Suppose, for example, that a passing point of 70% had been set for the scores plotted in Figure 3.There were 70 items on the test, making the cut score 49. The two conditional standard error of measurement figures for a score of 49 were 3.47 and 3.86; the first of these figures derives from application of Lord’s (1984) Method IV, and is a value appropriate in the usual case, when test items have been selected on the basis of specific criteria, such as subtopic, difficulty, or discrimination. The second figure, 3.86, is appropriate when items have simply been randomly sampled from an item pool, with no attempt to select them using a criterion.