Five Ways IPUMS-DHS Can Simplify Your Life

Have you ever formatted what you thought were your final models only to discover that:

• The survey question you used for your dependent variable had five rather than four variations across surveys?
• There are two other samples (not in your analysis) in which respondents were asked precisely the question that interests you?
• There is a better question on women’s employment than the one you’re currently using?
• A key question was asked about all daughters under 14 in one country but all daughters under 19 in another?
• The survey skip patterns differ significantly across surveys?

These are among the DHS equivalents of missing the nail and hammering your thumb. Ouch!

Fortunately, with IPUMS-DHS, you can put the metaphorical Band-Aids away. IPUMS-DHS, constructed at the Minnesota Population Center, is a web-based tool for accessing DHS data. It makes error-free comparative analysis (across time or countries) easy. IPUMS-DHS currently covers Africa and Asia and includes 23 countries, 101 samples, and 5000 variables. Why not give it a try?

1) See at a glance which surveys asked certain questions, how, and of whom.

Choose a topic from the drop-down list to see which samples include the groups of questions you want. Click on a variable name to see a comparison across countries. The tabs will guide you to codes and a description (which is especially great for constructed variables, like “Unmet Need”) and a discussion of comparability issues.

2) Compare the frequency of responses to questions and more without downloading a data file.

Clicking on the variable name will also bring up, for every sample, frequencies of responses, an explanation of who was asked the question (called the “Universe”) and an English-language version of the question text.

3) Trust that the same variable name and codes have same substantive content.

While the DHS standard variables simplify researchers’ work, even standard variables (such as V130, RELIGION) may have different responses or varying amounts of detail across samples. Non-standard variables’ names differ widely across DHS samples. IPUMS-DHS gives variables with the same substantive meaning consistent names and codes. This “integration” of the DHS data lets you analyze the data immediately, without investigating and resolving differences across samples.

4) Create a customized data file with multiple samples in minutes, and change it just as quickly.

With IPUMS-DHS, you can create a dataset tailored to your specific needs in a snap. Just log in using your existing DHS Program user ID and password, browse variables and samples, and add the ones you want to your “data cart.” (Despite the analogy, the data are completely free.) Indicate your preferred file format and, a minute or two later, your data will be ready to download, unzip, and analyze.

Did you forget a control variable? Want to add information from an additional sample? No problem. Just return to your data cart, click “Revise” and then “Change,” and you can instantly add or subtract variables and samples, and download the new, revised data file.

We encourage you to check out IPUMS-DHS. It could change your life (or at least your research).

Special thanks to our guest blog contributors, Elizabeth Boyle and Miriam King!

Elizabeth Heger Boyle, is Professor of Sociology & Law at the University of Minnesota. She studies the role of international laws and policies on women and children’s health around the world. She has written extensively on the impetus for and impact of laws related to female genital cutting, including the book Female Genital Cutting: Cultural Conflict in the Global Community. Her current research focuses on abortion policies globally and their effects; this includes a 2015 article in the American Journal of Sociology. Professor Boyle is currently co-Principal Investigator (with Dr. Miriam King) on IPUMS-DHS, a National Institute for Child Health and Development grant that integrates Demographic and Health Surveys over time and across countries to make them more user-friendly for researchers. Professor Boyle has a Ph.D. in Sociology from Stanford University and a J.D. from the University of Iowa.

Miriam L. King is a Senior Research Scientist at the Minnesota Population Center at the University of Minnesota.  She has managed data integration projects on the U.S. Current Population Survey, the U.S. National Health Interview Survey, and, most recently, the Demographic and Health Surveys.  Her research has focused on the history of the U.S. census, data integration methods, U.S. historical fertility differences, living arrangements, and disparities in access to insurance for same-sex couples.  Dr. King has a Ph.D. jointly in Demography and History from the University of Pennsylvania.