Authors:
Tom Snijders | University of Groningen; University of Oxford | Netherlands
Malick Faye | Rhine-Waal University of Applied Sciences; SciencesPo, Paris | Germany
Julien Brailly | Swinburne University of Technology, Melbourne; SciencesPo, Paris | France
This study considers a group of seven villages in rural North-Western Senegal. The villages have a common water supply which is governed by the villagers themselves by means of a managing board. The inhabitants of the villages are from two ethnic groups, a distinction corresponding – although incompletely – to the difference between sedentary crop farmers and nomadic cattle breeders. Ethnicity and the associated different interests of crop farmers and cattle breeders constitute an important underlying division for the governance of the water supply. The combined sociological research questions for this presentation are about how ethnicity and the crop farming – cattle breeding contrast is related to sociability and social support; how these social configurations (sociability, ethnicity, type of agriculture) are associated with the selection of board members; and how these social configurations influence the patterns of advice between the board members.
The network structure of this data set has, next to being multivariate (sociability and advice), two kinds of complexity. First, the actor set has a nested structure, with individual inhabitants in houses in villages; there are a total of 7 villages. Second, the board members are a subset of the inhabitants, chosen by the inhabitants and therefore endogenous; the sociability relation is observed between all inhabitants, and the board-related advice relation is observed between board members. This is summarized only inadequately by the heading of ‘multilevel network analysis’. The methodological research question for this presentation is how to specify this network data structure for a longitudinal analysis using a stochastic actor-oriented model implemented in RSiena. s
The structure chosen for the Siena analysis is a multivariate dependent network, with separate within-village sociability, between-village sociability, and board advice relations; and with board membership as a dependent actor variable. The structure is further specified by blocks of structural zeros to distinguish the within-village and between village networks, and by special effects restricting the board advice relation to the set of board members.
A first set of specific questions are concerned with homophily (concerning relevant individual attributes) of the sociability relation, and how this homophily differs between within-village and between-village ties, and between cattle farmers (Fulani) and crop farmers (Wolof). A second set of questions are about the ways in which the power basis of the board members, as reflected in the membership itself and in the board advice relation, depends on their attributes and on their social capital.