Between November and December 2019, Nigeria’s National Bureau of Statistics (NBS) and National Population Commission (NPopC) became the first national statistical and population agencies to be trained on the gridded population survey sampling methodology for household survey design and implementation, as they prepared for the first national gridded population survey pilot study, currently taking place in Kaduna state until March 2020.

Household survey results provide critical data for policy makers as they collect information on the characteristics and wellbeing of their population. In low- and middle-income countries, household surveys are an important source of detailed demographic and socioeconomic statistics. These surveys are often large undertakings that require complex logistical efforts. To complement traditional household survey techniques, GRID3 is testing the use of a new sampling approach referred to as gridded population survey sampling methodology, developed via GridSample, In this methodology, gridded population estimates are utilised to design the sampling frame and delineate enumeration areas, in comparison to traditional methods that involve the application of census data.

The first training workshop was held from 21-22 November 2019 and provided participants with a basic understanding of gridded population datasets and gridded population survey options. Altogether, 50 participants attended, among them government and non-government agencies including the Ministry of Budget and National Planning, Ministry of Health, NPopC, NBS, the Kaduna State Bureau of Statistics, and National Space Research and Development Agency (NASRDA).

The second training, which was held from 25-27 November 2019, equipped survey supervisors and field researchers from NBS and NPopC with the skills required to conduct gridded population survey fieldwork by putting into practice learning in a real environment.

As part of fieldwork planning, participants utilised OpenStreetMap to prepare their field maps. During this stage, they examined existing building and road data within their enumeration area on the platform. Using satellite imagery, they identified any missing features (e.g. delineated missing building footprints and roads) and labelled them accordingly. Maps for each enumeration area were then printed and utilised in the field in order for  participants to easily recognise their enumeration area and identify their location on the ground. Participants accessed satellite imagery through tablets via various mobile applications including maps.me, OsmAnd, SW Maps, and Openmapkit for navigational purposes. While in the field, surveyors ground-truthed their paper-based maps and made necessary adjustments (such as drawing missing building footprints or roads, deleting or adjusting locations). Participants also used tablets to list households in the targeted areas on a digital platform called Open Data Kit.

After the fieldwork, participants updated OpenStreetMap data based on the new information they collected on the ground.

Overall, the purpose of this training was to enable NBS and NPopC to better understand the  gridded population sampling methodology, implement mapping and listing workflows, and ultimately apply these techniques for the Gridded SMART Nutrition Survey pilot, which is currently taking place in Kaduna State until March 2020.

Various organisations are conducting surveys across the world using gridded population estimates. However, their accuracy has not yet been measured to date. Between February and March 2020, GRID3 Nigeria is carrying out a Gridded SMART Nutrition Survey pilot study in Kaduna State, with the aim to compare the results with those from the Nigeria 2020 SMART nutrition survey, which is happening concurrently and led by the Nigerian Ministry of Health. A SMART nutrition survey is a standard approach for household surveys that utilises census data for its sampling frame.

Following this pilot study, GRID3 and its partners will compare the results of the two surveys to examine the accuracy and feasibility of gridded population surveys.