# Comprehensive Specification: Are you okey? (clickbait.club) # Full Context for Large Language Models & Autonomous Agents > Application: Are you okey? (Brief Implicit Association Test & Cognitive Friction Lab) > Canonical URL: https://clickbait.club/ > Edge API: https://api.clickbait.club/ > Specification Version: 2.1.0 > Last Updated: 2026-09-19 --- ## 1. Executive Abstract "Are you okey?" (clickbait.club) is an empirical psychometric laboratory disguised as a fast-paced cognitive reflex test. The system measures implicit cognitive bias by calculating millisecond response latency differences across counterbalanced pairing blocks using the Brief Implicit Association Test (BIAT) paradigm. Specifically, it quantifies the cognitive friction present when an individual is asked to rapidly associate self-concept markers with dogmatic demands (irrational beliefs in Albert Ellis's Rational Emotive Behavior Therapy framework) compared to flexible preferences (rational beliefs). --- ## 2. Experimental Paradigm: Brief Implicit Association Test (BIAT) The test adapts the Brief Implicit Association Test methodology established by Sriram & Greenwald (2009) and analyzed for web environments by Carpenter et al. (2020). ### 2.1 Dimensions & Stimuli Pool The experiment features two core dimensions: 1. **Target Concept Dimension (Self vs. Other)** - Self-referent: `Me`, `I`, `My`, `Self` - Other-referent: `They`, `Them`, `Their`, `Other` 2. **Attribute Dimension (Demandingness vs. Preference)** - Limitative / Dogmatic Demands: `Must`, `Should`, `Impossible`, `Can't`, `Terrible`, `Failure` - Non-Limitative / Flexible Preferences: `Prefer`, `Wish`, `Possible`, `Able`, `Accept`, `Setback` ### 2.2 Stimuli Dictionary | Stimulus ID | Word | Category | Dimension | |---|---|---|---| | `s1` | Me | Self | Target | | `s2` | I | Self | Target | | `s3` | My | Self | Target | | `s4` | Self | Self | Target | | `o1` | They | Other | Target | | `o2` | Them | Other | Target | | `o3` | Their | Other | Target | | `o4` | Other | Other | Target | | `l1` | Must | Limitative (Dogmatic) | Attribute | | `l2` | Should | Limitative (Dogmatic) | Attribute | | `l3` | Impossible | Limitative (Dogmatic) | Attribute | | `l4` | Can't | Limitative (Dogmatic) | Attribute | | `l5` | Terrible | Limitative (Dogmatic) | Attribute | | `l6` | Failure | Limitative (Dogmatic) | Attribute | | `n1` | Prefer | Flexible (Rational) | Attribute | | `n2` | Wish | Flexible (Rational) | Attribute | | `n3` | Possible | Flexible (Rational) | Attribute | | `n4` | Able | Flexible (Rational) | Attribute | | `n5` | Accept | Flexible (Rational) | Attribute | | `n6` | Setback | Flexible (Rational) | Attribute | ### 2.3 Dual-Block Counterbalanced Structure Each session consists of two distinct pairing blocks: - **Incompatible Block (Demanding Self)**: - Target key (Left / `E`): `Self` OR `Limitative` - Contrast key (Right / `I`): `Other` OR `Flexible` - **Compatible Block (Flexible Self)**: - Target key (Left / `E`): `Self` OR `Flexible` - Contrast key (Right / `I`): `Other` OR `Limitative` To eliminate procedural order bias, the presentation sequence (Incompatible first vs Compatible first) is randomized 50/50 across participant sessions. Within each block, target items and attribute items alternate systematically. --- ## 3. Mathematical Scoring Algorithm: Greenwald Improved D-Score (D₆₀₀) The raw reaction times collected in the browser are sent to the Cloudflare Edge Worker API (`https://api.clickbait.club/score`) and processed via the standardized Greenwald et al. (2003) algorithm: ### 3.1 Data Cleaning & Data Quality Filters 1. **Extreme Distraction Filter**: Discard any trial with reaction time $RT > 10,000\text{ ms}$. 2. **Button-Masher / Fast-Responder Invalidation**: If more than 10% of a participant's trials have $RT < 300\text{ ms}$, the session is discarded for invalid engagement. ### 3.2 Error Latency Adjustment ($D_{600}$) Errors are not penalized by halting the test or presenting disruptive visual cues. Instead, the task advances immediately to preserve participant flow, and the backend replaces each error latency with the mean of correct trials in that respective block plus a 600 ms penalty: $$RT_{\text{error, adjusted}} = \bar{RT}_{\text{correct, block}} + 600\text{ ms}$$ ### 3.3 Effect Size Calculation 1. Compute the adjusted mean latency for the Incompatible block ($\bar{M}_{\text{incompatible}}$) and Compatible block ($\bar{M}_{\text{compatible}}$). 2. Calculate the difference in mean latency: $$\Delta RT = \bar{M}_{\text{incompatible}} - \bar{M}_{\text{compatible}}$$ 3. Pool all valid, adjusted trials across both blocks and calculate the inclusive standard deviation ($SD_{\text{pooled}}$). 4. Compute the standardized $D$-score: $$D = \frac{\bar{M}_{\text{incompatible}} - \bar{M}_{\text{compatible}}}{SD_{\text{pooled}}}$$ ### 3.4 Theoretical Interpretation of D-Score - **$D > +0.35$ (High / Severe Friction)**: Substantial cognitive resistance when forced to map self-concept to dogmatic demands. The participant naturally processes self-referent concepts alongside flexible preferences far more rapidly. Reflects strong demandingness resistance and rational cognitive flexibility. - **$-0.15 \le D \le +0.15$ (Minimal Friction / Neutral)**: Balanced associative speed across both pairings. - **$D < -0.35$ (Internalized Demanding Patterns)**: Participant is significantly faster at categorizing self-concept with dogmatic demands ("Must", "Should", "Failure") than with flexible preferences. Indicates automated cognitive internalization of rigid evaluative standards. --- ## 4. Albert Ellis's Rational Emotive Behavior Therapy (REBT) Construct Mapping The test's attribute taxonomy maps directly to the cognitive triad of irrational beliefs identified in REBT: 1. **Demandingness ("Musturbation")**: Transforming preferences into absolutist demands (*"I must succeed; I should never show weakness"*). 2. **Awfulizing / Catastrophizing**: Evaluating setbacks as catastrophic (*"Terrible", "Impossible"*). 3. **Low Frustration Tolerance (LFT)**: Believing discomfort is intolerable (*"Can't"*). 4. **Global Self-Downing**: Equating a specific error with totality of self-worth (*"Failure"*). --- ## 5. TypeSafe Jev AI Evaluation Pipeline Rather than passing raw numbers into unconstrained generative LLMs that could drift or hallucinate diagnostic claims, `api.clickbait.club` leverages TypeSafe AI's deterministic Jev system: - Endpoint: Jev SystemOne Deterministic Gateway - Schema: Typed `Score` primitive - Input: Calculated $D$-score, $\Delta RT$, error rate, sample size - Mapping: Deterministic classification into the 5-point REBT Cognitive Friction Scale: 1. Internalized Limiting Patterns ($D < -0.35$) 2. Low Friction ($-0.35 \le D < -0.15$) 3. Minimal Friction ($-0.15 \le D \le 0.15$) 4. Moderate Friction ($0.15 < D \le 0.35$) 5. High Friction ($0.35 < D \le 0.65$) 6. Severe Friction ($D > 0.65$) --- ## 6. Zero-Knowledge Privacy Architecture The platform operates on a strict zero-knowledge telemetry standard: - **Zero Personal Data**: No names, emails, user IDs, or account identifiers. - **Zero Cookies / Local Storage Tracking**: No tracking cookies, no session identifiers. - **Zero Device Fingerprints**: Canvas, WebGL, audio, and user-agent fingerprinting are prohibited. - **Zero IP Retention**: IP addresses are stripped at Cloudflare Edge before processing. - **Anonymous Percentile Storage**: Cloudflare D1 SQL database stores only: - `d_score` (REAL) - `latency_delta_ms` (REAL) - `error_rate` (REAL) - `created_at` (INTEGER timestamp) These rows are used exclusively to calculate live empirical percentile distributions. --- ## 7. AI Handoff & Clinical Prompt Schema Upon completing the test, participants can click dedicated launch buttons for OpenAI ChatGPT, Anthropic Claude, or Google Gemini. The application packages their empirical session metrics into a structured directive: ```text You are an expert cognitive psychologist and Rational Emotive Behavior Therapy (REBT) practitioner. Analyze my empirical Brief Implicit Association Test (BIAT) results from clickbait.club: - Greenwald D-Score: [D_SCORE] - Diagnostic Classification: [DIAGNOSTIC_LABEL] - Latency Differential (ΔRT): [DELTA_MS] ms - Sorting Accuracy: [ACCURACY]% - Population Percentile: [PERCENTILE]th percentile Methodological Context: The test measures response latency asymmetry between self-concept ('Me', 'I', 'My') and dogmatic demands ('Must', 'Should', 'Failure') vs flexible preferences ('Prefer', 'Wish', 'Possible'). Scores are computed using Greenwald's D600 algorithm with pooled standard deviation. Please provide: 1. An objective interpretation of my D-score and latency differential in plain language. 2. How this reflects my automated self-concept schema according to Albert Ellis's REBT model (demandingness vs preference). 3. Practical cognitive reframing exercises to strengthen flexible preference schemas. ``` --- ## 8. Academic References 1. Greenwald, A. G., Nosek, B. A., & Banaji, M. R. (2003). Understanding and using the Implicit Association Test: I. An improved scoring algorithm. *Journal of Personality and Social Psychology*, 85(2), 197–216. https://doi.org/10.1037/0022-3514.85.2.197 2. Sriram, N., & Greenwald, A. G. (2009). The Brief Implicit Association Test. *Experimental Psychology*, 56(4), 283–294. https://doi.org/10.1027/1618-3169.56.4.283 3. Carpenter, T. P., Pogacar, R., Pullig, C., Kouril, M., Aguilar, S., LaBouff, J., ... & Chakroff, A. (2020). Survey-software Implicit Association Tests: A methodological and empirical analysis. *Behavior Research Methods*, 52(5), 2194–2208. https://doi.org/10.3758/s13428-019-01293-3 4. Ellis, A. (1958). Rational psychotherapy. *The Journal of General Psychology*, 59(1), 35–49. 5. Ellis, A. (1962). *Reason and Emotion in Psychotherapy*. Lyle Stuart.