Alternative sampling designs for emergency settings: A guide for survey planning, data collection and analysis

This guide is tailored for programme managers, monitoring and evaluation specialists and survey leaders working with government agencies and humanitarian organizations. It provides them with instructions how to carry out rapid yet statistically reliable population-based surveys in difficult settings where data on the prevalence of acute malnutrition is needed. Government decision makers and humanitarian agencies need tools for rapid and effective prioritization of vulnerable areas and practical and reliable methods for on-going monitoring of the situation. In emergency settings, population-based surveys are conducted to assess the severity and magnitude of the situation and to obtain data for problem analysis and response planning.

The guide offers three different sampling designs appropriate for emergency settings, which offer limited time to collect required data. The data is a prerequisite to understanding the underlying causes of undernutrition and to the selection of the most appropriate actions to improve the health, nutrition and survival of the population. Carrying out a population-based survey can be time consuming and expensive, especially in emergency settings. The nature of an emergency exacerbates all the usual challenges associated with data collection while intensifying the urgency with which reliable data on the population are needed.The most commonly used survey method for emergencies is the two-stage 30X30 cluster survey. The guide gives the following three alternative sampling designs:

• the 33 clusters by 6 observations in each design.
• the 67 clusters by 3 observations in each design and
• the sequential design of up to 67 clusters by 3 observations in each but not exceeding the number 201.

This guide provides instructions for the planning, implementation, and analysis of data collected by the 33x6, 67x3, and sequential design. It reports that field applications of the above design innovations have shown that they:

• provide meaningful and valid results
• require less time and cost than is required for the 30X30 cluster design
• are sufficiently validated for wide-scale adoption by government and humanitarian agencies.

In conclusion, the guide hopes that the alternative sampling designs will become familiar and accessible to more users, and that the targeted ones will have the information they need to decide if the alternative sampling designs are appropriate for their purposes.