How to improve representativeness and cost-effectiveness in samples recruited through meta: A comparison of advertisement tools.

Abstract

The use of paid advertisements on social media, in particular Meta platforms, to create samples for online survey research is becoming increasingly common. In addition to researchers working on hard-to-reach populations, Meta's promise of unmediated, quick, and cheap access to a large pool of survey takers across the world is appealing also for researchers who want to create diverse samples of national populations for cheaper prices. Yet the design of Meta's advertisement optimization algorithm complicates the use of Meta advertisements for this purpose, as it generates a trade-off between cost-effectiveness and sample representativeness. In this paper, we rely on original online surveys conducted in the United Kingdom, Turkey, Spain, and the Czech Republic to explore how two primary tools determining the audience of Meta advertisements, i.e., campaign objectives and demographic targeting, affect the recruitment process, response quality, and sample characteristics. In addition to documenting the trade-offs between the cost and representativeness in Meta samples, our paper also shows that researchers can create high-quality, cost-efficient, and diverse samples if they use the right combination of Meta advertisement tools.