Premium
A kernel smoothing method of adjusting for unit non‐response in sample surveys
Author(s) -
Silva DamiãoN. Da,
Opsomer Jean D.
Publication year - 2006
Publication title -
canadian journal of statistics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.804
H-Index - 51
eISSN - 1708-945X
pISSN - 0319-5724
DOI - 10.1002/cjs.5550340402
Subject(s) - estimator , statistics , weighting , monte carlo method , variance (accounting) , mathematics , kernel smoother , smoothing , econometrics , replication (statistics) , kernel (algebra) , computer science , kernel method , artificial intelligence , medicine , accounting , combinatorics , radial basis function kernel , support vector machine , business , radiology
Non‐response is a common problem in survey sampling and this phenomenon can only be ignored at the risk of invalidating inferences from a survey. In order to adjust for unit non‐response, the authors propose a weighting method in which kernel regression is used to estimate the response probabilities. They show that the adjusted estimator is consistent and they derive its asymptotic distribution. They also suggest a means of estimating its variance through a replication‐based technique. Furthermore, a Monte Carlo study allows them to illustrate the properties of the non‐response adjustment and its variance estimator.