Speaker's Notes
The geographic distribution of Specialists throughout the country allowed for a broad sampling of establishments through all regions of the U.S. The choice of data collection locations, therefore, was based on the Specialists' geographical areas of responsibility and provided a reasonably convenient design for estimating national risk-related behaviors and practices.
An improved design for measuring trends within the retail food industry would be one based on probability sampling of the whole nation in which the number of establishments inspected within any given location would be related to the volume of food consumed within that location. This would have required the development of comprehensive establishment lists for randomly selected locations around the country and excessive travel, which would have been cost-prohibitive.
Comparison set establishment lists was the approach used by Specialists to select establishments. Using establishment inventories, most often provided by State or county health agencies, each Specialist developed 5 Comparison Set Establishment Lists for each of the 9 facility types. A comparison set list usually included 10-20 establishments. After preparing the 5 Comparison Set Establishment List, a table of random numbers was used to select one establishment from each list for inspection. The random selection of establishments from comparison set lists reduces the potential for selection bias.
