These data were collected in Chennai, India, and Machakos, Kenya, using two-day workshops in each location. The aim of the workshops was to explore perspectives on pathways out of rural poverty, in ways that do justice to the complexity and uncertainty inherent to such pathways. Using the interactive appraisal method called Multicriteria Mapping (MCM, http://www.multicriteriamapping.com/about), the workshops afforded the participants to foreground pathways (out of poverty) that they consider to be relatively more effective than others.Global poverty looks radically different in the 21st century as climate-related events, political-religious conflicts and economic growth-inequality nexuses add to persistent forms of social exclusion based on gender, race, and class. In this uncertain and unpredictable context, we require new approaches to understand complex pathways into and out of poverty, directing attention to poor people's collective capacity to bring about transformative change i.e., their agency, as constituted by social networks and relations with nature, and mediated by science and technology. Our aim is to develop the concepts and methods of an innovative 'relational agency pathways approach', drawing on theories from Science, Technology and Society studies and the 'pathways approach' to poverty reduction and social justice, which emphasise interactions between social, technological and environmental change. We will develop this new approach to understand diverse pathways out of poverty for smallholders and the landless in agriculture, in two arenas. First, we will study how small farmers and farmworkers adapt new technologies on the farm, as their cultivation practices are transformed due to technological and environmental change. Second, we will study how farmers turn a harvested crop into a commodity for the market, negotiating their relationships with credit providers and traders. Both these arenas played out dramatically under the 'Green Revolution', from the 1960s onwards, when technology, markets and government support were used to intensify agricultural production. The first geographical focus of our work will be on the North Arcot region of Tamil Nadu, India, a classic exemplar of the Green Revolution in Asia, where extensive historical data since the early 1970s are available. Collaborating with our co-investigators at Madras Institute of Development Studies, and collecting new life history data in the field, we will map long-term agency pathways into and out of poverty constituted by changing technologies, natural resources and social worlds, as lived by people of different genders, classes and castes. We will test the approach in Machakos County in Kenya (in collaboration with our co-investigator at African Centre for Technology Studies), where several attempts have been made to get a Green Revolution off the ground, but none have been sustainable. In addition to relying on archival data and collecting life histories using ethnographic engagement with the study's participants, we will use a workshop format to collect data on how people evaluate diverse pathways out of (and into) poverty along a range of criteria derived from conventional indicators of welfare and well-being as well as those designed by the participants themselves. To communicate our approach in other low-income contexts, we will develop a training programme for junior researchers. There will be broad-based participation from researchers, policymakers and farmers throughout the project, and we will organise a final workshop in Kenya, which will bring these participants together in a safe space for collective learning, where our findings and approach can be confronted with their different knowledges and experiences. We will present our work in academic and policy forums, produce policy briefs and web blogs and a short documentary film (to engage with audiences beyond academia and policy). We see our research to be of interest to at least five groups: a) government institutions attempting to intensify smallholder agriculture through better use of natural resources and new technologies; b) rural development organisations (including non-governmental ones), active in organising initiatives for poverty alleviation; c) academic researchers working on agricultural sustainability and poverty issues in the global south; d) environmental NGOs at international and grassroots levels; e) farmers' associations such as the East African Farmers' Federation.

As noted above, we used MCM for data collection. A plural and hybrid appraisal method, MCM allows the combination of quantitative rigour with qualitative flexibility, to enable the linking of less tangible (subjective, narrative) factors in poverty reduction, with more visible and dominant economic data (e.g. on incomes, wages, and changes in crop yields). Through MCM, we attempted to encompass the multiplicity of participants’ knowledges and values, manifesting in their appraisals of diverse pathways out of poverty. The aim was not to reach a consensus but rather to give equal attention to contrasting perspectives and assumptions. Thus, key disagreements about why particular pathways were favoured over others, were considered as important as seeking common grounds for policy-making.