Improving Representativeness in Probability-Based Panel Recruitment through Adaptive Survey Design

Citation

Ardiaca, J., Muñoz, J., Tormos, R. Improving Representativeness in Probability-Based Panel Recruitment through Adaptive Survey Design.

Abstract

This paper evaluates the application of Adaptive Survey Design (ASD) to improve representativeness during the recruitment phase of the Catalan Citizen Panel, a probability-based push-to-web+paper panel in Catalonia. Using randomized experiments and predictive modeling on over 120,000 sampled individuals across six waves, we estimate treatment-specific response propensities and allocate recruitment efforts accordingly. Our optimization framework aims to maximize representativeness, under realistic budget constraints. Simulations suggest that ASD-based treatment assignment improves representativity compared to random or uniform allocation, with minimal impact on overall response rates. The findings offer a practical path for implementing data-driven recruitment strategies in public opinion research.