Skip to content
America’s
Political Pulse

Reply to Van der Stigchel et al.: Empirical evidence that AI survey contamination is real and substantial

Proceedings of the National Academy of Sciences, 2026

← All corrected papers

The paper

Sean J. Westwood and Samuel Frederick (2026). Reply to Van der Stigchel et al.: Empirical evidence that AI survey contamination is real and substantial. Proceedings of the National Academy of Sciences 123(8), e2537420123.

doi:10.1073/pnas.2537420123 (opens in a new tab) · No public replication archive

Main text

  1. p. 1, second paragraph

    Published

    Our method flagged 3.1% of respondents; Prolific’s independent detection tool flagged 3.0%.

    Correction

    Our method flagged 3.1% of respondents; Prolific’s independent detection tool flagged 3.2%.

    Basis In Fig. 1, Prolific’s tool flagged 39 + 53 = 92 of 2,898 respondents (3.17%) and our method flagged 36 + 53 = 89 (3.07%), so Prolific’s tool flagged slightly more respondents than our method. The combined rate (4.4%) and the overlap (41%) are unchanged.