After every test sits a small pile of data most teachers never read: which questions everyone missed, which everyone aced, and which split the class cleanly. Item analysis is just the habit of reading that pile. It takes ten minutes and it tells you whether your paper measured what you think it measured.
Two numbers per question
You only need two figures to start. Difficulty is the fraction of students who got the question right. Discrimination is whether the strong students outperformed the weak ones on it. Together they sort every question into keep, fix, or drop.
- Very high difficulty with low discrimination — likely ambiguous or off-syllabus; fix the wording.
- Mid difficulty with high discrimination — the workhorses; keep and reuse them.
- Everyone correct — either too easy or a giveaway; fine for warm-up, useless for ranking.
You already have the data — it is in the marks. Item analysis is just deciding to look at it before you reuse the question.
Hassan Tariq
Hassan helps schools read their own test data. He writes about item analysis, difficulty, and designing questions that actually discriminate between students.
Published