Methodology
An honest explanation of how the assessment is structured and scored.
Structure of the assessment
The assessment consists of 26 questions across four reasoning domains: pattern recognition, numerical reasoning, logical reasoning and spatial reasoning. Questions are presented in a fixed order and increase in complexity as you progress, from straightforward warm-up items to considerably more demanding problems near the end.
Difficulty-weighted scoring
Not every question contributes equally to your result. Each item has an internal difficulty rating, and correct answers on harder questions contribute more to your overall score than correct answers on easier ones. Accuracy is the primary factor in your score — an incorrect answer never earns credit, regardless of how it was reached. Difficulty contributes additional weighting on top of that.
The role of response time
Response speed provides only a modest adjustment on top of accuracy and difficulty — it can never turn a wrong answer into a right one, and it can never let a faster respondent meaningfully outscore a slower, more accurate one. Answering correctly at a careful, deliberate pace earns close to full credit for that item. Implausibly fast responses (faster than a person could plausibly read and reason through the question) are denied any speed bonus, so rapid guessing cannot inflate a result.
From performance to an estimated IQ
Your difficulty-weighted, accuracy-driven performance across all 26 questions is converted into a single performance index. Scores are generated entirely server-side — the question bank, answer key and scoring logic never reach the browser. That index is then mapped onto an estimated IQ and an estimated percentile using a conventional IQ scale model (mean 100, standard deviation 15) and its normal distribution. This mapping is provisional — it has not yet been calibrated against a large, professionally representative normative sample.
What "provisional" means
The current scoring model is intentionally labelled and versioned as provisional. It is built to be replaced or recalibrated as we gather more response data, without needing to change the rest of the assessment experience. It is designed to give a directionally meaningful, honestly-labelled estimate of cognitive performance rather than a definitive, clinically validated score.
About this score
The Institute of Cognition assessment is an online reasoning assessment designed to estimate performance across several cognitive domains.
Results are estimates and should not be interpreted as a clinical, educational, or professionally administered intelligence diagnosis. Institute of Cognition is an independent online project — it is not a university, government institute, licensed medical provider, or academic research institute.