DETECTION OF EXPERIMENTAL EFFECTS IN SOCIAL NETWORK ANALYSIS

Descriptive and inferential statistical techniques exist for the analy sis of social networks, but to date the inferential methods have been limited to the comparison of one network to its hypothesized populatio n parameters (analogous to a one-sample t-test), or the comparison of multiple relational structures measured on the same group of actors (a nalogous to a correlation coefficient). In this paper, we explore tech niques for comparing network structures when each network comprised an entirely different set of actors (analogous to a two-sample t-test or a between-subjects analysis of variance). Such methods for between-gr oup comparisons are critical in theory testing, where a researcher var ies an experimental factor for the purposes of studying its impact on some dependent variable (e.g. the resulting network structure). This a bility to test the significance of manipulated factors (like ANOVA) wo uld provide another important means by which social network analyses w ould be an aid to researchers. We propose several such statistical met hods for comparing network interactions. There are many areas of subst antive research in which groups are examined under different operating conditions, and it seems reasonable to study these inter-related acto rs as players in a network, and test the significance of the factors t hat had been experimentally manipulated, thereby capitalizing on the s trengths of both network analyses and the logic of analysis of varianc e and experimental design. In this paper, we illustrate our proposed m ethods on data representing coalitions formed under different experime ntal conditions.