SNA review: A summary
(last update March 2013)
weADAPT authors reviewed several recent papers on quantitative social network analysis (SNA) in relation to adaptive governance, adaptive capacity and resilience. We wanted to get an insight into the potentials of the method for our work, and to review the lessons learned from other applications of SNA in this area. The following table summarises these works:
Reference (Tools used) |
Actors and their attributes | Links and link attributes |
Network boundary conditions |
Analysis methods (SNA and other) |
Stein et al., 2011 (UCINET and NetDraw) |
Water resources institutional spokespersons (direct and indirect) | Long term funding, information and collaboration, reciprocated links | Determined by recall list, threshold number of mentions as inclusion criterion | Multi-scale analysis |
Mikhail et al., 2010 (ORA) |
Agricultural land managers | Who they are talking to about agriculture | Snowballing, partial dataset | Comparative analysis, statistical testing |
Crona et al., 2010 (Not specified) |
Fishers, fish dealers and gear owners | Lending and knowledge sharing | Members of rural fishing community defined by location | Centrality measures, blockmodelling |
Downey 2010 (UCINET) |
Agricultural labourers | Demand for labour, reciprocated and unreciprocated ties | Farmer or labourers located in 5 villages | Comparison of villages and evolution of social system |
Sandstrom et al., 2010 (UCINET6 and NetDraw) |
Managers involved in rule-making | Communication | Determined by recall list and snowballing | Qualitative interview followed by survey. Comparative analysis: network heterogeneity and closure |
Cassidy and Barnes, 2012 (UCINET6 and NetDraw) |
Household heads or spokespersons in Habu village |
Labour, food, money and information exchanges among households during shocks |
Household location | Centrality and connectivity of networks; correlation with other indicators of social capital |
Garcia-Amado etal. 2012 (UCINET, NetDraw and SPSS) |
Household heads, tenure status and income group | Productive relations in coffee, palm, authorities and ecotourism | Households located in Sierra Morena | Centralisation and hierarchy, clustering, statistical correlations |
Luthe et al. 2012 (Visone and Gephi) | Stakeholders involved in tourism governance | Professional contacts | Stakeholders identified by marketing organisations | Network density and centrality; modularity |
Summary
We found a great deal of variety, and creativity in these applications of SNA with people finding new ways to circumvent conventional limitations of the method. As can be seen in the table above many different types of actors and linkages have been considered. While study design decisions (such as the drawing of network boundaries) varies depending on research objectives, the free recall and snowballing techniques are commonly used. Network centrality measures as well as degree centrality indices (five out of five papers) are used, but these are combined in interesting ways with a variety of other measures, structural and non-structural. Amongst the alternatives mentioned, UCINET is the most commonly used software program for data analysis (three studies).
Background
Three of the papers included in this collection formed part of the Special Feature of E&S edited by Crona and Hubacek in December 2010. This SF caught our attention because of its relevance to our work and so we selected what looked like an interesting subset of the seven papers published here. We also added two papers that were coauthored by colleagues at SEI that also looked interesting and relevant, and a number of other papers that investigate resilience concepts.
Here below follows a brief background on SNA and our understanding of its relationship to other methods – further information on weADAPT can be found here. In the next section we discuss the five publications detailing how SNA has been applied and we then add some concluding remarks.
SNA provides a great variety of quantitative measures/indicators to help describe the overall relational structure of a social system, as well as the roles of individuals within it. It can be correlated with other types of analysis (e.g an analysis of adaptive capacity) which, while it is generally not possible to establish causal relationships, can support an interpretation of the aspect of interest of the system.
There is a formal mathematical language (graph theory) and a sociological framework which is the basis for (social) interpretations for different kinds of networks. Various algorithms have been developed for visualising the networks which can also provide good insights, and these are presented in a similar way as social network mapping (SNM) techniques. Differences between SNA and SNM are that SNA graphs are ‘whole’ networks rather than ego-centric networks based on the perception of (usually) just one actor. They are also much more comprehensive (have more nodes and links) and can be quantitatively analysed with SNA software while standard statistical tests can also be run.
One of the interesting things about network approaches is that they capture the relationships that the respondents themselves recognise, and because of this there is also the possibility to obtain information on informal links in addition to formally sanctioned ones. Other interesting concepts are vertically and horizontally integrated networks and the multiple scales involved in social relations. In combination with other methods SNA can be very useful in understanding complex societies.
Limitations are that SNA offers a static view of the network and is not able to show dynamic aspects of social structure. Moreover, it requires a heavy investment of time and resources into designing and carrying out a study.