Encyclopedia of Social and Cultural Anthropology
Network analysis in anthropology is more than the study of ‘networking’—expanding and making strategic use of the stock of contacts a person has. Minimally, a network consists of points and lines. Points can be persons, marketplaces, or organizations, and lines can be social relationships, †commodity or information flows, or shared board members.
*Sociologists have been more interested in organizational networks and interlocking directorates, and often study them with existing documents. No particular centre to the network need exist, and the strength or weakness of connections among the organizations is the object of ‘block analysis’ (Scott 1991). Similarly, in anthropological studies of *market networks, no individual marketplace is given centrality; the place of each market in a hierarchical network is determined through fieldwork and quantitative analysis (Skinner 1964–5).
Networks that have a person at their centre are termed ‘egocentric’ networks, and this was the focus of network analysis as developed by social anthropologists in the 1950s and 1960s (Sanjek 1974; Scott 1991). In exploratory studies and conceptual papers, hope was expressed that charting and analysing egocentric networks would be as valuable in urban settings (where alters include many non-kin) as the *genealogical method was in kin-based societies.
The problem is how to determine that an alter is ‘in’ an ego’s network (Sanjek 1974). Some anthropologists see this as a cognitive question, and ask an ego to list their alters. One exhaustive study of two Maltese networks found 638 alters in one and 1,751 in the other; the first Maltese, however, exchanged conversation or visits with only 128 alters, and the second with 700. With no clear theoretical goal set for this data-gathering, however, the investment in research time was enormous, and few more such studies were undertaken.
Another methodogical tack is to observe (or recover through interviews) the alters actually encountered by an ego. The study of action-sets—the alters whom candidates interact with during an Indian electoral campaign, for example, or who appear at English christenings, Navajo ceremonies or Zambian funerals—is one way to isolate a discrete portion of a behavioural network, and to make comparisons among a set of egos. A second behavioural network method uses data on alters encountered by an ego over a delimited time period. Thus Sanjek (1978) compared networks of forty Ghanaians, each studied over four days, in terms of theoretical issues regarding the extent of ‘tribalism’ in urban life. He found that more than half the interaction scenes in the networks were polyethnic, and that in many others the co-ethnic alters were kin of the egos.
Since the 1970s, few network studies have been published by anthropologists. Network analysis requires diligent fieldwork, and even then needs to be contextualized and interpreted with other sorts of information. In sociology, a vast technical vocabularly related to formal measures now marks network analysis (Scott 1991); in anthropological discourse, network appears today more often as metaphor (when speaking about ‘networking’) than as method. But when theoretical questions that can be answered with network data are posed, network analysis remains a valuable if under-used research tool (Boissevain 1979; Sanjek 1978).
ROGER SANJEK
See also: markets, complex society, urban anthropology
Further reading
Boissevain, J. (1979) ‘Network Analysis: A Reappraisal’, Current Anthropology 20:392–94
Sanjek, R. (1974) ‘What is Network Analysis, and What is it Good For?’, Reviews in Anthropology 1:588–97
——(1978) ‘A Network Method and its Uses in Urban Ethnography’, Human Organization 37:257–68
Scott, J. (1991) Social Network Analysis: A Handbook, London: Sage
Skinner, G.W. (1964–5) ‘Marketing and Social Structure in Rural China’, Journal of Asian Studies 24:3–43, 195–228, 363–99
This is the complete article, containing 577 words
(approx. 2 pages at 300 words per page).
View More Summaries on Network analysis
