Census Tables-to-DHS Geography Crosswalks: IHGIS aggregate data

The IPUMS IHGIS and IPUMS DHS teams have created a series of crosswalk variables between DHS GPS sample clusters and IHGIS subnational geography units from census tables. These crosswalk variables enable IPUMS DHS users to attach aggregate population and agricultural census data as contextual variables to DHS households and individuals.

For example, a user might download aggregate tables on educational attainment by region from IHGIS, calculate the percentage of the population having secondary education or higher, merge that result onto their IPUMS DHS extract using the crosswalk variable as the linking key, and assess the effect of local educational attainment net of the respondent's own characteristics.

Computation

IHGIS-to-DHS geography crosswalks are based on a common methodology using Esri's ArcGIS software suite. A 2- to 5-kilometer buffer is drawn around each GPS cluster point to account for cluster displacement (a confidentiality measure employed by The DHS Program). A 2-kilometer buffer is used around the urban clusters, and a 5-kilometer buffer is used around the rural clusters. 10-kilometer buffers are deliberately not used in this analysis because this level of displacement affects only 1% of rural clusters.

Next, the buffers are overlaid on IHGIS geographic units. If the entire buffer is contained within a single unit, we assign a placement probability index of 100%. If the buffer crosses boundaries of two or more geographic units, we list the code for the unit with the highest placement probability index along with its index value (i.e., the probability that the cluster falls within that unit). IHGIS tabulation geographies with smaller units (e.g., districts) have lower placement probability index percentages, on average, compared to tabulation geographies with larger units (e.g., provinces).

According to DHS displacement procedures, cluster points are not displaced across certain administrative boundaries. For most DHS samples, cluster points are not displaced across first-level administrative unit boundaries; for some samples, cluster points are not displaced across second-level administrative unit boundaries. When calculating placement probability indices, portions of the buffer that fall outside the first- or second-level administrative unit in which the point is located are disregarded.

For further details on GPS clusters, see the documentation on cluster displacement. Users interested in performing spatial analysis with GPS cluster datasets may obtain DHS cluster shapefiles from the DHS Program website or from IPUMS DHS if they have obtained permission to download GPS cluster data from countries of interest. See the variable GPSLATLONG for more details.

Using census-to-DHS crosswalk variables for research

Researchers intending to utilize IHGIS geography crosswalks for contextual analysis should include the crosswalk variable for the corresponding census dataset in their customized IPUMS DHS extract. DHS data, however, should NOT be summarized using geographic units other than the geographic units provided in public use DHS files (typically only first-level administrative units). The DHS sample designs do not support analyses at geographic levels more specific than first level units.

Downloading census-to-DHS crosswalks (extract system)

Researchers may include the IHGIS-DHS crosswalk variables as part of their customized data file (extract) and use the codes to link to geographic units from IHGIS for all first-level geographic variables. Researchers may also use the codes to link to second-level geographic units for countries in which the second level units meet a minimum threshold of 20,000 total population for that year.
In the near future, we will be enabling a feature to allow users who have received additional authorization from the DHS Program to access the GPS data associated with specific DHS samples to download varibles for geographic levels finer than first level that do not meet the minimum population threshold. For the time being, these variables and crosswalks are not available.
To request permission to download GPS data from IPUMS DHS, follow the instructions here to apply directly with the DHS Program. In the box describing your research on the application page, make it clear why you need the low-level geography in GPS cluster points to do your analysis. For example, you might note that you intend to work with IPUMS IHGIS variables that need approved access to GPS data for your samples of interest.

Downloading census-to-DHS crosswalks (CSV files)

The table below lists the downloadable crosswalks between the DHS sample and IHGIS geography codes. Each crosswalk file contains the DHSID codes for a DHS sample and corresponding IHGIS codes you can use to link to a tabulation geography from an IHGIS dataset. If the buffer crosses boundaries of two or more geographic units, we list each unit along with its placement probability index (i.e., the probability that the cluster falls within that unit). The placement probabilities across all units for each cluster point sum to 1.

This icon indicates that login and country GPS authorization are required to download the crosswalk.

Note that the years used to designate a DHS survey sample may differ between the IPUMS DHS sample name and the crosswalk naming convention on this page. Some DHS samples were collected across two Gregorian calendar years, such as Malawi 2015-2016. In some instances, the DHS Program chose one year to name its DHSID identifiers while IPUMS DHS chose the other year. A similar difference may occur in the 2-character code for countries. For example, DHS Program uses "MD" to signify Madagascar, but IPUMS DHS uses "MG". In the interest of clarity because the following crosswalks use DHSID identifiers, the crosswalks are named according to the DHS Program sample names and not the IPUMS DHS names.

Back to Top