Abstract
Shawn Jacobson, Sylvia G. Leaver, and David Swanson (1998)
"Choosing a Variance Computation Method for the Revised Consumer Price Index,"
Proceedings of the Section on Survey Research Methods, American Statistical Association.
The Consumer Price Index estimation system currently uses a hybrid linearization random
groups method to compute variances. This method was compared to a hybrid linearization
method implemented using SUDAAN, a stratified random group method implemented using VPLX,
and balanced repeated replication and unstratified jackknife methods implemented using
WesVarPC. In this analysis, variances were computed for several index series using each
candidate method. The methods were evaluated on the basis of their stability, which was
used as a proxy for the variance of the variance estimates. The authors found that the
stratified random group method implemented using VPLX was superior to the other
candidates. This paper describes the candidate variance calculation methods. Results are
given for each combination of studied index series and variance calculation method. The
authors found several instances in which there was a breakdown in the assumptions that
underlie linearization. One instance is described in detail. In addition, several
non-numeric issues, such as computation time and variance computation cost are discussed.
Last Modified Date: July 19, 2008
|