Group Problem-Solving and Communication

Problem : The purpose of our investigation was to study the following four questions in order to confirm the main results of H. J. Leavitt's experiment (1).(1) Can we find out any difference in the method of problem-solving between groups in problem situation by changing their respsctive group structure successivevly?(2) Is there also any difference in communication frequency according to the change of group structure in the problem situation?(3) Is there any change in each group-member's position by the change of group structure in problem-solving situstion?(4) Do the member positions in each group pattern have any differential effects on the degree of their participation?Of each question, which centered on the problem of communication, we examined the time required for group problem-solving from the point of view of group as a whole, and secondly analysed communication frequency from that of individual member difference.Procedure : Experimental subjects were 20 boys of average chronological age 14.7, who were students of a junior high school in Kyoto.Experimental procedure and materials were mostly the same as which Leavitt used (cf. Fig. 1, 2.).We composed, however, only 4 groups and each group was successively put through all of 4 structural patterns (circle, chain, wheel and Y), according to Latin Square design which secured genuine pattern differences as distinguished from those of group indiveduality.As we see in fig. 3, each group was tested once a week and the experiment took 4 weeks for each group (cf. Fig. 3).Results : From our experiment the following results were deived ;(1) By means of analysis of variance on Latin Square design described above, we analysed the difference of time required for group problem-solving of groups in each pattern but we could not find out any significant statistical differeces in both patterns and groups.But we found that the total sum of time required for problem-solving in a group seemed to decrease with the increase of trails (cf. Table 1).(2) Also we found out no time differences by patterns both in the shortest and the longest time of the problem-solving through all trails (cf. Table 2).(3) On the communication frequency, i.e. the number of messages sent out, there was no statistical difference in patterns, but the number seemed to decrease with the increase of trials (cf. Table 3).(4) As for the positional effects in each group, significant statistical differences were found in problem-solving time among 5 positions in all groups except 2 or 3 in 16 group-pattern combinations (cf. Table 4).(5) Comparing the differences of commnunication frequency between positions of all groups in the 1st week, we found the clear positional differences between central and peripheral positions in Wheel and Y, but no differences in Circle and Chian (cf. Fig. 4)(6) By the analysis of the questionnaires after the experiment. we found that, the leader of a group tended to emerge in the most central position (cf. Table 7).(7) For each group, we computed Group Participation Indices which were rations of each position to the most Central one, C, on the time required for sending one unit of comunication.And we are led suppose that when any of these approsches I, the corresponding position-member tends to have almost the same degree of activity as the central one in each group (cf. Table 6).