Online Program

On Precision in Estimates of Change Over Time where Samples are Positively Coordinated by Permanent Random Numbers
*Annika Lindblom, Statistics Sweden 

Keywords: measures of change, sample coordination, survey design, variance estimation

In the area of economic statistics measures of change are key parameters. To improve the precision in estimates of change over time the majority of the business surveys at Statistics Sweden use stratified simple random samples (STSI) from separate sample occasions that are positively coordinated over time (overlapping) by permanent random numbers (PRN).

Sample coordination by PRNs makes the correlation between the included level estimates quite complicated to estimate. However, a method for estimating this correlation was developed at Statistics Sweden in the late 1990’s.

This method has been applied in a study on how the design of a repeated business survey affects the precision in estimates of change over time. In particular, detailed stratification promotes high precision in each level estimate but often results in smaller overlap between consecutive samples (meaning less correlation).

This paper focuses on results from the study as well as on empirical results from the practical implementation of the method for estimating the correlation. Statistics Sweden’s survey “Turnover in the Service Sector”, using STSI that are positively coordinated by PRNs, serves as an example.