The Adaptive IQ Test
Contents
Development and Norming
The Adaptive IQ Test was created by analyzing world-leading cognitive assessments with exceptionally high g-loadings between 0.8 and 0.96 - tests like the Stanford-Binet (SB-V) and WAIS that are typically only available through expensive in-person evaluations. We examined these gold-standard assessments to understand their underlying cognitive constructs and problem-solving patterns, then developed our own original pattern-based questions that measure the same abilities while being suitable for online administration.
Real IQ collaborated with licensed psychometricians to choose items, analyze norming data, and evaluate g-loading. The test underwent extensive norming with both online participants from a diverse global population and a smaller proctored baseline study, with both norming methods yielding remarkably consistent results in score distributions and psychometric properties.
Key features of the norming process included:
- Acceptance of only first attempts to prevent practice effects
- Automatic exclusion of entries where participants minimized the browser window or switched tabs
- Computer-generated questions that change with each page load, effectively preventing answer lookups
The norming data demonstrated high reliability, with analysis showing that participants typically required 7 or more attempts to achieve meaningful score improvements. Investigation revealed that most early score variations were attributable to guessed answers rather than genuine performance changes. The convergence of results between online and proctored administration provides strong evidence for the test's validity, reliability, and g-loading.