Choosing a Cluster Sampling Design for Lot Quality Assurance Sampling Surveys

Area of Science:

• Statistics
• Public Health Surveillance
• Survey Methodology

Background:

• Lot quality assurance sampling (LQAS) is vital for monitoring in resource-limited settings.
• Combining LQAS with cluster sampling offers timely and cost-effective data collection.
• Existing methods vary in how they accommodate clustering in survey designs.

Purpose of the Study:

• To compare cluster lot quality assurance sampling (LQAS) methodologies.
• To provide recommendations for selecting appropriate cluster LQAS designs.
• To clarify technical differences and address misconceptions in the literature.

Main Methods:

• Comparative analysis of three cluster LQAS methodologies.
• Evaluation of distributional assumptions and clustering parameterization.
• Illustration using vaccination and nutrition survey examples.

Main Results:

• Cluster LQAS methods are robust to distributional assumptions.
• Clustering parameterization significantly influences design (sample size).
• Current parameterizations show inconsistency with observed data, complicating method selection.

Conclusions:

• Choice of cluster LQAS method is not straightforward due to parameterization issues.
• Further research is needed to characterize clustering patterns for setting-specific best practices.
• Development of improved cluster LQAS designs tailored to specific applications is recommended.