1. Research framework J.Rusike Research Review & Planning Meeting Africa RISING – West Africa Project 23-25 October 2012, Tamale, Ghana

2. Context Sustainably intensify household food, cash crop and livestock production in FtF target areas (West Africa, ESA, Ethiopian highlands) In line with USAID missions In line with the CRP 1.1 and 1.2 logic Farm-level issues (beyond the plot and field to consumers) Integrate multiple stakeholders Staple foods within major farming systems with links to nutrition and diversification Research backstops FtF investments

3. Purpose Provide pathways out of hunger and poverty for small holder families through sustainably intensified farming systems that sufficiently improve food, nutrition, and income security, particularly for women and children, and conserve or enhance the natural resource base Research contributes to development impacts of FtF

4. Objectives: Research Identify & evaluate demand-driven options for sustainable intensification that contribute to rural poverty alleviation, improved nutrition and equity and ecosystem stability Evaluate, document & share experiences with approaches for delivering and integrating innovation for sustainable intensification in a way that will promote their uptake beyond the Africa RISING action research sites

5. Objectives: Development Create opportunities for smallholders (within Africa RISING action research sites) to move out of poverty and improve their nutritional status – especially of young children and mothers – while maintaining or improving ecosystem stability Facilitate partner-led dissemination of integrated innovations for sustainable intensification beyond the Africa RISING action research sites

6. Outcomes: Research Integrated innovations increase production & / or improve productivity in a sustainable manner for targeted households at Africa RISING research sites Aggregated impact of these farming practices at household level contributes to an improved understanding of ecosystem stability at the landscape level Dissemination of integrated innovations for SI leads to impacts beyond the Africa RISING action research sites

7. Outcomes: Development Wider adoption of innovations identified and tested by the program’s outputs within the Africa RISING action research sites enhances livelihoods through increased agricultural output, income diversity, reduced vulnerability to adverse environmental and economic challenges and improved nutrition and welfare; especially of young children and mothers Development community initiates programs, based on the knowledge tools and innovations developed and promoted by Africa RISING, that are directed at developmental goals that are consistent with the Africa RISING program purpose

9. Research design: hypotheses INTEGRATION: Integrating technological components into SI systems confers more benefits to small holder farmers than single components ADOPTION: Integrating technological components into SI systems stimulates adoption compared to single components TRADE-OFF: Offering interventions tailored to the context specific conditions lowers environmental damage SEQUENCING: Right sequencing of integration of component technologies provides SI pathways SCALABILITY: Agricultural SI interventions tailored to local context specific conditions are scalable to other settings

10. Research outputs Situation analysis and program synthesis Integrated Systems Improvement Scaling and delivery of integrated innovation M & E (Program-wide synthesis and co-learning)

11. Output: Situation analysis Determine development domains (agro-ecological potential, market access, and population density) Prioritize target areas (welfare, sustainability, farming systems, degradation, governments’ & USAID priorities) Develop farm household typologies Identify entry points for pathways Inventory of innovations Ex-ante potential of innovations Priority setting and planning for integrated systems improvement Program-wide synthesis and co-learning

12. Output: Integrated systems Identify research teams within R4D platforms to lead innovation activities related to system improvement Identify modeling/decision support tools for ex-ante technology identification, trade-off analysis, evaluation of the ex-ante sustainability and resilience of options, and guiding future research Participatory evaluation and adaptation of appropriate combinations of technologies and interventions Address new research challenges and opportunities emerging from the activities

13. Output: Scaling Assess scalability of integrated innovations (meta-analysis of options) Identify/develop scaling approaches for targeted integrated innovations Pilot test scaling approaches from action sites within project area Develop costed templates for scaling by development investors Evaluate aggregated impact of household level interventions at landscape scale and beyond

14. Output: M & E Sakana’s presentation

15. Methods On-station/on-farm, FSR, Participatory technology evaluation and adaptation, PTD, FPR, FFS Pilot technology on small test plots allocated randomly to treatment and controls (on-station/on-farm) to estimate gains of technology Farmers participating not representative of adopters (placement/selection biases) Researchers/Extension agents/NGO staff induce different behaviour from actual adopters Move away from on-farm trials to design that allows farmers to select into adoption

16. Methods: Randomized experiments (or natural) Natural or randomized experiments village, community is unit of randomization/individuals unit of observation Many villages in geographically distinct areas=>scalability Withhold treatment from some villages but data still collected…

17. Methods: Platforms Gundaa Trader Wumpini agrochemicals USAID ADVANCE 2,686 farmer outgrowers ADVANCE STANBIC BANK AGRISERVE Industrial maize buyer Accra Premier Foods – Maize-soy processor Kumasi CANDEL Agrochemical importer Business plans & record keeping Ghana Grains Council Credit Sales Services & inputs (ploughing, seed, fertilizer, thresher) Sales Seed supply supplier Training on GAPs

22. Methods: Roll out over time Match villages in pairs based on observables (using as many characteristics as possible) Randomize Treatment and Control within each pair Randomly choose say 5 pairs of villages Stratify the sample to obtain better randomization results Rollout across villages can start from the most favorable to the least favorable pair without imposing a bias on measured impacts

23. Thank you!