An Organizational Analysis of Collectivities

T ABLE 1 presents typology of collectivities. In Chart 1, these same types are arranged in cross-classification that specifies their defining properties. I can commend this classification, not on the grounds of its being complete or formally elegant, but, instead, as having grown from work on diverse problems, as being quite general in conception, and as correlating well with important variables. The analytic scheme that underlies this typology has developed with use. I want to characterize that scheme more fully than was possible in its earlier stages. Then I will consider two further questions: How is this type of classification related to other ways of looking at collectivities; what, in general, is the range of its power to order observations? I begin by reporting illustrative correlations in order to establish the desirability of undertaking the theoretical explication contained in this paper. We can get to these illustrative correlations by using the lowercase letters in the second column of Table 1. The first illustration will use the letter a which appears beside individuated heteronomy and certain other types and the letter b which appears beside commensalism and simple centralism. Child et al. (1958) have analyzed folk tales from sample of 54 primitive societies, giving each society score on the degree to which the theme of is present in these stories, (They follow McClelland and his associates (1953: 110-114) in defining need achievement as need to succeed in competition with standard of excellence.) McClelland (1961:66) shows that there is positive relationship (p-< .02) between these societies' scores on need achievement and the presence in them of full-time entrepreneurs. I had had occasion to classify 26 of the same societies according to the typology given in Table 1. For reasons reported later in this paper, I decided to see whether there was any relationship between this classification and the societies' scores on need achievement. As Chart 2 indicates, there is such relationship. The types of society indexed in Table 1 by the letter a are more likely than others to have high score. The value of chi square for this comparison is 12.32. (Chi-square was computed by the maximum likelihood method: df=2, p=<.Ol, C= .57.) Now compare the types of collectivities bearing the letter c with those indexed by d. These notations guide us to the results in Chart 3. Slater and Slater (1965) coded 90 primitive societies for signs of male narcissism and then performed centroid factor analysis on the resulting indices. They believe that their first factor (the first of two) is representative of the whole pattern of narcissism contained in their indices. The principal ingredients of this factor are indices of boasting and personal display. I had coded 38 of their 90 societies for another purpose, using the categories in Table 1.