Structural Versus Individual Effects

Methods for separating the effects of group structure or composition from individual effects have been proposed by Blau and by Davis, Spaeth, and Huson. Both methods are useful, but they do not always hold individual or group characteristics strictly constant as intended, thereby making it possible to obtain spurious group or individual effects. Several modifications of these techniques are proposed, making use of more precise matching and correlational techniques. The advantages and limitations of each are discussed. Several additional problems are considered, including the effects of deviants, overlap of distributions among groups, selection processes, and the conceptual definition of structural or compositional effects.