Replication Data for: Polling Across Borders: The Promise and Pitfalls of Convenience Samples in a Cross-National Context

This study evaluates the performance of two widely used survey platforms, Lucid and Morning Consult, across six diverse national contexts: Brazil, India, Japan, Nigeria, the Philippines, and the United States. We assess the impact of platform choice on sample composition, response quality, political judgments, and treatment effect estimation, focusing on a randomized corruption treatment embedded within the survey. Attention filter passage rates were similar and generally high across countries and platforms, while the percentage of high frequency survey-takers varied greatly across countries. Our findings reveal significant demographic skews, with both platforms consistently overrepresenting college-educated respondents. Despite these differences, political assessments and estimated average and heterogeneous treatment effects remain broadly consistent across platforms, and in Brazil our estimates largely tracked those from past research with a probability sample. We find some evidence of cross-national variation in the magnitude of treatment effects, but these differences were often platform-specific. These results suggest that convenience samples can provide reliable estimates of causal effects even in diverse contexts. Taken together, our research highlights the trade-offs between cost, speed, and representativeness in global public opinion research, offering insights into the challenges and opportunities of online survey platforms.