​​​​Problem

​

• The Target Audience: Emotionally vulnerable youth (ages 19–33) who feel highly insecure or rejected, making them deeply susceptible to the residual influence of digital feeds.
   

• The Digital Environment: Anonymous, low-accountability discussion boards (like Reddit) where a lack of identity removes a user's perceived obligation to self-regulate.
   

• The Habitual Loop: Users experiencing information overload or strong emotions undergo impaired cognitive processing, causing them to passively absorb, normalise, and share unverified rumours without conscious intent.

​​

Analysis

​

The study was structured around the PATH Model (Problem, Analysis, Test, Help)—a classic framework in applied social psychology used to translate real-world issues into concrete behavioural designs. It provided a formal roadmap to define the problem (vulnerable users falling into echo chambers), analyse the underlying psychological causes, design the simulation to test the intervention, and outline practical implementation strategies for platform owners.

​

The behavioural architecture relies on three intersecting pillars of social psychology to explain why anonymity alters our behaviour online:

 

• Social Identity Model of Deindividuation Effects (SIDE): Explains how the absolute anonymity of platforms like Reddit causes personal identity to take a backseat, driving users to heavily conform to aggressive group norms and echo chambers.

• Objective Self-Awareness Theory: States that introducing subtle, human-centric design cues can break automated scrolling habits, instantly forcing a user to self-monitor their behaviour.

• Theory of Normative Social Behaviour (TNSB): Clarifies how heavily a user's behaviour is dictated by what they perceive their peers endorse or find socially acceptable.

 

Behavioural Process Analysis

Rather than looking at social media habits in isolation, a rigorous Process Model was built to analyse the entire feed ecosystem

 

• The Map: The analysis mapped 8 key predictor variables against 1 primary outcome: Misinformation Engagement Rate.

• Competing Forces: The framework isolated two opposing behavioural pathways—reinforcing forces (where anonymity and depersonalization actively suppress user self-monitoring) and undermining forces (where targeted human cues and reputation incentives disrupt the cycle and restore digital accountability).

 

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Insight

 

Why do people spread misinformation in anonymous spaces like Reddit?  I targeted three specific behavioural quirks

 

The Crowd Effect: Anonymity makes users drop their personal filters and blindly mirror toxic group behaviours.

The Pattern Interrupter: Adding small, humanised prompts completely breaks mindless "zombie-scrolling" and instantly forces self-awareness.

Peer Approval: Even in anonymous spaces, users care deeply about their reputation and what their community rewards.

​

Intervention

​

The project proposes ReadIt, a simulated Reddit-like environment that embeds two interventions into the platform architecture:

• Human cues that make actions feel more personal and accountable

• Social norms prompt that highlight how others typically behave and what is socially acceptable in that space.

Both elements introduce short, deliberate pauses so vulnerable users have a chance to think before reinforcing misleading content.

​​​​

Mechanics

​

When a user tries to post, repost, comment, share or upvote on a post pre-classified as misleading, the interface introduces a brief friction moment. A short, conversational prompt appears with blurred peer avatars, for example: “Would you say this to someone’s face?”

Alongside this, a Vibe Karma system tracks behaviour:

• Helpful, accurate contributions earn a Vibe Karma badge next to the username.

• Reinforcing misleading content slowly reduces the Vibe Karma score and can remove the badge, visible only to the user.

The goal is to make responsible behaviour socially meaningful while avoiding public shaming, which can be especially harmful for emotionally vulnerable users.

 

​

​

​

​

​

​

​

​

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Measurement

​​

​​Reinforcing actions include commenting, upvoting, saving, following or clicking “show me more like this” on misleading posts.

A secondary measure, Vibe Karma, tracks whether behaviour shifts over time by capturing cumulative helpful vs misleading engagement. Three key indicators are proposed: misinformation engagement rate, Vibe Karma rate and trust points accumulated.

​​

​

​

​

​

​Study Design​

​

Participants: emotionally vulnerable 19–33 year olds who primarily browse Reddit-like platforms at least four days a week. They are recruited through colleges, mental health clinics and NGOs, and screened using tools such as PHQ-A or SMFQ.

Setting: a 7-day simulated feed containing posts pre-classified as misleading by an expert panel, consistent with similar social media interventions.

A power analysis with an estimated effect size of d = .25, alpha = .05 and power = .80 suggests a total sample size of 506 participants, with 253 per condition. A 10 percent pilot (N = 50) is proposed to detect large effects before running the full trial.

 

To test if the intervention would actually make sense to an anxious user, I didn't just guess. I used an AI simulation to act as a digital user profile. When the AI was exposed to the normal feed, it blindly shared the post. But when it hit my custom face-to-face prompt, the AI chose to Self-Monitor and closed the post, proving the design's behavioural logic works in a test sandbox.

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

​

 

 

 

 

 

 

 

 

 

 

 

 

 

​

 

 

 

 

 

 

 

 

 

 

 

 

 

Implementation ​

​

• Platform owners
  Build friction-based prompts, human-feeling language and Vibe Karma into the platform architecture. A foreseeable barrier is the perceived opportunity cost of adding friction, but the platform can reframing this as an investment in credibility by positioning itself as an “anonymity first, responsibility forward” community.

• Government authorities
  Support through campaigns, regulatory guidance and framing, for example using narratives such as “Anonymity ≠ Apathy” in line with existing obligations to show harm reduction for minors. Barriers include public mistrust or perceptions that this curbs freedom of expression.

• End users
  Initial friction can feel frustrating, but that frustration signals disruption of automatic behaviour, consistent with habit rupture models. Over time, Vibe Karma converts that disruption into motivation and belonging by making responsible behaviour visible and personally meaningful, while keeping negative feedback private.

 

Projected Impact

​​

If Vibe Karma scores and trust indicators show incremental improvement over a six-month period, it suggests that self-monitoring has started to shift from prompt-driven to more habitual behaviour in this simulated environment. The long-term ambition is to extend this design to broader platforms once the ReadIt simulation shows meaningful reductions in misinformation engagement among emotionally vulnerable users.