Scientific transparency
What this assessment
can tell you.
LabNivo Intelligence is an online assessment of performance on five types of cognitive tasks. It is in Beta. Its reliability, validity and population norms have not yet been established.
Five domains, with clear boundaries
Matrix tasks ask you to combine visual relationships. Figural sequences involve progressions and transformations. Quantitative tasks involve elementary numerical reasoning. Spatial tasks involve mentally transforming arrangements. Memory tasks involve retaining or reorganising short sequences.
These tasks do not cover everything meant by intelligence. They do not measure creativity, emotional understanding, practical judgement or personal worth. This is not a clinical instrument and should not be used for diagnosis, admission, hiring or other consequential decisions.
What the Beta score means
A session contains nine scored tasks in each domain, plus one unscored experimental task per domain. Each scored task earns one point for an exact correct response. Skipped, interrupted, timed-out and incorrect responses earn zero. The overall score is the percentage of the 45 scored tasks answered correctly, rounded to the nearest whole number. Domain profiles show correct counts out of nine.
No speed bonus is applied. Timing is recorded to study task performance and identify potentially unusable sessions. A 30–40 minute duration is a planning estimate, not a norm derived from participants.
Why this is not an IQ score
A raw score has no population meaning on its own. An IQ-style estimate requires suitable reference groups, evidence that the intended interpretation is valid, and a measured account of uncertainty. We do not convert a percentage to an IQ, invent a percentile or present a guessed confidence interval.
How the tasks are selected
Each session uses a balanced form covering five designed difficulty tiers. Designed difficulty is an editorial judgement about complexity. It is not empirically measured difficulty. Selection is not an adaptive intelligence algorithm. Alternate forms have not yet been statistically equated, so their scores should not be treated as directly interchangeable.
The item bank contains original combinations and arrangements built from established cognitive task types. Items are not drawn from clinical test manuals or commercial admission tests. Internal rule checks can catch malformed tasks; independent human item review and participant piloting are still necessary to establish their quality.
Conditions matter
Fatigue, distractions, familiarity, vision and comfort with the instructions can influence performance. Repeating similar tasks may improve scores through practice. Recent repeated sessions and interrupted memory tasks are flagged for exclusion from norm development.
Memory sequences have fixed display intervals and a 60-second recall period. A refresh cannot restart a sequence. Browser backgrounding and timing instability can affect these tasks; this is one reason Beta results require caution.
Building trustworthy norms
With separate permission, response data can support item analysis and research. Before enabling normed estimates, the process must include independent psychometric review, a defined target population, representative recruitment, adequate coverage of intended age and language groups, and a justified sample-size plan.
Analyses must examine item difficulty, discrimination, missingness, reliability, dimensionality, measurement error, form comparability and fairness across relevant groups and devices. Validation on held-out data and independent evidence from appropriate external measures are needed. A large self-selected online sample alone does not establish representative norms.
There is no universal participant count that automatically makes an IQ test valid. The required sample depends on the model, precision, groups and intended claims. A qualified psychometrician must specify and approve the study plan and evidence before norms are activated.
Any future standard score will use a documented norm version. Confidence intervals must come from measured error estimates. Original results retain their scoring and norm versions when methods change.
Reference framework
Our development approach is informed by the Standards for Educational and Psychological Testing and the ITC/ATP Guidelines for Technology-Based Assessment. These organisations have not validated or endorsed LabNivo Intelligence.