Measuring social networks in three Zambian villages

Abstract

The crucial role played by social context and social networks in exchange behaviour, and the social-embeddedness of resources in rural sub-Saharan Africa are well documented. Social-embeddedness means more than belonging to and participating in groups. This presents a challenge in formalising socially embedded market transactions within standard economics frameworks. Different social networks measures are constructed using data from three Zambian villages and analysed to demonstrate that network properties differ significantly for different people. The way we measure networks matters. Such fine-grained measures will help future work in analysing the role social networks and social-embeddedness play in people’s lives and decision-making.

Keywords:

• social networks
• measurement
• rural development
• mixed methods
• social-embeddedness

Acknowledgements

The author would like to thank Howard White, Julie Litchfield and Elizabeth Harrison (University of Sussex) for research support and guidance; Mr William Shimumbi, Mr Freddie Mpundu and staff at the Misamfu Regional Research Centre, Kasama, for ease of access to the communities, invaluable logistical support and insights; The RuralNet Associates (Lusaka) survey team: Mwila Mulumbi, Kelvin Munjile, Jessie Mukwaila, Fison Mujenje, Stuart Nsana, Geoffrey Siulemba, Muyaka Kampango, Mupeta Sokontwe, Dennis Chiwele, Stephen Tembo, Davison MendaMenda, Ngosa Chisupa and Rhoda Chiwele.

Notes

1. The data for this study were collected under the DFID ESCOR-funded project Linking Growth to Poverty Reduction through Rural Labour Markets, 2001–2004 (see White, Leavy and Seshamani 2005; White et al. 2005).

2. ‘Village’ is defined as the population under an individual headman.

3. Lufubu Resettlement Scheme began as a Rural Reconstruction Centre, established under the 1975 Zambian Rural Reconstruction Programme (RRP). This aimed to provide employment to unemployed youth through communal farming, forming part of National Service. Lufubu was subsequently absorbed into a later incarnation of the Government of Zambia’s resettlement programme in 1990.

4. It was never intended to obtain a sample representative of Zambia as a whole, rather to focus on ensuring the sample was representative of the population of smallholder farmers. Households were randomly sampled within villages in the case of Ngulula and Kabila, and full enumeration was attempted in Lufubu RS. The sample of households can thus be said to be representative of the village, or type of village (‘good’, ‘intermediate’, ‘poor’ market access), from which they are drawn.

5. Endogeneity may arise because of reverse causality between the social networks variables and other characteristics of respondents. For example, involvement in group labour may lead to larger networks but larger networks may also lead to more involvement in group labour, either through the information (for example, job search and screening) effects of networks or through greater obligation towards networks partners to participate in reciprocal group labour.

6. This is in line with methodological work in social capital measurement using data from The Netherlands, which finds different dimensions of social capital are covered by different measures: ‘the extensity of a network, the variety of persons and their attributes in a social network, and their resources all concern different phenomena’ (Van der Gaag and Snijders 2003b, 18). Furthermore, ‘different aspects of social capital indicators show up as significant predictors for different returns to social capital’ (Van der Gaag and Snijders 2003b, 18), which has implication for further research and analysis.