Monitoring Methods for Adaptive Design in the National Survey of College Graduates (NSCG): A Retrospective Appraisal

• Activity Number: 520
• Type: Topic Contributed
• Date/Time: Wednesday, August 7, 2013 : 10:30 AM to 12:20 PM
• Sponsor: Survey Research Methods Section
• Author(s): Stephanie Coffey*+ and Michael White and Benjamin M. Reist and Wan-Ying Chang
• Companies: US Census Bureau and U.S. Census Bureau and U.S. Census Bureau and National Science Foundation
• Keywords: propensity models ; monitoring ; adaptive design ; data quality ; r-indicators

Abstract

One goal of adaptive design is to allocate data collection resources efficiently, rather than exhausting money and time simply to increase response rate. Data monitoring is vital to this effort as it provides updated views of response information and data quality throughout the collection period. More up-to-date information can lead to interventions including: reducing contact attempts for low-impact cases unlikely to respond, switching contact mode to maximize response probability, or even stopping data collection. In the 2013 NSCG, data monitoring will help show the evolving state of data collection and inform interventions in a mode switching experiment.

This paper discusses several data monitoring methods explored for the NSCG using 2010 survey data including: R-indicators, response propensity by data collection type, and benchmarking. Some of these methods employ propensity models, and others rely on daily processing of survey data including nonresponse adjustments and weighting. In addition to discussing practical benefits and shortcomings of the methods, interventions are simulated to show their effect on the final state of data collection.