2016-Analysis of Complex Survey Data


Analysis of Complex Survey Data 
Chair: Karla Fox (Statistics Canada)
Organizer: Lenka Mach (Statistics Canada) 
[PDF]

YVES BERGER, University of Southampton
An Empirical Likelihood Approach for Complex Sampling  [PDF]
 
Survey data are often collected with unequal probabilities from a stratified population. We propose a new empirical likelihood approach for sample data selected with unequal probabilities. We show that the empirical likelihood ratio statistic follows a chi-squared distribution asymptotically. The approach proposed does not rely on variance estimates, re-sampling or joint-inclusion probabilities, even when the parameter of interest is not linear. Standard confidence intervals based on variance estimates may give poor coverages, when normality does not hold. This can be the case with skewed data and outlying values. The empirical likelihood confidence interval proposed has good coverages, even when the sampling distribution of the point estimator is not normal. 
 
ELISABETH NEUSY, Statistics Canada
Confidence Intervals for Proportions Estimated from Complex Survey Data  [PDF]
 
A large number of methods of constructing confidence intervals for proportions have been proposed and studied for non-survey data. A few of these methods have been adapted for complex survey data. A simulation study was undertaken to evaluate the performance of the logit transformation interval, the modified Clopper-Pearson interval, the modified Wilson interval and the bootstrap percentile interval in the context of stratified sampling and two-stage sampling. The paper will briefly review these methods, and present the results and conclusions of the study. 
 
CHANGBAO WU, University of Waterloo
Empirical Likelihood for Public-Use Survey Data [PDF]
 
In this paper we develop empirical likelihood methods for analyzing public-use survey data that contain only the variables of interest and the final adjusted and calibrated survey weights along with final replication weights. Asymptotic distributions of the empirical likelihood ratio statistics are derived for parameters defined through estimating equations. Finite sample performances of the empirical likelihood ratio confidence intervals, with comparisons to methods based on the estimating equation theory, are investigated through simulation studies. The proposed approaches make empirical likelihood a practically useful tool for users of complex survey data.