Using a "Match Rate" Model to Predict Areas where USPS-Based Address Lists May Be Used in Place of Traditional Listing

• Activity Number: 359
• Type: Contributed
• Date/Time: Tuesday, August 3, 2010 : 10:30 AM to 12:20 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Valerie Hsu*+ and Jill Montaquila and J. Michael Brick
• Companies: Westat and Westat and Westat
• Keywords: Address-based sampling ; Computerized Delivery Sequence File

Abstract

The time and resources associated with traditional listing have led survey practitioners to consider the use of United States Postal Service (USPS) based address lists as an alternative. In our recent investigation, we found coverage of the USPS-based address lists to be generally good in urban areas, but possibly inadequate (after geocoding) in rural and high-growth areas. We develop a "match rate" (the proportion of traditionally listed addresses that would have been obtained from a USPS-based list) model that identifies areas where USPS-based lists could be used in place of traditional listing. We use multiple regression to predict match rate with respect to characteristics associated with the geographic areas of interest. In this paper, we also discuss identifying a match rate threshold that is used to make an a priori decision of when it is acceptable to rely on USPS-based lists.