Modeling Networks

Latour’s keynote on Actor Network Theory as well as his introduction to Reassembling the Social provide an interesting frame through which to read Chapter 4 of Kadushin’s Understanding Social Networks. Kadushin (2011) points out in Chapter 4 that network statistics can only handle independent units of analysis. To take advantage of statistical analysis, then, means “cutting up large networks into separate non-overlapping pieces for analysis.” In light of Latour’s comments, this cutting process is no longer the simple process of making networks manageable for analysis, it is cutting them off from their environmental context. If networks are, as Latour suggests, fully dependent on the actors within it, this modular view skews the nature of the network being studied and of the individuals who make up part of that network. In talking about networks, Latour (2005) defines social as ‘a trail of associations between heterogeneous elements” (p. 5) and suggests the danger of mistaking the associations for the objects between which the associations lie. The modular view “interrupt[s] the movement of associations”  and “confuse what they should explain”(p.8).

One other concept I found interesting was etic groups. Kadushin describes these groups as being “identified by observers” not members, and without “firmly established boundaries.” In these groups membership, to a certain extent, is based on perception. The example that came to mind was race, since race is so frequently involved in Kadushin’s examples. Regardless of how an individual identifies him/herself, an individual is likely to be considered and treated as part of the race with which others associate his/her skin colour. In this context, perception defines the network reality of the individual.