Tuesday 14th July      Wednesday 15th July      Thursday 16th July      Friday 17th July     

Thursday 16th July, 09:00 - 10:30 Room: L-103

Survey Research in Developing Countries 3
Convenor	Dr Irene Pavesi (Small Arms Survey )

Session Details

This session explores the challenges involved in conducting survey research in developing countries and discuss best practices in sampling, questionnaire design and fieldwork organisation.

Even more often than in developed countries up-to-date data on population size and composition is absent. Mobile populations, scarcely populated areas and areas connected only by low quality roads and security issues complicate the creation of a sampling frame. What strategies have researchers used to deal with these challenges?

Response rates tend to be high in developing countries. This is in part because in rural areas trust tends to be high or a survey is seen as an interesting break from everyday life. However in some cases the consent of village heads or other local leaders is an order to people to participate. How does this fit with the idea of informed consent?

High poverty in some areas raises ethical questions on whether and how respondents should be compensated for their time; if respondents receive cash or in kind compensation this can lead to competition among households for inclusion in the survey. What are appropriate ways to compensate respondents?

Large household with complex structures can make collection of household data a time consuming and error prone process. How can data be collected in an efficient way?

High ethnic and linguistic diversity poses challenges to both questionnaire translation and selection of interviewers. How can these challenges be dealt with?

If the people who design the questionnaire are not from the country of data collection, what procedures can be used to ensure that concepts in the survey resonate with those of the target population?

We welcome papers on these and related topics, such as reaching female respondents, use of ICT in data collection, surveying in (post-)conflict areas, and surveys among populations with high illiteracy rates

Paper Details

1. Geospatial Grid Based Sampling in Developing Nations
Mr Jamie Cajka (RTI International)
Ms Jamie Ridenhour (RTI International)
Ms Safaa Amer (RTI International)

Designing an in-person survey in developing nations is a challenging undertaking. To overcome these challenges, RTI International created a geospatial grid-based sampling methodology. A 1km square grid derived from LandScan world population dataset was placed over the country’s boundary. Then a grid sample was drawn incorporating demographic information. These squares were further subdivided into 50m or 100m square units which were randomly sampled based on the presence of a residence. This methodology has been successfully implemented in Thailand, India, Colombia, and Kenya. The design is transferable, adaptable, and scalable, and can be used anywhere in the world.

2. Facilitating Survey Sampling in Low- and Middle-Income Contexts Through Satellite Maps and Basic Mobile Computing Technology
Mr Marco Haenssgen (Oxford Department of International Development, Hertford College)

This case study describes how satellite mapping services such as Google Maps and Bing Maps can facilitate survey sampling in low- and middle-income contexts where researchers face challenging administrative and resource conditions (Google Inc., 2015; Microsoft Corporation, 2015). The paper outlines the process of selecting villages and households using satellite maps in rural China (Gansu province), and the associated technical, logistical, and methodological challenges. Other researchers can replicate this approach in contexts where resources for household listing are limited, where sampling frames cannot be produced from administrative data, and where residential structures are homogenous and distinctive.

3. Localised Censuses: comparing methods, findings and policy uses of National and Local ‘census’ surveys in a South African semi-rural near-mining community
Dr Tara Polzer Ngwato (Royal Bafokeng Administration)
Mr Martin Bekker (Royal Bafokeng Administration)

This paper assesses three ‘census’ datasets covering all 48000 households of a small South African community. This includes the 2011 South African government census, and two surveys commissioned by the local traditional/quasi-municipal administration (2011 PULA & 2014 Ntshegetse). Through these three independent datasets covering the same population in a short time period, we assess the impacts of differences of field work method (field entry, community acceptance, refusal rates, check backs, quality controls, geo-spatial referencing, etc.), question formulation (for example on household income), and final data weighting. The paper also considers the local politics of data ‘ownership’ and origin

4. Climate Refugees: Survey Research with Bangladeshi and Nepalese Migrant Communities in India
Ms Abigail Blue (UNFCCC)

This paper explores the methods, research design, tools and ethical considerations in conducting survey research within environmentally displaced, migrant communities resettling in India. Conducting survey research in vulnerable populations, with complex family and community structures and high socio-cultural and linguistic diversity requires an interdisciplinary research approach. This two-year longitudinal study employed in-depth exploratory interviews with self-settled migrants in specific urban and rural regions 0 the Bangladeshi slums of Mumbai and the Nepali slums and of the Himalayan Kedar Valley. This study provides methods and recommendations for conducting survey research in migrant communities of the developing world.