Abstract
This study examined whether Rest's (1979) objective test of moral comprehension could yield comprehension stage data that form a Guttman scale. If so, the test's construct validity and its usefulness for certain research purposes would be enhanced (Rest, 1973). Subjects were 197 adults in junior college, college, or graduate school. Each completed Rest's (1979) Defining Issues Test and an 11-item version of Rest's multiple-choice test of moral comprehension, which assesses comprehension of Kohlberg's Stages 4, 5A, and 5B. For each comprehension item the subject chooses one of four multiple-choice alternatives. A subject was scored as comprehending Stage 4 if he answered at least three out of four Stage 4 items correctly, Stage 5A if he answered three out of four Stage 5A items correctly, and 5B if he answered both Stage 5B items correctly. These criteria were chosen because the probability of reaching or exceeding each criterion by randomly guessing is low and quite uniform across the three stages: 5% for Stage 4 and Stage 5A, and 6% for Stage 5B. Using Henry's (1952) method for assigning scale scores to non-scale-types (see Nie, et al., 1975), the coefficient of reproducibility was .95 and the coefficient of scalability was .84. Thus, the comprehension stage data do form a Guttman scale.