Reliability is defined as what?

• A. The degree to which a test is free from errors of measurement across forms, rates, and time.
• B. The degree to which different assessors give similar results.
• C. The degree to which scores match a normative sample.
• D. The degree to which a test predicts future performance.

Answer: (A)

Reliability is about consistency of measurement. A reliable test yields stable, repeatable results across different forms, across different scorers, and across time. That comprehensive sense of being free from measurement errors in multiple conditions is what the described option captures. In other words, the observed score should stay similar no matter which form is used, who scores it, or when it’s taken.

The other ideas describe different concepts. Matching a normative sample relates to how scores compare to a group, which is about norms rather than consistency. Predicting future performance is about predictive validity, not reliability. Interrater similarity is indeed a type of reliability, but it’s only one facet; reliability as a whole encompasses consistency across forms, raters, and time.