Lessons from a randomised experiment with mixed-mode designs for a household panel survey

This article describes a randomised experiment with mixed-mode survey designs in the context of a household panel survey. The experiment was designed to allow comparisons between two alternative mixed-mode designs (telephone interviewing plus face-to-face interviewing) and a unimode design (face-to-face interviewing only) in terms of response rates, sample composition, costs and other key outcomes, as well as to permit assessment of the feasibility of implementing the mixed-mode designs.