Systematic observation methods in the study of small groups

Abstract

A comparison was made between the Bales and the Carter-Haythorn-Meirowitz-Lanzetta methods of systematic observation. The comparison was limited to the categorizing of verbal behaviour. It was found that there are several deficiencies in the latter method with respect to the classes used, and that satisfactory observer reliability could not be obtained. By using the Bales method, on the other hand, fairly high total observer reliabilities were achieved, and even the category reliabilities in various classes turned out to be satisfactory, provided the category frequencies were large enough. In a further study using the Bales method, the relative frequencies in several classes were found to be outside the limits determined by Bales. The main reason for this is presumably the dissimilarity between the American material used by Bales and the Finnish material of the present study. An attempt was also made to assess to what extent differences in the quality of the task presented to the groups cause differences in the categorical distribution of the acts observed. Discussions and problem situations tended to call forth very similar results, while the number of emotional reactions was found to decrease with increasing age and higher social level. Attention is drawn to the fact that intensity differences cannot be adequately taken into account when either of the methods discussed are used. Finally it is suggested that the categorizing of behaviour could be made more accurate if a two-dimensional system of categorizing were adopted in place of the present unidimensional ones.