Cluster Sampling -A Modified Method In An Urban Area

Abstract

Research Question: Whether use of census enumeration block instead of ‘Ward’ or ‘Census Charges’ as a unit for cluster sampling in urban area has any advantages? Objective: To Study the feasibility of starting survey from the 1st household in sample clusters by using census enumeration block as a unit for cluster sampling in urban areas. Study design: Cross Sectional Study setting: National Capital Territory Of Delhi. Participants: House holds with children 12-23 months of age. Sample size: 15 clusters and 20 children (12-23 month) in each cluster. Results: To cover 20 children 12-23 months of age each an average 98.6 households (70.4-82.2% of total householders per cluster were surveyed. Population-wise census enumeration block are small units and are usually homogeneous and since majority of the households in a block were covered to complete the required sample size, error introduced by starting with first householder instead of a random starting point can be assumed to be very small. Alternatively systemic pattern can be used, starting with I st and last household in alternate blocks. The method has an added advantage that underserved areas amongst the surveyed clusters can be identify.