Time Series Analysis of Repeated Surveys: The State–space Approach |
| |
Authors: | Moshe Feder |
| |
Affiliation: | Research Triangle Institute, 3040 Cornwallis Road, P.O. Box 12194, Research Triangle Park, NC 27709-2194, USA |
| |
Abstract: | Cross sectional estimates from repeated surveys form a time series { yt }. These estimates can be viewed as the sum y t = Y t + e t of two processes, { Y t }, the population process and { e t }, the survey error process. Serial correlations in the latter series are usually present, mainly due to sample overlap. Other sources of data such as censuses, administrative records and demographic population counts are also available. The state–space modelling approach to the analysis of repeated surveys allows combining information from different sources, incorporating benchmarking constraints in a natural way. Results from these methods seem to compare favourably with those from X-11-ARIMA in filtering out survey errors. |
| |
Keywords: | Kalman filter survey error benchmarking multivariate model |
|
|