Psychometrics

Item difficulty (P-value)

How hard a question proved, expressed as the proportion of candidates who answered it correctly.

What is item difficulty?

Item difficulty, usually reported as a P-value, is the proportion of candidates who answered a question correctly. It runs from 0 to 1: a P-value of 0.9 means 90% of candidates got it right (an easy item), while 0.3 means only 30% did (a hard item).

Counter-intuitively, a higher P-value means an easier question. For objective questions the value is often chance-corrected, so that guessing (for example, a one-in-five chance on a five-option question) doesn't inflate how easy an item appears.

Why it matters

Item difficulty helps you build a balanced exam and spot questions that are too easy, too hard, or not behaving as expected. On its own it isn't enough, though — a question can be a reasonable difficulty yet still fail to separate strong candidates from weak ones, which is why it's read alongside the discrimination index.

Synap reports the P-value on every question — part of its psychometrics and standardised testing tooling.

See rigorous assessment in action

Book a tailored demo and we’ll show you how Synap puts these concepts to work in real exams.