Adaptive Management Helps Researchers Shift Gears During the COVID-19 Pandemic
Best laid plans
In our last post, we shared some of the Phase 1 findings of the U.S. Agency for International Development’s year study exploring how camel leasing in Somalia may impact household resilience. Completed in January 2020, the analysis from our desk review and retrospective focus group discussions deepened our understanding of camel leasing in the Woqooyi Galbeed region of Somaliland and helped us identify key questions to explore in household surveys. We geared up for Phase 2 of our work in March 2020, with an in-person survey of more than 200 households and almost 50 dairies. Our sampling frames were developed, we had received our exemption from the institutional review board and our survey training materials were in the works.
Upended by the pandemic
Then, of course, the COVID-19 pandemic struck. We hit the pause button on our survey plans to protect the safety of our enumerators, participants and researchers until we had enough information to craft a new strategy. Pulling on our adaptive management hats, we worked closely with USAID to devise a new methodology for this study phase; safety came first, and it was time to try a phone-based data collection model. The context of our study area and our participants’ demographics (camel herding pastoralists and dairies in Somaliland), mixed with the challenges of COVID-19, meant that we risked lower response rates due to the potential of non-working phone numbers and less contact with village elders — a critical link to working with pastoralist communities. On the upside, we would be able to more fully ensure that no one was exposed to the virus as a result of this work. Also, moving trainings to a virtual setting led to our U.S.-based research team being able to answer enumerators’ questions in real time.
Learning and adapting
As we implemented our phone-based surveys with pastoralists that July and with dairies that September, we learned many lessons about conducting resilience research via phone-based surveys in rural contexts, and sought creative solutions to overcome challenges from connectivity to quality control.
Methodology: Our original questionnaire was designed to help us understand participants’ experiences with camel leasing, assess their resilience capacities related to camel leasing and establish baseline levels of food security, our main outcome of measurement. Shifting to a phone-based survey meant making hard decisions about what to cut so that the original 45-60 minutes it took to answer would be cut to around 20 minutes. We went back to our camel leasing theory of change and study questions, and focused on our main outcome with enough demographic questions to characterize differences in our study groups. Our experience during data collection showed that even in person, the original survey was likely too long, so our streamlined version provides a strong foundation for more effective data collection in person in the future.
Virtual training: While we had been utilizing Zoom with our data collection partner throughout our study, having each enumerator access Zoom from their homes in Somaliland required several rounds of testing to ensure adequate connectivity and comfort with the platform. Connectivity proved challenging for a few enumerators who ended up socially distancing in our partner’s office to access a more stable internet connection. Moreover, most enumerators were not accustomed to spending long days on video conferences and trainers were unused to a virtual format. Together, we sought opportunities to shorten sections, increase interactions via breakout groups and provide plenty of screen breaks. The U.S.-based research team also shifted their schedules to be present during Somaliland working hours. Virtual training had the added benefit of helping the enumerators prepare for phone-based data by practicing interactions that were not face-to-face.
Conducting phone calls with pastoralist communities: Though our sampling frame included previously validated phone numbers, many potential respondents had inactive numbers or were not in possession of their phones. Fortunately, replacing those respondents with new, randomly selected respondents was easier than it would have been for in-person data collection. During the household survey, enumerators reported facing challenges such as keeping respondents on the phone to complete all sections in full. These challenges were less common with the dairy owners, a different demographic that may be more comfortable conversing via phone with unknown parties. In the future, we will build in additional time for the enumerator teams to reach out to village elders and remind them of the survey to reestablish their credibility.
Despite the challenges, our dairy survey yielded interesting insights for further exploration. For example, some leasing dairies reported leasing several years before the 2016/2017 drought in the area, but the practice seemed to accelerate after the drought ended. We learned that dairies are leasing from both camel-herding pastoralists as well as others who are settled in towns, suggesting that non-pastoralists are seeking out camel leasing as a livelihood strategy. As we look to the remainder of our study, we are shifting back to in-person interviews while using COVID-19 safety protocols. We are also adding new methods to deepen our contextual understanding of camel leasing through a value chain analysis and further focus group discussions.