Methodology -Collecting Geographic Data

Geographic data in the DHS surveys

Two types of geographic information are included in the DHS and AIS surveys. All survey data is presented both nationally and by sub-national reporting area. These reporting areas are often, but not always, provinces or groups of provinces, and are included in all recoded data files.

In most recent DHS surveys, the location of communities that participated in the survey is georeferenced using a GPS receiver. Data on the location of these communities, or clusters, is released as a separate DHS data file. More detailed information on the survey cluster coordinate data is given below.

Cluster coordinate data: sources and data precision

Survey cluster coordinates are collected in the field using GPS receivers, usually during the survey sample listing process. The collected coordinates are always checked for accuracy before they are released to the public. In cases where a GPS reading is missing, the location of the center of the survey cluster is approximated using a gazetteer or a map.  The source of the coordinates (GPS, map, gazetteer) is reported in the geographic data file that is released to the public.

The precision of the data is also affected by the quality of the GPS units and the environmental conditions at the cluster location. In general, the coordinates reported in most surveys without HIV testing are accurate to no less than 15-20 meters. Nonetheless, the GPS coordinates are often more accurate than most publicly-available administrative boundary data, and may not necessarily fall within the boundaries of the appropriate administrative unit.

Finally, many recent DHS and AIS surveys collect blood samples to test for HIV infection. Because of confidentiality issues related to HIV status, additional precautions are taken to ensure that it is impossible to identify specific communities in surveys that test for HIV. Up to 2 kilometers of random error in any direction is added to cluster locations in urban areas, and up to 5 kilometers of random error is added to cluster locations in rural areas. Additionally, one point in each survey with HIV testing is displaced up to 12 km in any direction.

Guidance on how to use Geographic Information

• Incorporating Geographic Information Into MEASURE Surveys: A Field Guide To GPS Data Collection (PDF, 1.1 MB)
• Cluster Waypoint Database Structure (DOC, 28 KB)