By Engels Del Rosario, Erik Thulin, Sania Ashraf, and Philipe Bujold (Rare’s Center for Behavior and the Environment).
Imagine you’re tasked with conducting a baseline assessment in a remote coastal community. How can you ensure your survey sample represents the broader community without reliable local population data or a complete household registry? Conservation researchers often deal with limited and outdated data, especially in low- to middle-income countries where “data divide” is more pronounced 1. Consider our experience in Southern Leyte, Philippines, where we sought to identify and survey fisherfolk aged 18 to 65. We were lucky enough to access a fisher registration database, but it proved to be an inadequate sampling frame as it lacked age information. Later, we learned that the fisher registration database captured less than half of the fisherfolk population in our study site. As a result, we had to explore other sampling strategies, highlighting the need for more adaptable approaches in similar contexts.
Figure 1. Mangroves and Del Carmen’s coastline on Siargao Island, Philippines. Photo credit: Ferdz Decena for Rare.
This situation is not unique. Conservation and development professionals regularly face hurdles when sampling remote communities. Researchers typically rely on traditional systematic sampling methods like random walk or cluster sampling with listing operations (where every eligible household is counted), but these methods come with challenges. Random walk sampling can lead to biased representation as it involves interviewing, for example, every fifth or tenth household, and often oversamples in denser areas. Cluster sampling partially overcomes this bias by randomly selecting clusters like villages or neighborhoods and surveying every household within them. While this method is an improvement over random walk sampling, it requires extensive initial fieldwork to create accurate lists within each cluster, significantly increasing cost and complexity.
Rare’s Center for Behavior & the Environment developed an innovative yet practical alternative: geospatial sampling using freely accessible tools such as Google My Maps.
***
How Geospatial Sampling Works:
Geospatial sampling leverages satellite imagery and online mapping platforms to identify households and create a visual, unbiased sampling frame without extensive preliminary fieldwork and technical expertise.
Step 1: Create a customized map of your research site in Google My Maps (free with any Google account). (Video source: https://youtu.be/gNjOG4dr3lo).
Step 2: Delineate research site boundaries to define your sampling area. Using the polygon drawing tool, ensure that the boundaries you draw accurately reflect those of the regions, districts, or villages of interest. (Video source: https://www.youtube.com/watch?v=lyAU58pwepQ&feature=youtu.be)
Step 3: Place markers on visible buildings and label them accordingly. (Video source: https://www.youtube.com/watch?v=KAh8a4IbGLg&feature=youtu.be)
Step 4: Export the data as a CSV file to create your sampling frame or list of sampling units. (Video source: https://youtu.be/J0YtNJOmUjg)
Step 5: Select your random sample from the list of housing structures, for example, using common software (e.g. Excel, R, Python).
Step 6: Share the map with enumerators using mobile devices equipped with your map.
This method eliminates reliance on incomplete or biased government data and reduces the costly and time-consuming pre-survey field visits.
Field Lessons and Practical Tips
Our experience implementing geospatial sampling in rural coastal communities in the Philippines highlighted several critical lessons.
Population validation matters: Not every dwelling will contain an eligible respondent. To avoid wasted time and effort on ineligible households, conduct limited preliminary field visits to ensure your map reflects all possible target areas and to estimate eligibility rates to inform field plans.
Recognize technology limitations: Google Maps imagery isn’t real-time and requires internet connectivity. Have alternative references for areas with poor connectivity and seek assistance from locals to locate structures obscured by vegetation.
Collect GPS coordinates: Use digital tools like KoboToolbox to facilitate return visits and improve location accuracy.
Benefits of Geospatial Sampling
Geospatial sampling offers significant advantages for conservation social science.
- Accessibility: Requires only basic digital skills and free tools.
- Cost-effectiveness: Reduces expensive preliminary field visits for household listing.
- Adaptability: Works in various ecological and social contexts.
- Enhanced accuracy: Provides better representation than convenience or quota sampling.
Geospatial sampling can bridge the gap between rigorous methodological standards and practical field limitations for conservation practitioners working in remote or challenging environments.
Moving Forward
As conservation increasingly incorporates social science methodologies, researchers need practical, reliable, and cost-effective approaches suited to challenging field conditions. Geospatial sampling offers a middle path between methodological rigor and practicality, enabling scientifically valid surveys without high costs or logistical burdens. Using widely available technology to create robust sampling frames, conservation researchers can produce high-quality, credible social science insights — even with limited data infrastructure.
This blog highlights practical insights from our experience at Rare. For more information about behavior change approaches to environmental challenges, visit behavior.rare.org.
***
About the authors:
Erik Thulin
Erik is a Senior Director of Behavioral Science at Rare Center for Behavior and the Environment. He works on bridging the academic-practitioner gap in environmental behavior change through research collaborations and science communication. He holds a PhD in Psychology from University of Pennsylvania and a B.A. in Cognitive Systems from University of British Columbia. Prior to joining the BE. Center, Erik worked with the Social Norms Group at Penn on behavior change problems like encouraging toilet use in low-income countries.
Sania Ashraf
Sania is a mixed-methods research scientist at Rare, specializing in translating behavioral theory into practical behavior change interventions. She works with Rare’s Fish Forever program supporting their research to maximize adoption of pro-social behaviors. Holding a PhD from Johns Hopkins and an MPH from Emory, her prior work focused on global health, designing and evaluating WASH and norm-centric interventions in low-resource settings.
Philipe Bujold
Philipe is a senior behavioural scientist at Rare’s Center for Behavior & the Environment, working to better understand and address environmental challenges through the lens of human behaviour. He works closely with NGOs and government partners to translate behavioural insights into scalable behaviour-change interventions for climate-smart agriculture projects, regenerative grazing, and sustainable fishing, among others. Philipe holds a PhD in behavioural neuroscience from the University of Cambridge, where he previously studied the neural algorithms that shape decision-making.
Engels Del Rosario
Engels is a research manager at Rare’s Center for Behavior and the Environment, currently focusing on evaluation of Rare’s sustainability fisheries intervention in the Philippines. His prior work includes managing and implementing the monitoring and evaluation systems of national big-ticket programs on poverty reduction and youth workforce development, among others.
- World Bank. 2021. World Development Report 2021: Data for Better Lives. Washington, DC: World Bank. doi:10.1596/978-1-4648-1600-0 ↩︎
