classic test theory

n. Theory of testing in which each test score is supposed to be the sum of random error and true measurement. As random error is normally distributed, it allows computation of the reliability of scores either through repeated testing as in test-retest reliability or through examination of internal consistency of items as through part-whole correlations or more recently factor analysis. Later versions of test theory considered test scores to be the sum of truth, random error, and bias (consistent error), thereby facilitating calculations of validity.