these mechanisms before and during exposures.
Clinically, this means: if you are delivering exposure therapy for social anxiety and seeing slow progress, the bottleneck is almost always one of the three maintaining mechanisms not being addressed, not the dose of exposure.
Tracking SAD-specific outcomes
Track SPIN (Social Phobia Inventory) or LSAS (Liebowitz Social Anxiety Scale) every 4 sessions. The Self-Focused Attention Scale and Safety Behaviors Questionnaire are useful for tracking the model-specific mechanisms. Practice management platforms with built-in SAD workflow tooling (CBT Assistant Pro includes the Clark and Wells session structure template and SPIN/LSAS tracking) make this straightforward to maintain across the 14-16 session protocol without manual data entry.
Frequently asked questions
What is the success rate of CBT for social anxiety disorder?
The Clark and Wells protocol produces large effect sizes around d = 1.5 and superior outcomes to medication in head-to-head trials. Approximately 70-80% of completers achieve clinically significant change on SPIN or LSAS.
Does CBT for social anxiety include exposure therapy?
Yes, but the exposures are structured as behavioral experiments dropping safety behaviors with attention training, not as generic graded exposure. The cognitive framing of the exposure is what produces the change.
How is the Clark and Wells protocol different from CBT for general anxiety?
It is highly specific. It includes video feedback (unique to SAD), attention training for self-focused attention, and a behavioral experiment structure that specifically targets safety behaviors. Generic anxiety CBT does not include these and produces smaller effects for SAD presentations.
Can CBT for social anxiety be delivered in group format?
Group CBT for SAD shows efficacy but produces smaller effects than individual Clark and Wells protocol. For severe presentations, individual treatment is preferred. Group can be a cost-effective option for mild to moderate cases.
Is medication needed alongside CBT for social anxiety?
In Clark's head-to-head trials, cognitive therapy alone outperformed both medication alone and combined treatment. Medication is appropriate when CBT is unavailable or has not produced sufficient response, but is not required as first-line treatment.
Ready to speed up your CBT documentation?
CBT Assistant Pro helps therapists build formulations 3× faster with AI-assisted documentation. HIPAA compliant. Free trial, no credit card.
Start Free Trial →