This document contains three papers from the Methodology Project of the Center for the Study of Evaluation. Methods for characterizing test accuracy are reported in the first two papers. "Bounds on the K Out of N Reliability of a Test, and an Exact Test for Hierarchically Related Items" describes and illustrates how an extension of a latent structure model can be used in conjunction with results in Sathe (1980) to estimate the upper and lower bounds of the probability of making at least k correct decisions. "An Approximation of the K Out [of] N Reliability of a Test, and a Scoring Procedure for Determining Which items an Examinee Knows" proposes a probability approximation that can be estimated with an answer-until-correct (AUC) test. "How Do Examinees Behave When Taking Multiple Choice Tests?" deals with empirical studies of AUC assumptions. Over two hundred examinees were asked to record the order in which they chose their responses. Findings indicate that Horst's (1933) assumption that examinees eliminate as many distractors as possible and guess at random from among those that remain appears to be a tolerable approximation of reality in most cases. (Author/PN)