Reasoning Challenge
Methodology
Every form is manually authored, scoring stays on the server, and the claims stop where our evidence stops.
How forms are authored
All nine forms are written by hand against per-form blueprints — no bulk generation and no reuse of any live, leaked, or licensed test content. Quick forms carry 18 scored items; Full forms carry 54 scored items plus 30 reserves; Language-Light forms carry 36 scored items plus 24 reserves. Parallel forms within a family are authored to the same blueprint but share no scored items and no underlying item rules, so NRC-Q2 is a genuine second attempt rather than a reshuffle of NRC-Q1. Reserve items let us retire or rotate questions without shrinking a form.
What informed the design
Before building, we reviewed the Standards for Educational and Psychological Testing (AERA/APA/NCME), International Test Commission guidance, WCAG 2.2, and official Mensa material (accessed 2026-07-28). We used those sources for what they are good for — cautious claim limits, accessibility expectations, and content security — not to imply the challenge meets standardization requirements. It does not, and we say so.
What we deliberately do not claim
A real IQ score is a standardized score: it requires a representative norm sample, published reliability and validity evidence, and controlled administration. The Reasoning Challenge has none of those, so it reports a plain result — questions correct, total, and percentage on that one form. We do not convert it to a deviation quotient, assign percentiles, or compare you to a population. That is why the product name says “reasoning challenge” and the result is a Novus Reasoning Score, never an IQ. It cannot be used for Mensa qualification, diagnosis, giftedness identification, educational placement, employment selection, disability applications, accommodations, legal matters, or medical decisions.
How scoring works
Your browser receives only question stems and answer choices. Correct answers live in server-only modules that are never bundled for the client. Submissions are scored on the server against those keys, session deadlines are enforced server-side for timed forms, scoring endpoints are rate-limited, and score responses are sent with no-store cache headers. The scored result is then saved to your device only — see the privacy page for the full data model.
Corrections
If you find an ambiguous item, a defective key, or a claim on these pages that outruns the evidence, use the report links on assessment pages or see the assessment editorial policy for how corrections are handled.
Reasoning Challenge hub · Open NRC-Q1 · External testing options