Choices of Alpha Value in Regression Composite Estimation for the Canadian Labour Force Survey
Auteur : Edward J. Chen
Date de publication : 2002
Éditeur : Statistics Canada, Methodology Branch, Household Survey Methods Division
Nombre de pages : 36
Résumé du livre
The Canadian Labour Force Survey (LFS) is a continuous monthly survey with a complex rotating panel design where there is a 5/6 sample overlap between any two consecutive months. After extensive studies, including the investigation of a number of alternative methods for exploiting the sample overlap to improve the quality of estimates, the LFS has chosen and implemented a regression composite estimation method. Currently, a compromise linear estimator between level and monthly change driven estimates with the a value of 2/3 is implemented. This study is to evaluate a broad range of values on many different LFS characteristics and its impacts on the final survey weights.