Abstract:
Multi-sourcing, or using multiple sample providers to recruit respondents into a single survey, is becoming a necessity for large research agencies and panel companies alike. The paper discusses the results of a comparison study of nine US sample providers, and shows that, beyond demographics, each sample source has an inherent bias because of the way that they recruit, manage, reward, and communicate with their members. It shows that multi-sourcing can provide a stable and representative sample under certain conditions, and gives concrete examples of how to manage multiple sample partners.
