Since the 1970s, scientists have suggested that avoiding the worst impacts of climate change will require preventing more than a 2 degree Celsius warming over pre-industrial levels. With the growing likelihood of such an increase, science has turned to questions of adaptation, or how people can reduce the impacts of coming changes. However, recent research shows that the majority of climate adaptation science has relied on a relatively narrow set of disciplinary and methodological approaches. Of more than 450 articles on climate adaptation, smallholder agriculture and climate change published from 2007 to 2015, for example, roughly a third drew exclusively on natural science. Of those that included social science, less than 25% drew on any social scientific techniques outside quantitative farmer surveys. As Deborah Davidson argues, the corollary of these disciplinary and methodological gaps is a weak understanding of how gender, institutions and social norms may shape people’s responses to climate change – potentially one reason why, despite years of investment in adaptation science, there are few reports of systemic transformation in adaptive capacity.
Figure 1: Categorization of 201 journal articles on climate change adaptation in agriculture published in 2015. Source: Davidson, D. Nature Climate Change 6, 433–435 (2016)
This suggests the need to methodologically enrich existing approaches to understanding both generic and specific pathways of adaptation. This post reflects on our research team’s efforts to do this through a National Science Foundation supported study in Nicaragua. As is true of many other studies drawing on social science to understand the linkages between climate change and adaptation, household surveys are a key piece of this research. Over the past five years, households in northern Nicaragua’s coffee-growing regions have been hit by drought and coffee leaf rust, both of which further strain household food and water security and both of which are shaped by increasing hydro-climatic variability in the region. Surveys with ~350 households in the region in both 2014 and 2017 help to understand household experiences and responses over time as well as how experiences in Nicaragua compare to dynamics reported elsewhere. The use of a stratified random sampling approach in survey data collection also helps to isolate the role of institutional affiliation in shaping household responses. However, household surveys are only one piece in a larger data puzzle. Below we reflect on how embedding household surveys within a mixed-methods and participatory research approach can enhance their value and inform more policy-relevant insights.
- Embedding the qualitative in the quantitative (and vice versa): The relative merits of qualitative vs. quantitative data are extensively debated, often overlapping with and confused by ancillary debates over the benefits of different disciplinary approaches (e.g. anthropology vs. economics). While these debates rarely engage the benefits of using qualitative and quantitative insights in tandem, we have found a mixed-methods approach central in enhancing the insights derived from surveys. For example, within our household survey, we asked about household coping responses to hazards in a structured manner best suited for quantitative analysis. However, to make sure these analyses support situationally-appropriate policy responses, we weighted household actions using in-depth interviews and focus group discussions that focused on what constitutes a more vs. less severe response locally (Table 1). We are currently repeating this approach using gender-disaggregated rankings to identify how the perceived severity of different actions may vary between men and women.
Table 1: Reported household coping strategies from 2014 household survey weighted using focus group discussion data. Source: Bacon, C., Sundstrom, Stewart-Frey, Beezer. 2017
- Using survey data in sequence with case study research. In addition to an integrated analysis of qualitative and quantitative data, we are also using both data types sequentially to enhance their value; for example, using survey data to “scale” institutional insights surrounding water governance generated through interviews and participant observation. Early interviews suggested the importance of understanding resource conflict and the role of common property institutions in moderating scarce water resources in the dry season. While understanding conflict and institutional management is arguably best accomplished through in-depth, long-term qualitative research, this approach can be difficult to scale across many different research sites given limits in time and funds. We adopted a hybrid approach: developing an in-depth understanding of water governance dynamics in three case study communities and using these insights to develop general questions on water use rules and practices in the survey distributed to households spanning 51 communities. This approach will also be useful to interpret survey findings. As a next step, we are also piloting a survey for community water committees.
Figure 2. Participatory research process: From collaboration and goal-setting to data analysis and action to create strategic change.
Figure 3. Participatory research process in action: Survey team gets started in northern Nicaragua. Credit: Chris Bacon.
- Embedding survey in a community-based participatory action approach. Household surveys are often performed by a team of “drop-in” enumerators unfamiliar with the research context and racing to complete surveys within a set period of time. In many instances, survey results are also not returned to those people the research is theoretically designed to support. Before drafting each of this project’s surveys, the project team consulted farmers, cooperatives, university researchers and government officials in Nicaragua on their research priorities, key questions, and needs. In some cases, this took the shape of formal presentations on past research and collaborative discussion on future questions relevant to the area. In other cases, this involved asking households what key challenges they were facing and what information they lacked at present.Survey enumerators were hired and trained from the villages in which the research was conducted. These individuals later helped to share information on specific findings to each of the households surveyed as well as their local cooperatives and other members (Figure 4) while the research team organized formal presentations with regional leaders of farmers associations and national universities. These practices increase the likelihood that the research will translate into actionable insights. They also improved the science, for example, ensuring that the questions would be clear, intelligible and meaningful in context. (While components of participatory action research have been mainstreamed into development practice since at least the 1980s, adoption of the full cycle of participatory action research remains rare).
Figure 4. Example of research findings returned to communities following the 2014 household survey. See link for a copy of the full pamphlet shared with surveyed farmers.
- Building on existing survey instruments and sharing data: The goals of this project are complementary to a number of other research efforts related to climate change, land use and household food and water security. Thus, to maximize the value of those household surveys that are conducted, we included questions that would enable a comparison of findings between Nicaragua other regions and countries. In our case, we drew on surveys from CGIAR’s Climate Change, Agriculture and Food Security Research Program, Nicaragua’s Census, and the International Forestry Resources and Institutions Methods, as well as those openly shared through Harvard’s Dataverse and the University of Michigan’s ICPSR.
- Considering what questions are not best addressed in survey research: While this post has reflected on how to enhance the value of household surveys in understanding linkages between climate change, smallholder agriculture and adaptation, it is important to mention that it is always worth considering what research questions and topics are not best addressed by survey research – and what other approaches will be needed to assess these dynamics. For example, because only one household respondent is surveyed, surveys are generally not well suited to capturing gendered differentiation or power relations within the household. Although surveys (often coupled with mapping activities) can offer an effective tool for documenting agricultural land use, they are not adequate for explaining collective histories of conflict and change, or assessing bottom up struggles for access to land, water and food. In response to these types of questions, we rely on ethnographic research, including interviews and focus groups.
Addressing how political economic shocks as well as climatic variability and change impacts agriculture and food security requires social science and the deeper integration of social and natural science approaches as well as insights from local knowledge, engineering, and the humanities. Farmer surveys will continue to yield insights, but they are more useful when linked to community-based participatory action research processes and complemented with an integrated analysis of institutions, gender, biophysical processes, and agrarian change.