Using design effects from previous cluster surveys to guide sample size calculation in emergency settings

Disasters. 2006 Jun;30(2):199-211. doi: 10.1111/j.0361-3666.2006.00315.x.


A good estimate of the design effect is critical for calculating the most efficient sample size for cluster surveys. We reviewed the design effects for seven nutrition and health outcomes from nine population-based cluster surveys conducted in emergency settings. Most of the design effects for outcomes in children, and one-half of the design effects for crude mortality, were below two. A reassessment of mortality data from Kosovo and Badghis, Afghanistan revealed that, given the same number of clusters, changing sample size had a relatively small impact on the precision of the estimate of mortality. We concluded that, in most surveys, assuming a design effect of 1.5 for acute malnutrition in children and two or less for crude mortality would produce a more efficient sample size. In addition, enhancing the sample size in cluster surveys without increasing the number of clusters may not result in substantial improvements in precision.

MeSH terms

  • Afghanistan / epidemiology
  • Child Mortality / trends
  • Child, Preschool
  • Cluster Analysis*
  • Cross-Sectional Studies
  • Emergencies*
  • Humans
  • Infant
  • Infant, Newborn
  • Nutrition Surveys
  • Outcome and Process Assessment, Health Care
  • Research Design*
  • Sample Size*
  • Yugoslavia / epidemiology