On Stratum Factor Structure: Containing the Critics and Theorizing of Semantic Differential Method

The concept of stratum factor structure which explains the gap between connotative and denotative meanings is introduced after the re-evaluation of problems on SD method.This concept revealed that most of the problems were due to the differences of between-subjects variance on SD scales.The stratum factor structure of TAT No.I was discussed and the necessity of re-examination of personality factors was also proposed.Though the fundamental methodologies in SD studies were often discussed by several authors (Heise, 1969; N ordenstreng, 1969, etc.), there are many problems yet unsolved.The purpose of this article is to show by mean of a concept of stratum factor structure that those problems were pseudo-ones.E-P-A structure: In Osgood, Suci, and Tannenbaum (1957), the factor analytic results of adjectives revealed that the three factors can describe most of the variances.The first factor was labeled evaluation (E), the second potency (P), and the third activity (A).These were considered to be the fundamental dimensions of the meaning system.The evaluation factor is represented by the scales of good-bad, pleasantunpleasant, etc., the potency of strong-weak, heavy-light, etc., and the activity of quick-slow, active-passive etc. Osgood (1962) asserted that the generality of the E-P-A structure existed in at least 6 languages through the comprehensive cross-cultural studies of the meaning systems, and that an identical psychological process is assumed to exist in spite of their cultural and language differences.Though the E-P-A structure hypothesis was admitted, it does not bear the strict examinations which are discussed in the following items.The biopolarity hypothesis: This hypothesis can be divided into smaller ones, one of the SD scales and the other of the factor structure.Mordkoff (1963Mordkoff ( , 1965) ) revealed that the bipolar adjectives did not always function oppositely though the SD scales were supposed to be composed of two adjectives of reverse meanings.He selected 16 SD scales from Osgood et al. (1957), and 32 adjectives were rated on 10 SD scales by 41 students.Of the 16 SD scales, ten were proved to have functional bipolarity and 3 other scales had no evidence of it.Of the remaining, the scales of complex-simple differed functionally but only at 5%, masculine-feminine at 1 %, and hard-soft at 1 % levels.