ABSTRACT
It can be safely said that social network research has a relatively long past, yet a relatively short history. Although studies using social network methodology date from the early part of the 20th century (for instance, see Criswell, 1939; Evans-Pritchard, 1929; Moreno, 1934), such methods seem to be sparsely used in modern research. Such scarcity could stem in part from several problematic hurdles which researchers must overcome when choosing social networks as a method to study developmental phenomena, the biggest of which are inherent in data collection and data analysis. Data collection difficulty stems from the amount and type of information needed to fully assemble a network of relations among individuals. Additional concerns regarding consent of individuals named within a network complicate such efforts. Once network data have been collected, however, additional challenges meet researchers in the form of selecting the right type of statistical analysis to cull information from the network data. The following chapter seeks to present a survey of tools used for extracting groups of similarly behaving individuals from social network data. The methods presented in the chapter are aligned into two general categories: nonstochastic groups of individuals formed by meeting definitional requirements and stochastic groups of similarly behaving individuals. Following a brief introduction about statistical methods used in social network analysis, methods for each type of category are presented, framed in the context of the analysis of a social network data set.
