Spatial arrangement of social and economic networks among villages in Nang Rong District, Thailand

This paper examines the spatial arrangement of social and economic networks among villages in Nang Ž . Rong district, Thailand. We use spatial information from a geographic information system GIS for the Ž . district to help interpret the patterns of movement of agricultural equipment large tractors between villages, of people into villages for temporary labor, and of people to village temples and to elementary and secondary schools within the district. Once social networks have been incorporated into the GIS they can be mapped in relation to geographic features of the district, such as topography, landcover, and locations of roads, rivers, and villages. Not only does geographic information about village locations allow us to properly orient the graphs of these networks, but the resulting visual displays reveal strikingly different spatial arrangements for the five networks. Networks of shared temples and elementary schools link small sets of villages in close geographic proximity whereas tractor hiring, labor movement, and secondary school networks bring together larger sets of villages and span longer distances. Information on landcover from satellite digital data provides insights into the patterns of network ties throughout the district and shows a clear relationship between tractor hiring networks and type of agricultural activity in the district. The spatial analytic capabilities of the GIS also allow us to assess the impact of the administratively defined district boundary on our measured relations and to evaluate whether rivers and perennial streams create barriers to network ties between villages. q 1999 Elsevier Science B.V. All rights reserved. ) Corresponding author: Tel: q1-803-777-6848; fax: q1-803-777-5251; e-mail: [email protected] 0378-8733r99r$ see front matter q 1999 Elsevier Science B.V. All rights reserved. Ž . PII: S0378-8733 99 00014-3 ( ) K. Faust et al.rSocial Networks 21 1999 311–337 312 1. Social networks and spatial networks How do networks of social and economic engagements link communities into larger regional-level systems? How are spatial proximity, features of topography, and land use related to economic and social networks between communities? This paper addresses these questions by looking at the spatial organization of social and economic linkages among villages in Nang Rong district, Thailand. We incorporate networks of five kinds of ties — sharing temples, elementary and secondary schools, labor movement, and Ž . tractor hiring — into an existing geographic information system GIS for the district to examine variation in spatial patterning of social networks and the relationship between these networks and such geographic constructs as proximity and land use. Doing so provides insight into how spatial aspects of land use are associated with patterns in networks of social and economic relationships among villages. Also, representing social networks spatially makes potential data error more visible, providing a valuable tool in checking data quality, and makes spatial pattern and distribution more obvious as a possible indication of landscape form and function. The proper spatial location of network actors aids network visualization. Well-desigŽ ned sociograms draw attention to important features of networks Klovdahl, 1981; . Freeman, 1996, 1997; McGrath et al., 1997 and proper geographic locations for nodes can provide additional insights into spatial aspects of network patterns. Interactions between people and between communities are influenced by opportunities and barriers to contact. The frequency of travel and visiting across locations depends on proximity, accessibility, and the characteristics of available travel routes. Some of these opportunities and barriers are spatial or geographic, yet social network analyses have seldom studied either spatial configurations of relations or how geographic features might shape social networks. In fact, most social network data sets do not contain information about the geographic locations of actors. Even analyses of networks in which actors have Ž spatial locations such as networks of economic and political transactions between . nations or alliances between communities do not incorporate geographic locations and accompanying spatial information into network models. Over the past several decades a handful of studies have looked at regional-level networks among communities where nodes for communities can be spatially referenced. These studies have incorporated spatial information into their representations of networks but have paid little attention to spatial aspects of networks in their models. River routes of trade between villages have been used to study how Moscow’s centrality in Ž . this network affected its rise as an influential city within the region Pitts, 1965, 1978 . Ž . Intervillage networks have also been used by Halpike 1970 to study patterns of Ž . alliance formation between Konso, Ethiopia towns, by Crump 1980 to compare the communications infrastructures between villages in Zinacantan and Chamula, Mexico, Ž . and by Barkey and Van Rossem 1997 to study contentious legal actions between Ž . Ottomon Empire towns. Morris et al. 1996 use networks of social interactions and visiting patterns between villages in Uganda as part of their analysis of the social and spatial basis for the spread of HIV. Networks within towns have been used by Ž . Ž . Hammond 1972 and Doreian 1987; 1988 to study the spatial arrangement of Mayan ( ) K. Faust et al.rSocial Networks 21 1999 311–337 313 Ž . ceremonial plazas. On a larger scale, Hage and Harary 1991; 1996 use network models to study regional exchange systems among islands in Oceania. Although these examples represent networks using sociograms in which points can be Ž spatially referenced nodes for villages are in appropriate geographical locations; ceremonial plazas are placed properly within the villages; positions of islands in Oceania . correspond to the map of the area subsequent analyses of these networks employ only the abstract graph — that is, they study graph theoretic properties based on linkages between nodes — but do not incorporate geographic distance or other properties of the Ž . landscape, topography, or spatial relationships topology among points. Nor do they investigate how locations of points, proximity, travel costs, landscape, topography, land use, spatial density or other geographic features are associated with characteristics of the social networks. In contrast to the lack of attention to spatial features in social network research, social geographers have long employed graph theoretic ideas and spatial network models to Ž represent and analyze flows between locations Haggett and Chorley, 1969; Evans et al., . Ž . 1995; Walsh et al., 1997 . Nodes represent locations cities, communities, etc. and edges represent flows between points, such as roadways, rivers, or air traffic. Geographers then use graph theoretic concepts to model flows between locations and to calculate optimal locations for services such as fire stations or hospitals. However, with Ž the exception of movements of people or commodities e.g., commuter traffic flows . from suburbs to cities , geographers have only minimally incorporated social networks or information on explicitly social transactions between occupants of different locations into their spatial models. Nor have geographers employed the wealth of social network analytic concepts and models. Geographers have been concerned with the spatial patterning of social networks, organizations, and institutions particularly as those patterns relate to environmental factors and landscapes. Much of this work has been fostered by cultural ecologists and political ecologists in geography who have relied on qualitative ethnographic analyses, case study sampling, and intensive field ethnographies to study social systems and their Ž influences on environmental systems e.g., Denevan, 1989; Forsyth, 1996; Bebbington, . Ž . 1997 . For instance, Bebbington 1997 analyzed the spatio-temporal relationships between agricultural intensification and local social–spatial network developments associated with cooperative water management strategies, institution building for market development, and other social capital development initiatives in the Andean Highlands. Geographers have also recognized the opportunity to integrate these field-oriented ethnographic studies with quantitative studies that account for broader-scale processes in both the political and the biophysical environments. Geographers have called for ‘‘hybrid research approaches’’ in order to link investigations of global change processes Ž . to sustainable development studies Batterbury et al., 1997 . Such approaches incorporate quantitative and qualitative research methodologies through the study of land usercover change, which represents the most fruitful opportunity for the political ecologists and global change scientists to integrate quantitative and ethnographic reŽ . Ž . search approaches Turner, 1997 . Gould 1996 argues that physical geography and human geography have always brought the spatio-temporal perspective to analyses of process. ( ) K. Faust et al.rSocial Networks 21 1999 311–337 314 Although both social network analysis and social geography employ graphs and graph theoretic concepts to analyze relations between units, the two traditions have proceeded as relatively separate enterprises. On one hand, social network studies seldom have actors with known geographic locations. On the other hand, although geographers employ graphs to represent flows between locations, for the most part they do not incorporate social transactions or social networks into their analyses. Recent advances in GIS’s, with extensive capabilities for storing and analyzing spatial data, provide a means for joining social networks with spatial information and geographic models. Our intent in this paper is to begin to bridge these somewhat separate traditions by incorporating social network information into a GIS, thus spatially referencing the social transactions between actors in the network — in our case, villages in a district in Thailand. This will allow us to use the wealth of spatial information and spatial analysis capabilities available in the GIS to help understand network patterns. We can then model spatial aspects of social and economic networks between villages and relate these networks to information about spatial proximity, topography, land use and other spatial properties. Spatially locating and viewing social network data also allow for an assessment of the quality of the social network data in a manner that would not be available without the utility of the GIS. To our knowledge this is the first example incorporating a rich social network data set with a fully developed GIS.