Best Practices for Matching Address-Based Samples to Telephone Numbers

• Activity Number: 17
• Type: Topic Contributed
• Date/Time: Sunday, July 31, 2011 : 2:00 PM to 3:50 PM
• Sponsor: Section on Survey Research Methods
• Author(s): Ashley Amaya and Ben Skalland and Becky Ruthenburg*+ and Karen Wooten
• Companies: NORC at the The University of Chicago and NORC at the The University of Chicago and NORC at the The University of Chicago and Centers for Disease Control and Prevention
• Keywords: address-based sampling ; National Immunization Survey ; telephone interviewing ; address to telephone matching ; match accuracy

Abstract

As concerns about declining representativeness of landline random digit dial (RDD) surveys have increased in recent years, many survey organizations have begun to explore alternative methods such as address-based sampling (ABS). Use of telephone as a primary interviewing mode can increase efficiency of ABS designs if addresses are reliably matched to telephone numbers. Several matching strategies were evaluated as part of the 2009 National Immunization Survey - Address-Based Sampling (NIS-ABS) Experiment. Addresses drawn from the USPS Delivery Sequence File were matched to telephone numbers by two sample vendors. Results indicate that the inclusion of "inexact" matches (i.e., matches to the street address, but not to the unit or building number) can increase match rates dramatically, especially in multi-unit buildings. However, the quality of these matches was shown to be relatively low. With the tradeoff between quantity and quality of matches, an approach that exclusively uses "exact" matches is more efficient in terms of cost and interviewing time than using inexact matches.