The estimation of change in a rotation panel design survey has been of interest in NSIs for a long time. The variance estimation of differences, ratios, etc. between different time points are important in order to be able to decide if the observed change is due to sampling variation or not. The rotation scheme means that point estimates from different time points will be dependent because of the partial overlap between the samples, this dependency should be taken into account in the variance estimator of the estimated change. In this paper different methods of doing the variance estimation of the estimated changes are investigated, the relatively simple method used for LFS within Statistics Sweden is compared with the method that was suggested by Berger (2004). In LFS the generalized regression estimator is used and the goal is to find a more reliable variance estimator for change that is also possible to use on a regular basis in a large scale statistical production environment. The methods are illustrated by a simulation study with the quarterly sample from the Swedish LFS as a population.
Berger, Y. G. (2004). Variance Estimation for Measures of Change in Probability Sampling. Canadian Journal of Statistics, 32, pp. 451-467.
Qualité, L. & Tillé, Y. (2008). Variance Estimation of Change in Repeated Surveys and its Application to the Swiss Survey of Value Added. Survey Methodolgy, 34, pp.173-181.
Tam, S. M. (1984). On Covariance from Overlapping Samples. The American Statistician, 38, pp. 288-289.
Keywords: Survey sampling; Rotation panels.; Estimation of change; Variance estimation
Biography: Senior methodologist at Statistics Sweden. Main interest is variance estimation and software in complex surveys