Geospatial approaches to sampling frame development for household surveys: A systematic review of methods, limitations and future directions
This systematic review evaluates how geospatial technologies are transforming household survey sampling frames in low- and middle-income countries to address outdated census limitations, identifying four key methodological approaches while highlighting persistent challenges in validation and implementation that require further comparative evidence and standardized protocols.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer
Imagine you are trying to take a perfect photo of a bustling city to understand how its people live. But instead of a camera, you have a list of addresses to visit. This list is your "sampling frame." If the list is outdated—missing new apartment blocks, ignoring a neighborhood that grew overnight, or including houses that burned down years ago—your photo will be blurry and unfair. You might accidentally ignore the people living in the new slums or the nomads who move around, giving a distorted picture of reality. This is the core problem in the world of household surveys, which are the giant data-gathering missions governments and organizations use to figure out how many people are hungry, sick, or poor. To get a good answer, you need a map that matches the world right now, not the world from ten years ago.
Now, imagine that old map is like a paper blueprint that hasn't been updated since the 1990s. In many places, the cities have changed so much that the blueprint is useless. That's where "geospatial approaches" come in. Think of these as high-tech, satellite-powered magic glasses. Instead of relying on a dusty paper list, these tools use satellites to see buildings from space, count them with computers, and draw new, fresh lines on a digital map. This paper is a massive detective story that reviews how scientists are swapping out those old, crumbling paper maps for these shiny, satellite-updated digital ones to make sure no one gets left out of the count.
The Detective Work: What the Paper Found
The authors of this study, Nazim Gashi, Andy Tatem, and Sarchil Qader, decided to play detective. They didn't just look at one city; they hunted down 1,206 research papers and narrowed them down to the 45 best ones that showed how people are actually using these space-tech tools to build survey lists. They also looked at the real-world stats of how often countries update their official population counts (censuses) and how old the lists are when they finally get around to doing a survey.
Here is the big reveal: The old way is struggling. The paper found that in many places, the "master lists" used for surveys are like expired milk. In the sixth round of a major global survey program called MICS, nearly half (43%) of the surveys were using lists that were older than five years. In some regions, like West and Central Africa, the lists were often a decade or more out of date. It's like trying to navigate a city using a map from before the internet existed; the roads are there, but the new skyscrapers and shortcuts are missing.
The paper identifies four main ways people are trying to fix this broken map problem:
- The Old School (Census-Based): This is the traditional method. You wait for the government to count everyone, draw lines around neighborhoods (called Enumeration Areas), and use that list. The problem? By the time the survey happens, the neighborhood might have doubled in size, or a war might have moved everyone. The paper shows this method is still the most common, but it's often the least accurate because the world moves faster than the census.
- The Grid Game (Gridded Populations): Imagine taking a giant checkerboard and laying it over the whole country. Instead of using neighborhood names, you just count how many people are in each square of the grid based on computer models. It's great because it covers the whole map, even the places with no official names. But the paper warns that if the computer guesses wrong about how many people live in a square, your whole survey is off. Also, a square grid might cut right through a river or a road, making it a nightmare for the survey team to walk through in real life.
- The Satellite Snap (Satellite-Based Mapping): This is like using a drone to take a picture of a town and counting the roofs. It's super fast and can find hidden villages or refugee camps that no one knew existed. The paper highlights that this is a game-changer for finding people in conflict zones or informal settlements. However, it has its own glitches: clouds can hide the ground, trees can block the view, and sometimes the computer mistakes a shed for a house.
- The Digital Architect (GIS-Based Delineation): This is the "best of both worlds" approach the paper seems to like the most. It uses the satellite photos to see where the buildings actually are, then uses smart computer software to draw new neighborhood lines that follow real-world features like roads and rivers. It creates a list that is both up-to-date and easy for survey teams to navigate. The paper suggests this method produces more balanced groups of people and saves time, but it requires a lot of technical skill and good data to work.
The Catch: It's Not a Magic Wand
The paper is very careful not to say that technology has "solved" everything. While these geospatial tools are powerful, they aren't perfect. The authors point out that we don't have a standard rulebook yet for how to check if these new digital maps are actually better than the old ones. Sometimes, the high-tech maps miss the smallest, poorest houses because they are hard to see from space. Sometimes, the computer models make mistakes that we can't see until we get to the field.
The study concludes that while the old census-based lists are still the "gold standard" if they are fresh and complete, they are often too old to be useful. The future, the paper suggests, lies in mixing methods: using satellite images to update old lists, or using the digital architect approach to build new lists from scratch. But to make this work, countries need to invest in training, better technology, and a way to keep these digital maps constantly updated. Until then, the map of our world is still a work in progress, and getting the count right remains one of the hardest puzzles in science.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.