Continued from the previous report where the interaction between P (the chooser) and o (the choosee) was mainly investigated, he present report deals with the interaction including q (the third person). This situation contains three relations : P's attitude towards q (R
1), P's attitude towards o (R
2) and P's perception of q's attitude towards o (R
3). Though they are considered as mutually interdependent, let us, in the present study, suppose that R
3 is dependent on R
1 and R
2 which are independent respectively.
We got a clue to the present study from F. Heider's hypothesis on the relationship between attitude and cognitive organization. When every relation considered was divided into two categorigs (+) and (-), he assumed that there existed a balanced state if either none or two of the three relations were (-). Then we nypothesize that R
3 will be determined as to bring a balanced state in Heider's sense according to the “type” (the combination of signs of R
1 and R
2). But we believe it is impractical to divide interpersonal attitude or its perception into two categories as Heider does. So we take a trichotomy by adding a neutral category (0).
The S's in the present study were children of six classrooms from fifth to seventh grade. In each class, twelve pairs of children of the same sex each were selected on the basis of data from a near-sociometric test and a test where S's where asked to mention some pairs which they thought either remarkably intimate or hostile in their class. These stimulus-pairs were arranged in random order and presented to S's. They were asked to judge, and to rate the attitude the child on the left side of a (q) pair should have towards the child on the right side of the pair (o).
The responses of R
3 thus gained were compared with those of R
1 and R
2 gained from the near-sociometric test. The results of statistical testing showed that our hypothesis led by Heider's might be said to be verified on the whole. The verifiablity of the hypothesis, however, was found to vary with the six “types” of the independent variable ; the results in some “types” accorded well with the hypothesis, while those of the others did not. But a further analysis of the data showed that our results differed from the Heider's hypothesis in two important points. (1) R
1 and R
2 contribute differently in determining R
3. 2) the distribution of R
3 is more or less contradictory to what is expected by the hypothesis, when R
1 is (-).
These two facts can, however, be included harmoniously in the generalized expression of our results that
the sign of R3 tends to be same as that of R2 independent of R1, though the Powerfulness of the tendency is dependent on R1. It may, therefore, be concluded that we need not rely, in order to explain our results, on the Heider's hypothsis which is universal and abstract, but on two factors which have a more concrete psychological meaning : the general tendency described above and the factor of “assumed similarity” presented by F. E. Fiedler.
抄録全体を表示