Abstract

This study proposes a taxonomy of causal attributions appropriated for marital relationship. Study I examines the content of spouses' causal attributions most common in marriage and the relevance of the direct attribution inducing method which was used to investigate causal attributions. Attribution data from 338 married people were classified into 183 attribution categories through content analysis. A frequency analysis of 183 attribution categories showed that the actor attribution had the highest frequency for both positive and negative behavior of spouses, followed by the third parties/circumstances/stimulus attribution, the self attribution that is the object of the behavior of spouses, and the attribution to both sides. In order to examine the relevance of the direct attribution inducing method, two attribution inducing methods, a direct method in which an actor directly described causes for the behavior and an indirect method that extracted attributions from the data of thoughts or feelings that arose sponstaneously when partner behavior occurred, were used. .72 agreement rate on attribution category was achieved. This result indicates that there is little difference between direct and indirect methods. Study 21 investigates dimensions underlying attributions derived from study 1. First, among 183 attribution categories, 70 terms were selected on the basis of frequency and representativeness for each category. Then 69 undergraduate and graduate students classified them into 10 categories according to similarity of meaning. A multi-dimensional scaling analysis of these data showed three dimensions(stress=.07, R²=.96). The interpretation of these dimensions was guided by data from 60 additional subjects who rated the 70 causes on 24 bipolar scales. As a result, the first dimension was interpreted as 'self or spouse-other than self or spouse'(relationship-other than relationship), the second as 'spouse-other than spouse' and the third as 'both sides-other than both sides'. The result of the cluster analysis also shows that the interpretation of each dimension acquired from the analysis of the multi-dimensional scaling was relevant. Finally, the significance, implications, and limitations of this study were discussed, and issues to be investigated in the future studies were mentioned.