The design of this study was straightforward. The plan consisted of requiring subjects to rate 50 concepts on a large number of GD scales as well as on the 12 verbal scales of the SD. These ratings would then be intercorrelated and factor analyzed so QS to identify those GD scales that acted most like the E-P-A scales of the SD instrument. The factor structure, the size of the loadings of their defining scales, and intercorrelations between GD and SD composite scores would form the criteria by which we could evaluate the existence of cross-culturally valid pictographic scales.
The Test Booklets. The selection of GD scales was done on a purely intuitive basis. We asked a number of our cross-cultural colleagues to send us simple abstract designs which we either replicated in their original form or adapted to the designs we made up. We retained a total of 64 pairs and these are reproduced here as Fig. 1. The size of the squares enclosing each pictograph was 25mm x 25mm, so that eight scales could fit on a standard 8" x 11" page. The designs were originally drawn with black ink in large size on individual cards then reduced to the final size through photographic process. The left-right orientation of each pair of pictographic alternatives was determined by a random process. Stencils were cut through an electronic process and run off on a standard Gestetner mimeograph machine. The mimeographed sheets were then mailed to each field location where they were assembled in booklet form after the 50 concepts were added (written in, typed, or printed) in the appropriate language. These concepts were randomly chosen from a set of 100 words previously used in the work on the development of the pan cultural SD and had already been translated in the various languages. The words in English are listed in Table 1.
Each field location received a standard set of instructions which were to be followed in the execution of the task. There were to be three booklets and two testing sessions for each subject. The first booklet was to consist of 48 pages as follows: six concepts were to be rated on all 64 GD scales, there being eight scales per page; thus, each concept appeared on top of eight successive pages, The second booklet U85 made up of 32 pages: four concepts each being rated on the 64 GD scales. The third booklet consisted of the appropriate SD scales in each language. There were 12 scales per page with the concept to be rated appearing on top. These consisted of the ten concepts of booklets 1 and 2. The six concepts for booklet 1 were to be chosen at random and half the booklets for each group were to be assembled in one random order and the other half in the reverse order. Similar random procedures were used for booklets 2 and 3.
Each subject rated all three booklets: booklet 1 during session one and booklets 2 and 3 (in that order) during session two, The testing was done in groups, there being at least 2325 subjects in each group. Since each subject rated only 10 concepts, five groups were needed for the 50 concepts. Each session lasted about an hour. The subjects in each location were drawn from a school population of young men and women within the age range of 17 to 21 (in the US these were college freshmen and sophomores).
Instructions. Written instructions appeared in the front of each booklet. These were adapted from the usual instructions with the semantic differential technique (Osgood, Suci, and Tannenbaum, 1957). Clarifications about the task were given orally by the experimenter in charge whenever these seemed desirable.
The Participating Sites. In the selection of sites for data collection an attempt was made to diversify as much as possible both culture and language, at least, as much as our network of 28 cooperating language/culture communities allowed for. The present report summarizes the findings for the following five locations: American-English (AE; Urbana, Illinois), Delhi-Hindi (DH; India), Finland-Finnish (FF; Helsinki), Germany-German (GG; Munster, Westfalia), and Japan-Japanese (JP; Tokyo). The pancultural SD scales for each language are given in their English (approximate) translations in Table 2. The numbers 55-7 give the scale identification for subsequent discussion. The letters in parenthesis preceding each scale refers to the affective dimension it taps: E = Evaluation, P = Potency, and A = Activity. The scales are oriented in the random direction presented in the test booklets. For each scale, that qualifier is underlined which represents the positive aspect of the factor keyed towards the American-English scales (viz. for E: helpful', nice, good, sweet; for P: big, strong, powerful, deep; for A: young, fast, alive, noisy).