Propagation of nonresponse weights for censoring in multi-phase screening in complex sample designs

Daniel F. McCaffrey, Naihua Duan, Sally C. Morton

Research output: Contribution to journalArticlepeer-review

Abstract

We explore an analogy between survival analysis and nonresponse analysis for complex sample designs that include multi-phase screening procedures to determine the eligibility of sampled units. Units excluded due to ineligibility are analogous to mortality; nonresponse that occurs at multiple phases of the screening procedure is analogous to censoring. Based on this analogy, we discuss methods for estimating the response rate, the nonresponse weights, the eligible population size and the eligibility rate. In particular, we propose a method that estimates nonresponse weights at each phase of screening, and propagates these weights forward as the analysis weights for units that screen eligible at each phase. The analysis weights are used in nonresponse analyses at later phases of screening and for the analyses on the final sample of eligible units. This method is analogous to using a life-table estimator to estimate the probability of eligibility; the sum of the analysis weights for the final sample estimates the total number of eligible units in the population.

Original languageEnglish (US)
Pages (from-to)213-231
Number of pages19
JournalHealth Services and Outcomes Research Methodology
Volume1
Issue number3-4
StatePublished - 2000
Externally publishedYes

Keywords

  • Eligibility screening
  • HIV Costs and Service Utilization Study
  • Life-table
  • Respondent coverage rate

ASJC Scopus subject areas

  • Health Policy
  • Public Health, Environmental and Occupational Health

Fingerprint Dive into the research topics of 'Propagation of nonresponse weights for censoring in multi-phase screening in complex sample designs'. Together they form a unique fingerprint.

Cite this