7 Statistical analysis of censored environmental data

Michael G. Akritas, Thomas F. Ruscitti, G. P. Patil

Research output: Contribution to journalReview articlepeer-review

19 Scopus citations

Abstract

Censored data commonly occurs in environmental studies when pollutant levels fall below the detection (or reporting) limits of instrumentation. Estimation of population parameters, inference, and other analyses of censored data sets are problematic. Various methods for parameter estimation are surveyed, including simple substitution of detection limits; maximum likelihood estimators; and probability plotting. Estimation of location difference in the 2-sample case is presented in the framework of extensions of the nonparametric Hodges-Lehmann estimator. Various regression methods for censored data are discussed, including maximum likelihood; Buckley-James; least absolute deviations; and Theil-Sen regression. Selected examples, proposed new methods, and an extended bibliography are included throughout.

Original languageEnglish (US)
Pages (from-to)221-242
Number of pages22
JournalHandbook of Statistics
Volume12
DOIs
StatePublished - 1994

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Modeling and Simulation
  • Applied Mathematics

Fingerprint

Dive into the research topics of '7 Statistical analysis of censored environmental data'. Together they form a unique fingerprint.

Cite this