Wedgewood admits most election polls lacked real data

Wedgewood Polls acknowledged on October 20 that most of its election surveys lacked genuine data, after researchers questioned raw data from a Mississippi Senate poll. The firm described the operation as an experiment to test whether a fabricated pollster could gain recognition among polling aggregators and news organizations. Its stated purpose remains disputed. Wedgewood said it had conducted one genuine, crowdfunded poll in Nebraska and that most of its other releases were not supported by real survey responses. The firm said it wanted to expose weaknesses in polling-industry oversight and the vetting of data. Cooper Burton, a data scientist at FiftyPlusOne, said the explanation was misleading and argued that Wedgewood issued it after learning researchers had uncovered the fabrication. FiftyPlusOne researchers said they compared respondent and transaction identifiers in Wedgewood’s Mississippi survey with data from a genuine Pollfish survey. They reported patterns inconsistent with authentic interviews. Wedgewood had said it used Pollfish to conduct the survey. The researchers’ account also said the firm supplied two names, Jack Sims and Andrew Prater, and that Sims had...