Distinguishing among Distributions Using Data from Complex Sample Designs

Abstract

This article presents the statistical methodology for estimating stratum cumulative distribution functions (cdf's) when the sample is drawn according to a complex design. Justification for the use of this methodology is offered, and examples are presented to demonstrate the expanded inference possible when estimated cdf's are used in addition to stratum means. The methods are applied to data from a survey of the patterns of care in radiation therapy facilities in the United States.

Key Words:

• Estimated distribution functions
• Empirical cdf's
• Cumulative distribution functions
• Inference for quantiles
• Inference for quantiles