Applies to evaluating reliability. A statistical estimate of the amount of random error in the assessment results/scores (from classical test theory). Inversely related to the Standard Error of Measurement (SEM). In theory, it is the correlation between scores of two strictly parallel forms.