Union Conventions and Executive Boards: A Formal Analysis of Organizational Structure

Abstract

choose among these in some rational manner? Perhaps we may have to fall back on the criterion of simplicity, as ambiguous as this may be in many instances. If a theory such as that of status inconsistency can yield a large number of predictions about diverse dependent variables, and if no other single alternative can do so, then one would ordinarily prefer the simpler more all-inclusive theory. For example, if it were predicted that inconsistencies between ascribed and achieved statuses should produce strains, whereas those between achieved statuses should not, then one could test such a theory by examining different pairs of status variables to see whether or not the interactions occurred for the predicted combinations (e.g., interaction for race and occupation but not for education and occupation). We have seen that the status inconsistency formulation leads to unexpected complications in the form of identification problems. Obviously, if a simple additive model predicts almost as well to a given dependent variable as does this more complex theory, the additive model is to be preferred. Social theory is complex enough as it is, without our introducing further difficulties. In particular, one must guard against theories with too many unspecified parameters, or loopholes, that can be adjusted ex post facto to account for almost any set of empirical results.