連想法による文字刺激とそれに対応する線画刺激の意味の研究 : 1.自由連想法による

Purpose: To study the associative difference (or similarity) of some concrete concepts when presented through verbal mode (printed words) and pictorial mode (line drawings) of stimuli. Free association method was taken. According to the previous studies and Tada's hypotheses, the following hypotheses were formulated; 1. Sense impression responses (SIR) will be much evoked by words (W) than by pictures (P). 2. Nouns : W>P. 3. Adjectives : W>P. 4. Verbs : P>W. 5. P will take longer reaction time (RT) than W. Method: 1) Materials : A list of Japanese common nouns were pretested, and 40 nouns were selected to be used in the experiments. 2) Subjects : Exp. I; A total of 611 college students (freshmen and sophomors) were used, 271 for W-group, 340 for P-group. Exp. II; A total of 50 freshmen, 20 for W-group, and 30 for P-group. 3) Procedure : Experiments were conducted by groups using intact classes. In Exp. I stimuli were projected on a screen by an Overhead Projector, for 8 sec. with 2 sec. intervals. In P-group, instruction was given to respond with words "other than the names of pictures". Exp. II was run in a "Language Laboratory Room". Stimuli were presented through TV and responses were recorded. Other procedures are the same as Exp. I. Results and Discussion: 1) Rate of SIR was W(8.04%)>P(5.73%). Hypothesis 1 was supported, which was explained mainly from the difference between words and pictures in their denotative fields. 2) Rate of nouns was P(78.01%)>W(70.01%), which was contradictory to Hypothesis 2, and mainly explained from Paivio's dual coding hypothesis in its possible relationship to the effect of instruction to P-group. 3) Rate of adjectives was W(9.02%)>P(7.51%), which supported Hypothesis 3, Partly, the same explanation as 1) was presented, but there remained a problem of connotative affective responses. 4) Rate of verbs was W (16.91%) > P (11.17%) ; contradictory to Hypothesis 4. This was explained from the difference between W and P in their denotative fields of actions of referents. 5) Mean RT was longer with P (3,83 sec.) than with W (3.57 sec.), which supported Hypothesis 5 ; interpreted according to Paivio's dual coding hypothesis. 6) K-value was defined as the number of different kinds of responses. When K/N were compared, a little lower value was shown with P-group (though not significant), which was considered to suggest the narrowness of their denotative fields. 7) In all the indices above, rank corelations between W and P stimuli were significant, which suggested that associations of concepts were similar over W and P mode. 8) For a few pairs of stimuli, opposite tendencies were shown to the results as a whole, which suggested the possible exceptions of the explanations presented above.