Our Methodology
How questions are designed, how raw answers become a score, and what that score does and doesn’t mean.
Question Design
Each question on the core test is written to target exactly one of four reasoning categories, spatial, numerical, logical, or applied, at one of nine difficulty tiers. Across the full 36-question set, every category appears once per tier, so the final composite score is never skewed toward one type of reasoning simply because it had more questions.
Difficulty tiers are calibrated by construction rather than after-the-fact: early tiers use short sequences and simple one-step logic, while later tiers require multi-step inference, larger working-memory load, or less intuitive numerical relationships.
Scoring Formula
Each answer is scored as correct or incorrect, weighted by its difficulty tier. Your raw result is the weighted share of total possible points earned across all 36 questions, not a simple count of correct answers.
That weighted fraction is converted to a z-score using the inverse cumulative distribution function of the standard normal distribution, then rescaled to a mean of 100 and a standard deviation of 15, the convention used across most published cognitive assessments. This is the same conversion formula, applied consistently, across every scored tool on this site.
What the Score Does Not Measure
A 25-minute, untimed, self-administered online test cannot replicate the controlled conditions of a proctored clinical assessment, which also evaluates factors like verbal comprehension and working memory span through methods a web format cannot reproduce.
Treat your result as a well-grounded estimate of fluid reasoning ability on the day you took it, not a clinical diagnosis. See our about page for more on our editorial standards, or read a sample results report.