Premium
Adjustment procedures to account for non‐ignorable missing data in environmental surveys
Author(s) -
Munoz Breda,
Lesser Virginia M.
Publication year - 2006
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.776
Subject(s) - missing data , statistics , weighting , estimator , inverse probability weighting , respondent , econometrics , sampling (signal processing) , population , mathematics , computer science , demography , medicine , filter (signal processing) , sociology , political science , law , computer vision , radiology
Methods for non‐response are well‐known techniques used in survey practice for handling missing data. In this approach, sampling units (respondents and non‐respondents) are classified in non‐response classes, and the sampling weight for each respondent unit is weighted by the inverse of an estimate of its response probability (also known as propensity score). Optimal weighting classes are selected using variables associated with the response but uncorrelated with the response indicator. We explore the assumptions needed to construct optimal adjustment classes in the case of non‐ignorable missing data in environmental surveys. We propose a modified Horvitz–Thompson non‐response estimator for the population total of the spatial random process of interest, and study some of its properties. By using the weighting class adjustment, we will account for the non‐ignorable missing data. Copyright © 2005 John Wiley & Sons, Ltd.