Researchers often rely on convenience samples via social media like Facebook, but these aren't typically representative of the broader population. This article introduces a new quota sampling method using targeted Facebook advertisements to achieve more balanced representation.
Data & Methods: Used randomized controlled trials and probability weighting within Facebook's Marketing API for semi-automated recruitment.
Key Findings: Compared survey results on climate change opinions from their Facebook-sampled group (2,432 US respondents) with a contemporaneous national panel survey using random sampling.
- The findings largely align across key metrics despite being convenience samples.
- The Facebook-recruited data approximates official government statistics like the American Community Survey for several validation questions.
Why It Matters: This approach provides political scientists and social researchers with a more accessible alternative to traditional probability sampling methods while still achieving population-level insights.
