For Teachers: Setting a CBC Exam That Actually Measures Strands

Most weak exams are weak before the first question is written: they sample whatever topics came easiest to set, in whatever mix of difficulty emerged by accident. A CBC exam is an instrument for measuring strands, and instruments are designed. The design tool is a blueprint table, and it takes twenty minutes that repay themselves at marking time.

The blueprint: decide the sampling before writing

The rules the table enforces: every strand taught appears in roughly the proportion it was taught; no strand is examined only at recall level; and application items, where the learner uses the strand on an unfamiliar case, carry real marks. An exam of thirty recall questions measures memory, whatever the paper's title says.

Write the marking scheme first

Writing the scheme before polishing the question exposes bad questions instantly: if you cannot write the acceptable answers and their mark splits, the learners cannot divine them either. A good scheme states the points that earn each mark, the acceptable alternatives, and what earns nothing, which is exactly the discipline our rubrics guide applies to performance tasks. Report language for the results belongs in the BE, AE, ME, EE guide, and lesson-level alignment in the lesson plan guide.

Borrow a working model

The fastest way to calibrate is to mark against a professionally built pair. The Grade 9 Integrated Science Term 2 exam with its matching marking scheme shows the blueprint method executed end to end: strand sampling, mixed demand levels, and a scheme that teaches while it marks. Set your own paper beside it once, and the twenty-minute blueprint becomes a habit your learners feel in fairer marks.