IQ — Wechsler-style age-adjusted scoring
The IQ test runs 25 items across five reasoning categories: verbal, numerical, logical, spatial, and working memory. Each item is calibrated to a difficulty level (medium / high / expert) with a corresponding score weight.
Raw scores are converted to a deviation IQ on a mean-100, standard-deviation-15 scale — the same convention used by the Wechsler Adult Intelligence Scale (WAIS) since 1939. The conversion is age-adjusted across five bands so a 19-year-old’s 130 means the same thing as a 45-year-old’s 130: both sit roughly two standard deviations above the age-cohort mean.
We don’t claim our scores are interchangeable with a clinical Wechsler battery. We do claim they’re a defensible, reasonable estimate of general fluid reasoning in a 25-minute online format.
Personality — the Big Five (OCEAN)
The personality test is built on the Five-Factor Model — Openness, Conscientiousness, Extraversion, Agreeableness, and Emotional Sensitivity (the same trait often labelled Neuroticism elsewhere; we use the less stigmatising term).
Items are drawn from the International Personality Item Pool (IPIP), a public-domain instrument curated by Lewis Goldberg and colleagues since 1999. The short form (IPIP-50) takes about seven minutes; the deep form (IPIP-120) takes about fifteen and produces a more reliable profile. Both are scored using the original published reverse-coding patterns.
Optional career-interest items follow the Holland (RIASEC) typology— Realistic, Investigative, Artistic, Social, Enterprising, Conventional. The model has been the basis of vocational guidance research since 1959 and underlies the U.S. Department of Labor’s O*NET Interest Profiler.
Personality results are presented as a profile, not a diagnosis. We avoid type labels like introvert or extrovert in favour of percentile bands, and we explicitly frame the result as a snapshot rather than a verdict.
General Knowledge — domain-balanced sampling
The GK test draws 25 items spread across ten domains: Indian history and polity, world history, geography, general science, technology and computing, current affairs, Indian economy, arts and culture, sports, and general awareness. The spread is roughly proportional and adjusts to the candidate’s difficulty level.
Current-affairs items are tagged with their applicable year so stale content can be retired quarterly. We don’t pretend a static question pool stays accurate forever; we manage decay actively.
The no-repeat guarantee
Every question a candidate sees is recorded in a per-user history table. Subsequent attempts draw from the remaining pool. A candidate retaking the IQ test will not see a single item from any previous attempt until the unseen pool is exhausted — and the bank is large enough that exhausting it requires hundreds of attempts.
Generation and review
The IQ and GK question pools are produced through a hybrid pipeline. Generated items are validated against a strict schema, embedded with OpenAI’s text-embedding-3-small model, and rejected if their cosine similarity to any existing active item in the same category exceeds 0.92. Surviving items enter a pending-review queue; an editor approves or rejects each one before it goes live.
Personality items are not generated. They’re seeded once from the public-domain IPIP and RIASEC instruments and remain stable, since modifying validated psychometric items invalidates the framework.
Limits we’re honest about
- Online assessments cannot perfectly substitute for proctored, clinical evaluation.
- A single attempt is a snapshot; mood, sleep, and distraction all affect results.
- Cognitive scores are stable over time; what changes on retake is test-taking skill.
- Personality scales describe averages, not your behaviour in any specific situation.
- No assessment captures wisdom, creativity, ethics, or character.
If you want to dig deeper, the IPIP scales are documented at ipip.ori.org and the RIASEC framework is summarised in any standard career-counselling text.
Last updated: April 2026