AidData-Afrobarometer partnership produces geocoded public-attitude survey data for 37 African countries
The new dataset represents the most precise and consistent geocoding of Afrobarometer's surveys to date.
For more than a decade, Afrobarometer and AidData have independently provided policymakers, practitioners, and researchers with granular data about what is happening on the ground in Africa. Two years ago, our organizations recognized that a broad array of new insights would be discoverable if we could combine our data on development investments with data on local governance, public service delivery, and socio-economic conditions in African towns and villages.
Today, AidData and Afrobarometer are pleased to announce the first fruit of this partnership: a subnationally geocoded dataset of all six rounds of Afrobarometer’s surveys in 37 African countries between 1999 and 2015. With generous support from the U.S. Global Development Lab, AidData spent more than 16,000 hours mapping Afrobarometer survey responses to specific villages and towns within these 37 countries.
This dataset provides hyperlocal, time-varying information about the priorities, preferences, experiences, and opinions of more than 200,000 African citizens in 28,000 localities. More specifically, it provides village- and town-level data on:
- the most important problems that citizens would like to see their governments address;
- the perceived performance of local and national institutions in solving specific problems, such as the provision of basic education and health services;
- access to information via newspapers, radios, televisions, and mobile phones;
- levels of civic engagement and political participation;
- engagement with and confidence in local and central government institutions;
- the physical presence or absence of state institutions (e.g. police stations, health clinics, schools, water and sanitation systems);
- the reported quality of local public services; and
- local perceptions of and experiences with various forms of corruption.
This effort represents the most precise and consistent geocoding of Afrobarometer surveys to date, offering the greatest coverage across all six survey rounds.
Accessing the data
To protect the privacy of individual survey participants, those who wish to access the respondent-level data must request access from Afrobarometer and comply with the network’s data use policy.
AidData plans to additionally incorporate some aggregate geocoded Afrobarometer survey data into its GeoQuery platform, which will allow users to aggregate and access some of these data at various geographic levels, including municipalities, districts, and provinces. GeoQuery integration will also allow users to quickly and easily merge geocoded development investment data from AidData with local governance and development outcome data from Afrobarometer, thereby opening up new opportunities for impact evaluation and other analyses.
Data from Afrobarometer Round 7 surveys, which are in the field in 2017, will be added if sufficient funding is secured.
Ariel BenYishay is AidData’s Chief Economist, Director of Research and Evaluation at AidData, and Assistant Professor of Economics at the College of William & Mary.
Dan Runfola is AidData's Senior Geospatial Scientist and leads the GeoQuery research project, a joint Ph.D. and undergraduate lab at William & Mary which develops new geocomputational methods for the processing of large-scale spatial datasets in highly parallelized environments. He is an Assistant Professor at William & Mary, and Director of its Data Science Program.
Renee Rotberg is a Junior Program Manager at AidData.
Jessica Wells is a Senior Program Manager at AidData who manages the AidData Research Consortium and AidData’s geocoding work.
Miranda Lv is a GIS Analyst at AidData.
Seth Goodman is a Data Engineer at AidData, and a PhD Candidate in Applied Science at William & Mary.
Lidia Kovacevic is a 2017 graduate of the College of William & Mary.
The views expressed here are those of the authors alone, and do not necessarily reflect the views of the institutions to which the authors belong.