This paper (1) proposes a new method for studying attitude by revised application of a series of statistico-mathematical models, i.e. the “Quantification Theory” which were developed by Guttman, L, and Hayashi, C.; (2) presents an empirical des-cripiton of the structure of political attitude and the pattern of political behavior of the Japanese based on the new method; and (3) tests the various kinds of models that have been proposed so far on the structure of political and social attitude.Of these models, a great many tend to be either subjective or speculative in their methodology. On the other hand, are usually based on a measurement developed through factor analysis.In an attempt not only to avoid subjectiveness or speculativeness, but also to revise many of the theoretical deficiencies in factor analysis, the present paper applied “Quantification Theory” to structural models of political attitude.As a result, a description was made concerning both the structure of politcal attitude and the pattern of political behavior.Method:Subjects: Electorate of both sexes over 20 years of age.Sample size: 800 subjects, Tokyo (23 wards), and 840 subjects, Chiba City, Chiba Prefecture.Questionnaires: A total of 44 items are included which are to be broken down as follows:(1) Political behavior (7 items)(2) Political attitude (25 items)(3) Social attitude (4 items)(4) Demographic information (8 items)Procedures:(1) The subjects were interviewed individually and answered to the questionnaires that were given orally.(2) To meet the limitation of the computor capacity, 29 items were selected out of the aforementioned 44 items on the basis of the significance of the items, which ranged over 100 categories. Calculation was made through Guttman-Hayashi multidimensional quantification method.Result: Three dimensional vectors were allocated to each of these categories (Table 1 and 2). For the cluster of items on the “party identification” the regressions of the metric component weights, the second component weights, and the third component weights on content percentile (direction of party identification) showed monotonic, U-shaped and N-shaped curves (Fig. 5, Fig. 6 and Fig. 7).Disscussion: The above results lead the author to make the following interpretations.1. The structure of the political attitude of the Japanese is to be observed most efficiently in terms of the three basic dimensions; (1) Progressive vs. Conservative (Direction), (2) Indifferent vs. Interested (Indifferency), and (3) Progressive Critical vs. Conservative Critical (Stability).2. The “content”, “intensity” and “closure” are both mathematically (theoretically) and psychologically (empirically) identified by Guttman, L. as basic functions of the attitude when clusters of items are perfectly scalable. The present study successfully extacted the same three basic functions by applying newly proposed methodology despite the fact that not all theitems were scalable. Furthermore, it was assumed that the above three basic dimensions, i.e., (1) Progressive vs. Conservative, (2) Indifferent vs. Interested, and (3) Progressive Critical vs. Conservative Critical, are themselves the basic functions of attitudes in the area of “party indentification”. In other words, they are “content”, “intensity” and “closure” in the field of “party identification”.In the subsequent discussions, an attempt will be made to describe and consider, by applying the above three basic dimensions, a pattern of specific political behavior and a structure of political attitude.
Read full abstract