Fig 1.
Operational areas of the multi-stakeholder platforms.
Source: [13].
Table 1.
Concepts and measurements in network analysis.
Table 2.
Factors and variables used to explore the changes in multi-stakeholder network characteristics.
Fig 2.
Stakeholders in livelihood and agricultural innovation systems.
Dots represent different stakeholders and the circles surrounding them represent the group of stakeholders operating in multi-stakeholder platform (a), innovation network (b), innovation system (c), and livelihood system (d). MSP targets a sub-group of an innovation network (orange circle) with its events and influences, and is influenced by, the characteristics of that network (blue circle).
Table 3.
Typology of stakeholders in livelihood and innovation systems based on their involvement in interventions with MSPs and the influence of the intervention on livelihood systems.
Table 4.
Differences in MSPs in Burundi, DRC, and Rwanda.
Percentages represent the characteristics of the factors between surveys. DRC received the least funding support, and Rwanda received the most. Types of problems targeted by the MSPs increased in Burundi and DRC and stayed the same in Rwanda. Rwanda has the highest number and highest ratio of innovation-generation, innovation-diffusion, and innovation-use events.
Fig 3.
Maps of multi-stakeholder innovation networks in Burundi, DRC, and Rwanda in t1 (left) and t2 (right). Node size represents the degree centrality. Dark green (upper left) nodes represent organizations based in Burundi, blue (below) represents DRC, light green (upper right) Rwanda, and orange supranational organizations. Dark green coloured ties represent organizational connections in Burundi, blue represents DRC, and light green represents Rwanda. Collaboration in innovation networks was positioned around locally central actors (a) in each country and contained sub-clusters with both national and supranational organizations (b). After the MSP, some sub-clusters (c) left the collaboration.
Table 5.
Changes in the collaboration, knowledge exchange, and influence spread characteristics of multi-stakeholder networks in Burundi, DRC, and Rwanda.
Fig 4.
Knowledge exchange in innovation networks in Burundi, DRC, and Rwanda in t1 (left) and t2 (right). Node size and boldness represent the degree of knowledge exchange centrality. White nodes are parts of innovation networks but not knowledge exchange. An orange tie colour represents connections in Burundi, purple in DRC, and green in Rwanda. During the MSP, all knowledge exchange clusters that were not initially connected to the lead organization (a) left the network. New knowledge exchange connections were generated either by participation of national organizations (b) or by establishing cross-boundary connections (c). Isolated clusters in the initial network (d, e) connected to the main clusters, and some new organizations (f) joined the network.
Fig 5.
Influence spread in innovation networks in Burundi, DRC, and Rwanda in t1 (left) and t2 (right). Node size and boldness represents the degree of influence centrality. White nodes are parts of innovation networks but not influential. An orange tie colour represents connections in Burundi, dark blue in DRC, and green in Rwanda. During the MSP, some existing influence clusters (a) left the networks, some organization (b) joined the influence spread networks, and some existing organizations (c) increased their influence.
Table 6.
Results of logistic regressions explaining the factors that affect multi-stakeholder innovation network configurations.
Initial characteristics of the innovation networks and funding were significant both in term of incumbent stakeholders’ decision to stay and new stakeholders’ decision to join.
Table 7.
Changes in innovation networks, factors influencing the changes, and the implications for scaling innovations following an R4D intervention with MSPs.