Social Network Patterns of Sharing Information on Land Use and Agricultural Innovations in Ethnically Heterogeneous Communities in Ecuador

A case study in South Ecuador serves as an example to understand the dynamics of adoption of agroforestry species. Agroforestry species become important in land that is totally devoid of forest cover. They can reduce soil degradation, increase local biodiversity and at the same time increase household income. Their use can be seen as a new practice that spreads throughout a region by means of social networks. Sharing information about this potentially sustainable agricultural innovation has become important in a region like southern Ecuador. In this region there are mainly two ethnic groups, namely the Saraguro and the Mestizo-colonos. In the first group are indigenous people and the latter are migrants from other places of the country that went to settle the region during the agrarian reform. The two ethnic groups do have different traditions and land use practices. Qualitative research shows that there are potential differences in adoption of agroforestry tree species between two ethnic groups. The low adoption rate of Saraguro communal leaders may be an indicator of lower contagion than Mestizo-colonos. Contagion refers to actor i adopting the same innovation as actor j with whom actor i had contacted. It is proposed a heterogeneous diffusion model that addresses network exposure effects and a generalized blockmodel for relational data analysis. It was first hypothesized that the Saraguro indigenous group may have lower access to the information necessary for the adoption of the innovation than Mestizo- Colonos. Nonetheless, the results show that Mestizo-colonos have higher adoption rate than Saraguros. The network actors’ behaviour is shaped by local characteristics, for instance their ethnicity. For the analysis it was taken into account the search for advice for the main agricultural activities for the household. On this base, it is proved that indeed there are structural differences on search for advice between ethnic groups. By analyzing ethnic subgroups, the subgroup of Mestizo-colonos has the actors with highest degree of centrality and betweenness. These actors are in the highest cohesive block in cohesive block analysis. Furthermore, a high Burt’s constraint shows that low access to structural holes in the network create a differentiation of information sharing between groups. Higher Burt’s constraint indicates less access to the information located in other clusters in the network. The periphery of the network plays an important role as a source of advice, as shown by cohesive bloc analysis, which we complement with a k-core analysis. The k-core analysis suggests that different and more diverse information is accessed by Mestizo-colonos as compared to Saraguros. In applied terms this suggests that Saraguros tend to share more similar information within their group than do Mestizo-colonos. In order to understand patterns of network structure the Exponential Random Graph Models (ERGM) become important. For example, ERGMs help to valuate the importance of reciprocity in the network relations depending on how many reciprocal dyads are found in the observed network. Fitting an exponential random graph model (ERGM) to the network, it is shown that indeed there are differences between two ethnic groups in the way they share information. We explore the network patterns in the sharing on agroforestry species. Sharing information with similar others is limited to the less commercial tree species. The most commercial tree species information is concentrated in one actor, who is Mestizo-colono. Transitive network effects are identified within ethnic groups as long as they share information on different tree species (others than pine). The two ethnic groups share information of different tree species. This shows a different knowledge, where Mestizo-colonos are more related with commercial species.