Journal of Official Statistics, Vol.8, No.2, 1992. pp. 149–166
Time Series Modeling of Sample Survey Data from the U.S. Current Population Survey
Richard B. Tiller
Abstract:The signal extraction approach to repeated sample surveys is potentially an effective way to reduce high variances in conventional sample estimators arising from small sample sizes. A signal-plus-noise model of labor force estimates from U.S. Current Population Survey is formulated as a structural time series model with explanatory variables where variance-covariance information from the survey sample is used to place restrictions on the time series model. This model is fit to a statewide series. Model-based estimates are compared to the observed sample data and the effect of controlling for sampling error is explored.
Keywords:Small area estimation; correlated sampling error; signal extraction; Kalman filter.
Copyright © Statistics Sweden 1996-2018. Open AccessISSN 0282-423XCreated and Maintained by OKS Group