Assessing Uncertainty When Combining Administrative Data to Estimate Population Counts

• Activity Number: 134
• Type: Contributed
• Date/Time: Monday, August 10, 2015 : 8:30 AM to 10:20 AM
• Sponsor: Government Statistics Section
• Author(s): Dilek Yildiz* and Peter Smith
• Companies: University of Southampton and University of Southampton
• Keywords: Bayesian methods ; census ; patient register ; log linear models with offsets ; population estimates

Abstract

The aim of the research is to develop a methodology to make use of already collected administrative data to estimate population counts for England and Wales in the absence of a traditional census. The key goals are to produce population counts by age, sex and region (local authority) by combining administrative data, and estimate associated measures of precision. Previous research (Yildiz and Smith 2015) showed that combining administrative data by log linear models with offsets in a classical framework is useful to decrease, to a certain extent, the bias in the administrative data sources. However, they did not provide the estimates of uncertainty for the final estimates. Therefore, this research moves away from the frequentist approach, extends the previous work, and focuses on combining data sources in a probabilistic framework. To illustrate the proposed methodology the aggregate level patient register is combined with auxiliary data by log linear models with offsets in a Bayesian framework.