How Social Media Reshapes Social Anxiety in Young People: New Oxford Research (2026)

SocialAnxiety.co Research Summary | Based on OxCADAT research, Department of Experimental Psychology, University of Oxford

Research Attribution: This article summarizes findings from Skjerdingstad, N. & Leigh, E. (2026), “Youth social anxiety in the digital age: Reconceptualising cognitive-behavioural processes,” published in Nature Mental Health. The original research was conducted at the Oxford Centre for Anxiety Disorders and Trauma (OxCADAT) and the Cognitive and Behavioural Approaches to Mental Health in Young People (CAMY) research group, University of Oxford. SocialAnxiety.co is an independent psychoeducation platform and is not affiliated with the University of Oxford.

Why This Research Matters

The Clark and Wells (1995) cognitive model of social anxiety was developed before smartphones existed. Before Instagram. Before TikTok. Before the average teenager spent several hours daily navigating social evaluation through a screen.

That model — which remains the theoretical foundation of NICE-recommended cognitive therapy for social anxiety — describes how social fear is maintained through three interlocking processes: self-focused attention, safety behaviors, and post-event processing. It was designed to explain what happens when a person walks into a room, gives a presentation, or has a conversation.

But what happens when the “room” is a comments section? When the “presentation” is a posted photo? When the “conversation” is a group chat where your message can be screenshotted, shared, and evaluated by an audience you cannot see?

This is the question Oxford researchers Nora Skjerdingstad and Dr. Eleanor Leigh address in their 2026 Nature Mental Health paper. Their work reconceptualizes the cognitive-behavioral processes that maintain social anxiety disorder in the context of how young people actually live now — in digital environments that fundamentally alter the nature of social evaluation.

The Core Argument: Digital Environments Change the Maintenance Mechanisms

The original Clark and Wells model describes social anxiety as a cycle. A socially anxious person enters a social situation. Their attention turns inward — monitoring their own performance, scanning for signs of failure. They use safety behaviors to manage perceived threat. Afterward, they ruminate — replaying the interaction, searching for evidence of negative evaluation.

Skjerdingstad and Leigh’s contribution is not to replace this model but to demonstrate how each of these maintenance mechanisms operates differently in digital social environments — and in many cases, more intensely.

Self-Focused Attention Becomes Literal

In face-to-face interaction, self-focused attention is an internal cognitive process. The person imagines how they appear to others — the “observer perspective” described by Clark and Wells.

On social media, this imagined observer perspective becomes a literal one. Young people see themselves on camera during video calls. They review their own photos before posting. They monitor their own profiles as others would view them. The gap between the internal self-perception and the external presentation collapses — the person is simultaneously the performer and the audience.

For a young person with social anxiety, this means the self-monitoring that was previously intermittent (occurring during social encounters) becomes continuous. The digital self is always visible, always evaluable.

Safety Behaviors Evolve into Digital Strategies

In the original model, safety behaviors include avoiding eye contact, rehearsing statements, holding objects to hide trembling. In digital environments, the equivalent safety behaviors are technologically amplified:

Excessive editing of photos, messages, and posts before sharing — the digital equivalent of over-preparation
Delayed responding in group chats — waiting to see what others say before contributing, avoiding the risk of an unpopular first response
Lurking without participating — the digital equivalent of attending a party but standing in the corner. The person is present in the group but invisible, which prevents corrective social learning
Curating a “safe” online persona — presenting only a carefully filtered version of the self, which prevents authentic social connection and reinforces the belief that the real self would be rejected

These safety behaviors share the same clinical problem as their offline counterparts: they prevent the young person from learning that authentic social participation does not lead to catastrophe. But they are harder to identify — and harder for therapists to target — because they are embedded in normal digital behavior.

Post-Event Processing Becomes Permanent

In face-to-face interaction, post-event processing is reconstructive. The socially anxious person remembers the interaction — imperfectly, biased toward negative details — and ruminates on what went wrong.

In digital environments, the “interaction” is permanently recorded. A young person can re-read their own messages. They can check how many likes a photo received. They can see who viewed their story and who did not respond. They can screenshot conversations and analyze them hours later.

Post-event processing is no longer limited by memory — it is fueled by an infinite archive of social data. The rumination cycle does not decay over time. The evidence is always available for re-examination.

The Audience Becomes Invisible and Infinite

Perhaps the most significant digital amplification is to the nature of the social audience itself. In face-to-face social anxiety, the feared audience is visible and bounded — the people in the room, the colleagues at the meeting, the strangers at the party.

Online, the audience is invisible and potentially infinite. A comment, a photo, a video can be seen by unknown numbers of people. The socially anxious young person cannot assess the audience’s reaction in real time — there are no facial expressions, no nods, no reassuring smiles. The absence of feedback is filled by the anxiety’s default assumption: negative evaluation.

Clinical Implications: What This Means for Treatment

Skjerdingstad and Leigh’s reconceptualization has direct implications for how cognitive therapy for social anxiety should be adapted for young people in 2026.

Assessment must include digital behavior. Clinicians asking about avoidance patterns need to assess not only whether a young person avoids parties or presentations, but whether they avoid posting, participating in group chats, or turning on their camera during video calls. Digital avoidance is functionally identical to in-person avoidance — it prevents corrective learning.

Safety behavior identification must extend to online strategies. Excessive message editing, lurking, and persona curation are safety behaviors. They need to be named, tracked, and gradually eliminated in the same way traditional safety behaviors are addressed in CBT.

Exposure hierarchies should include digital challenges. Posting a photo without excessive editing. Sending a message in a group chat without waiting for others to go first. Turning on the camera during a video call. These are the modern equivalents of raising your hand in class or speaking to a stranger — and they need to be incorporated into graduated exposure protocols.

Post-event processing interventions must account for digital evidence. Teaching a young person to stop ruminating is harder when they can re-read the exact conversation. Therapeutic strategies may need to include deliberate “digital detachment” periods after social interactions — reducing access to the archive that fuels rumination.

About the Researchers

Dr. Eleanor Leigh is Associate Professor of Clinical Psychology at the University of Oxford and Deputy Director of the Oxford Centre for Emerging Minds Research. Her work focuses on understanding and treating social anxiety disorder in adolescents, applying and extending the Clark and Wells cognitive model to younger populations. She is a member of OxCADAT (Oxford Centre for Anxiety Disorders and Trauma).

Nora Skjerdingstad is a DPhil student at the Department of Experimental Psychology, University of Oxford, supervised by Dr. Leigh and Professor Cathy Creswell. Her research examines how digital environments and social media shape the maintenance of social anxiety in young people, using ecological momentary assessment and experimental methods.