Extended Data Table 2 Contribution of offline recruitment and weighting schemes to discrepancies between surveys

From: Unrepresentative big surveys significantly overestimated US vaccine uptake

  1. A portion of each Axios–Ipsos wave is recruited from a population with no stable internet connection; Ipsos KnowledgePanel provides tablets to these respondents. In the Axios–Ipsos 22 March 2020 wave, the offline panellists (n = 21) were 24 percentage points less likely to be vaccinated than online panellists (n = 974). Weighting the same Axios–Ipsos data (n = 995) to the age and gender target distribution implied by Delphi–Facebook’s weights make the vaccination estimates higher by 1 percentage point. However, this number is still lower than Delphi–Facebook’s (responses from 14–20 March 2020, n = 249,954) own estimate of 46%. During this time period, the CDC benchmark vaccination rate was 35.2%. This suggests that the recruitment of offline respondents and different weighting schemes each explains only a small portion of the discrepancy between the two data sources.
Back to article page