Abstract:
|
Adaptive design theories and procedures are being developed in reaction to declining response rates and increasing data collection costs. Adaptive design uses survey paradata to guide data collection strategies in order to achieve higher quality estimates per unit cost. In theory, given specified data quality and cost functions, adaptive design should attempt to optimize the tradeoffs between quality and costs through mathematical programming. In practice, however, quality functions may be multidimensional, and often times there are no clearly defined quality and cost functions at all. In lieu of response rate, the R indicator has been proposed as an alternative indicator of nonresponse bias, which has led to adaptive design strategies that focus on improving the balance of the responding sample through judicious resource allocation during data collection. However, the R indicator has limitations as a data quality indicator, which could compromise the effectiveness of R-based adaptive design approaches. The goal of this round table is to share our experience in implementing adaptive designs, explore practical adaptive design options, and discuss future research directions.
|