Citation

Johnsen EC. Soc. Netw. 1986; 8(3): 257-306.

Copyright

(Copyright © 1986, Elsevier Publishing)

DOI

10.1016/0378-8733(86)90007-9

PMID

unavailable

Abstract

In this paper we study the relationship between certain fundamental social microphenomena and the micro- and macrostructures they generate in a social group or network. The microphenomena of interest concern the effect of agreement or disagreement between members of the group on the friendship, respectively, nonfriendship between them and vice versa. We examine various versions of this type of microprocess and focus our attention on five which, at equilibrium, are in close structural relation to the five affect models arising in the Davis--Leinhardt set of empirical sociomatrices. Each version determines a micromodel consisting of the sets of triads which are permitted and forbidden in the group, and each of these in turn determines a macromodel of typical macrostructures of the group in which all of the permitted triads appear but none of the forbidden ones. We derive these micro- and macromodels and determine their structural relations to the five empirical affect models. One of these process models corresponds to a modification of Newcomb's "positive balance", for which the corresponding micro- and macromodels define the "direct positive influence" model of Johnsen and McCann. Because of the structural closeness of two of these process models to the empirical affect models, we see that in groups presumed to be moderately close to social equilibrium, processes by which friendship induces agreement (or disagreement induces nonfriendship) much more nearly account for these affect structures than do processes by which agreement induces friendship (or nonfriendship induced disagreement). In fact, it appears that the empirical macrostructures for affect in social groups which arise in the Davis--Leinhardt data may be accounted for to a considerable extent by the inducement of agreement by friendship (or nonfriendship by disagreement), and this is also true for the last week's data in the second year of Newcomb's well known study of the acquaintance process. These two models appear fundamental in that both contain Heider's "balance" and Davis' "clustering" as special submodels, that sandwiched between them is Holland and Leinhardts' "transitivity" model, that the less restrictive of the two also contains Davis and Leinhardts' "ranked clusters of cliques" as a submodel, and that the more restrictive of the two, the modified "positive balance" or "direct positive influence" model, is also a submodel of Johnsen's "hierarchical cliques". The theoretical results of this paper, together with substantive results from earlier empirical and experimental studies, are used to construct a broad description, with some detail, of how friendship structure can emerge and decline in social groups for which other persons, issues and values are important to the members.