Norming converts raw task performance into scores that can be compared within a reference group. The Japanese version of BrainTypeIQ uses 841 eligible first-time completers as its reference sample.
The 841-person reference sample was selected from more than 2,000 completion records
The Japanese version had accumulated more than 2,000 completed test records. These included repeat attempts and results produced with items or scoring conditions that differed from the current test.
The records were reviewed for response conditions, testing history, items, and scoring rules. The analysis then retained 841 people who had completed the current nine tasks for the first time under the same conditions. The raw-score distributions from these 841 completers became the reference for all nine tasks.
The same participants were used in the BrainTypeIQ Japanese-version reliability analysis.
In the Japanese version, BrainTypeIQ scores and percentiles show a test taker's relative standing among these first-time completers. The relationship between IQ 100, 120, and 130 and their percentiles provides a guide to reading that common scale.
Norming converts raw scores into comparable IQ scores
Twenty correct answers do not necessarily represent the same level of ability when tasks differ in question count or difficulty, or when the comparison group changes. For the Japanese version, BrainTypeIQ therefore maps the raw score for each of its nine tasks against the distribution from the 841-person reference sample and converts the results to a common scale.
How BrainTypeIQ scores are calculated explains how task scores become five domain scores, GAI, CPI, and overall IQ. IQ test measurement quality explains the distinct roles of norming, reliability, and validity.
Age trends were estimated primarily from 739 participants aged 18–54
The raw-score distributions use all 841 participants. The age model was developed primarily from the 739 participants aged 18 to 54.
| Domain | How age is handled |
|---|---|
| Gc (verbal comprehension) | No age adjustment |
| Gf, Gv, Gwm, and Gs | Age trends observed in the data are reflected in the task scores |
The treatment differs because cognitive abilities have different relationships with age. Intelligence and aging explains how the five abilities tend to change across adulthood.
Across 100 split-sample validations, the largest age-group mean difference was 3.77 points
The 841-person reference sample was divided into 70% training data and 30% validation data while preserving its age composition. Conversions built from the 70% training subset were applied to the unused 30% validation subset. This procedure was repeated 100 times to test whether the result depended on one particular split.
Across the four age groups, the largest remaining difference between mean overall-IQ scores was 3.77 points. The difference stayed below 4 points even in unused data, supporting the stability of the norms after accounting for age trends.
The English version uses provisional scoring while its own reference data continue to accumulate.