📣 What the paper proposes

A novel procedure uses the advertising option on Facebook to sample party activists across countries and over time. This approach targets different parties via ads, allows quota controls, and collects large samples at relatively low cost—making it possible to improve representativeness through poststratification and to check robustness with subsamples.

🧭 How activists were recruited and why it matters

• The method leverages Facebook's targeting and ad-delivery system to reach activists affiliated with specific parties and movements.
• Quotas and large sample sizes enable statistical adjustments (poststratification) and subsample-based robustness checks that mitigate common biases in online samples.
• The procedure is suitable for repeated cross-sectional and time-series cross-sectional designs focused on politically active populations.

📌 Three real-world tests that demonstrate the approach

• A Facebook sample approximated intraparty decisions and matched the outcome of a leadership contest of the Alternative for Germany.
• A weighted Facebook sample produced estimates similar to those from a representative local-leader survey of the Social Democratic Party of Germany.
• Two Facebook samples analyzing subgroups of key demographics in Thailand—where party population parameters were unknown—showed that the color-coded conflict was driven by different concepts of regime type rather than by a left–right divide on economic policy-making.

🔑 Key takeaways

• Facebook sampling can produce results comparable to traditional surveys of party activists and local leaders.
• The approach is low-cost, supports quotas, and facilitates poststratification and robustness checks.
• It is especially promising for time-series cross-sectional studies of political activists across countries and periods.

🌍 Why this matters for political science

The study addresses a gap in intraparty research caused by the difficulty of random sampling among activists. By offering a scalable, affordable sampling strategy that performs well in diverse cases, Facebook sampling expands the toolbox for studying parties and social movements over space and time.