KR-20 is an estimate of which property in tests with dichotomous items?

• A. Content validity
• B. Test-retest reliability
• C. Inter-rater reliability
• D. Internal consistency reliability

Answer: (D)

KR-20 measures internal consistency reliability for tests that use dichotomous items (such as right/wrong). It assesses how well the items on a test work together to measure the same underlying construct, by looking at how item scores relate to the total score. A high KR-20 means the items tend to be answered in a way that cohere with the overall trait being measured, indicating consistent measurement across items; a low KR-20 suggests the items may be tapping different constructs or adding noise. This is different from content validity (whether the test covers the intended material), test-retest reliability (stability of scores over time), and inter-rater reliability (agreement between scorers). For binary items, KR-20 is the appropriate way to estimate how reliably the test measures the same construct across items. Cronbach's alpha can be used more generally, but KR-20 is the dichotomous-item special case.