How the skills assessment is built
Where the questions come from, how they are tested before they count and how the result is calculated.
The questions
The questions are drafted with the help of AI, six per domain, from the RAIF and RAIP syllabi. Each draft only moves on once a person has reviewed the stem, options, answer key and rationale.
The pilot
For 4 weeks, each approved question appears inside the RAIF and RAIP open assessments, two per sitting, marked on screen and without counting towards the score. The answer is stored together with the score the person got on the questions that count.
The statistics
Every week two numbers are calculated for each question. Difficulty (p) is the share of correct answers. Discrimination is the point-biserial correlation between getting the question right and the sitting score: a good question is answered correctly more often by those who do well on the rest.
The selection
At the end of the pilot, 4 questions per domain remain, thirty-six in total. Only questions with at least 30 answers, difficulty between 0.25 and 0.85 and discrimination of 0.1 or more qualify; among them, those with the best discrimination. If a domain does not have 4 such questions, the selection does not happen and the pilot continues.
The result
Each domain gets a band: below 50% to develop, 50% to 74% consolidating, 75% or more solid. The recommendation looks at both parts: below 60% in foundations, RAIF; with solid foundations and below 60% in application, RAIP; above 60% in both, RAIP as the path to RAIL.
The result is indicative. Four questions per domain point to where to study, but they do not measure anyone’s competence and do not replace a certification exam with a published blueprint and pass mark.
Status today
0 questions in pilot, 0 selected and 0 answers recorded.