@article{oai:nagoya.repo.nii.ac.jp:00002069,
 author = {内田, 敏夫 and UCHIDA, Toshio and 内田, 良男 and UTIDA, Yosio},
 journal = {名古屋大學教育學部紀要. 教育心理学科},
 month = {Nov},
 note = {The purpose of this paper is to investigate, as a first step, the context effect in the problem of visual similarity of figures. This paper consisted of 4 experiments. In each experiment, standard stimuli were simple geometric figures and the influence that the context conditions had on observed similarity between each of standard stimuli was analyzed. In Experiment I, one working hypothesis was made. In this experiment, the procedure that "different" stimuli (random figures) were introduced into the group of geometric figures (circles & ellipses) was carried. Ss were asked to measure similarity with insight by arbitary number from 0 to 100. Ss were 6 adults. As a result, the content of hypothesis was as follows : (1) The more "difference" is, the more observed similarity between each of standard stimuli would be. (2) The more the ratio of "different" stimuli in context stimuli is, the more observed similarity between each of standard stimuli would be. In Experiment II, the purpose was to prepare 4 context conditions consisted of various "difference" to verify the hypothesis. After the classification data were transformed into proximity data, each of Kruskal's nonmetric MDS and hierarchical clustering was applied to the proximity data. Ss were 9 adults. As a result, the order of "difference" of each context stimuli condition was as follows : (1) Area reduced stimuli (2) Apparent cube stimuli (3) Area reduced & randomized stimuli (4) Randomized stimuli (Standard stimuli were rectagle figures) In Experiment III, the hypothesis (obtained in Experiment I) was verified. The context conditions were the ones formed in Experiment II, and the procedure was almost the same as in Experiment I. Ss were 8 adults. As a result, only hypothesis (2) was proved. In Experiment IV, multidimensional analysis on the context effect was tried. The context stimuli were the ones used in Experiment III, and similarity was measured as order scale value. The proximity data were analyzed with Kruskal's nonmetric MDS. Ss were 8 (partly 4) adults. The results were as follows : (1) The number of dimensions was 2 or 3,and no systematic change with context condition was found. (2) It was the dimension of area rather than of form that was influenced by context effect. (3) The more "difference" of context condition was, the more the original pattern formed with standard stimuli changed. (4) On the Minkowski's r-metric, Euclidean model was most suitable for all cases., 国立情報学研究所で電子化したコンテンツを使用している。},
 pages = {55--66},
 title = {<原著>視的類似印象の研究 : 図形の類似性認知における刺激文脈効果について},
 volume = {24},
 year = {1977}
}