METHODOLOGY · VERSION 1.0
OMIB methodology: source, scoring and limits.
Here are the origins of the tasks, our procedure and the conclusions this first version supports.
Where the tasks come from
Marco Koch, Frank M. Spinath, Samuel Greiff and Nicolas Becker published OMIB in 2022. It provides 220 matrices with parameters and research information. The item bank is licensed under GPLv3. We retain the notices and provide corresponding source materials.
Our form: a 22-task pilot
We use the 22 items from set 1 and two practice examples from set 10, without overlap. The published study used 28-task forms: 22 set items and six anchors. Our version does not reproduce that full procedure. We therefore do not attribute the published alpha of 0.92 to this adapted form.
How we score
Build the missing figure by selecting from 20 components. A response earns one point only if it matches the complete solution. There is no partial credit; an omission earns zero. Practice examples are not scored. Your report shows correct answers, completed and omitted tasks, and the form version. Elapsed time is reported but does not change the score.
What still needs to be studied
We have not yet validated or normed this application in English or Spanish. The original German sample does not automatically represent a global audience. We do not convert correct answers into IQ, population percentiles, diagnoses or giftedness labels. Translating an interface does not validate a new population.
Conditions and accessibility
This pilot is for adults aged 18 or over. There is no time limit. The figures require visual pattern recognition; accessible component labels do not establish equivalence for people who cannot see the matrices. Use a comfortable screen and zoom if necessary, recognising that these conditions still need study.
Transparency and repeat attempts
The bank is open: people can know its solutions, and practice can affect correct answers. We do not use this test for clinical, educational or employment decisions. Licence notices and source materials contain no participant data or email credentials.
How to check the procedure
If a task requires two elements, selecting only one earns zero. Selecting exactly both earns one point; adding an unrelated element earns zero. This rule distinguishes a partially similar selection from a complete response.
Each result identifies version omib-set1-pilot-v1. Time is reported without bonuses or penalties. Practice examples include feedback and are not part of the 22 possible points.
Check the two worked examples · Understand the result