Model-Based Estimation in Official Statistics: Applying Small Area Estimation to the Dutch National Crime Victimisation Survey

• Activity Number: 615
• Type: Contributed
• Date/Time: Thursday, August 7, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Elke Moons*+
• Companies: Statistics Netherlands
• Keywords: small area estimation ; social statistics ; crime

Abstract

Over the past decade, the need to monitor and measure effectiveness of policies at different geographical levels has grown extensively. Often local authorities require information at very small regional levels, while most sample surveys at Statistics Netherlands are designed to provide statistics at the national and larger regional levels, using traditional design-based estimation procedures. Sample sizes are generally too small to provide sufficiently precise estimates at low regional levels. In such situations, model-based estimation techniques can be used. Statistics Netherlands' Social Statistics Division started the implementation of a small area estimation program to accommodate the demand for detailed regional statistics on labor force, crime and public safety. The availability of strong correlating register variables that are known for the entire population is crucial to improve the efficiency of small area estimates. This paper summarizes the experiences and results obtained with setting up this program and assessing the validity and plausibility of the results. The Dutch national crime victimization survey (NCVS) is used as an illustration.