Item difficulty

The proportion of a reference sample that answers a particular item correctly. Easy items (proportion correct > 0.7) discriminate poorly among above-average test-takers; hard items (proportion correct < 0.3) discriminate poorly among below-average test-takers. Modern adaptive tests select items at difficulty levels matched to the test-taker's estimated ability.

This term appears throughout the cognitive ability literature and across this site's articles. Understanding it is essential for interpreting any IQ score or cognitive subtest result. Modern psychometric textbooks (such as those by Anne Anastasi or Susan Embretson) cover the term in significant additional depth and document the empirical findings that justify its prominence in the field.

In the context of online IQ testing, the implications of this term are usually that the test-taker should be cautious about over-interpreting brief screener results. Most of the published precision claims for major IQ batteries do not transfer directly to short online instruments, and the relevant adjustments — wider confidence intervals, more conservative band assignments — are best made explicitly rather than ignored.

For further reading on this term, consult the related entries in this glossary and the deep-dive articles linked in the Related Reading section. The American Psychological Association's task force report 'Intelligence: Knowns and Unknowns' (1995) and its follow-ups remain the most authoritative summary at an accessible technical level.