Test Methodology

Non-verbal format. All questions use abstract visual patterns — matrix reasoning, spatial analogies, and sequence completion. This eliminates language bias and makes the test appropriate across linguistic backgrounds.

Culture-fair design. Items avoid culturally specific knowledge, relying solely on figural reasoning. This approach is consistent with Cattell's Culture Fair Intelligence Test framework and the Raven's Progressive Matrices tradition.

Five cognitive dimensions. 4 questions per dimension (20 total) across pattern recognition, visual logic, processing speed, sequential reasoning, and mental flexibility — representing the major factors of fluid intelligence (Gf) as defined in the Cattell-Horn-Carroll (CHC) model. This 4-items-per-factor design targets an estimated internal consistency (Cronbach's α) of ~0.78, meeting the APA/NCME 2014 threshold of ≥ 0.70 for screening instruments.

Calibrated difficulty. Items span three difficulty levels (foundational, intermediate, advanced) per dimension. Scoring weights reflect item difficulty, with harder items contributing proportionally more to the final score — consistent with item response theory principles.

Raw scores are normalized against a reference distribution approximating a standard IQ scale (mean = 100, SD = 15). The resulting estimate reflects where the test-taker's performance falls relative to this distribution.

Percentile rankings indicate the proportion of the reference population that scored below the test-taker. A percentile of 84 corresponds to approximately one standard deviation above the mean (IQ ≈ 115).

Limitation: A 20-item screening test has lower reliability than a full clinical battery (e.g., WAIS-IV, 100+ items). Results should be treated as indicative estimates, not clinical diagnoses. The standard error of measurement is ±5 IQ points.