Abstract
289 Air Force students were presented in counterbalanced order within-Ss two forms to gauge their attitudes toward military technical training. The content of each of 55 items was structured according to two different formats, a Likert-type and a Guttman-type configuration. The forms were similar in the sense that for each item written in the Likert format there was a corresponding item written in the Guttman format which had essentially the same content. Not only were all items within one form of the same structure, but all were composed in such a manner to reduce as much as possible variability among items due to phraseology. Each item was randomly placed in its sequential position among the 55 items of each form; and each item maintained the same sequential position within both forms. Items were scored according to customary Likert and Guttman procedures. A multiple discriminant analysis and its associated statistics were computed for these data between measurement formats. Ss demonstrated significantly more favorable attitudes toward the same content areas on the Guttman-structured items than on the Likert-structured items. Evidently, item format did affect the degree of the evaluative assertions ascribed to the attitude universe. These results were explained in terms of several alternative theories, namely, anchoring, adaptation-level, and intensity interpretations.