Bridging Livestock Survey Results to Published Estimates Through State-Space Models: A Time-Series Approach

• Activity Number: 125
• Type: Contributed
• Date/Time: Monday, August 1, 2011 : 8:30 AM to 10:20 AM
• Sponsor: Section on Government Statistics
• Author(s): Stephen Busselberg*+
• Companies: National Agricultural Statistics Service
• Keywords: State-Space ; Time Series ; ARIMA ; Kalman Filter

Abstract

The status quo in survey sampling is publication and dissemination of survey results which are directly calculated as a function of the sample design and the survey inclusion probabilities. This process is appropriate for population parameter estimates for the case in which the only information known about the population of interest comes from the sample. In other cases, however, components of a system related to the population may be available from sources outside the survey sample. This external information may allow a deterministic inferential solution to the population parameters of interest which are targeted by the survey. For this reason, the survey results may be incongruous with external data and publishing the survey results leads to contradictory information. Such is the case with livestock inventory surveys collected by the National Agricultural Statistics Service (NASS) of the United States Department of Agriculture (USDA). This paper details a solution to this issue of incongruity through the application of a State-Space model system.