A Pew Research Center study evaluates three methods for removing bogus respondents from online opt-in polls—trap questions, automated prescreening, and voter file matching—finding that while all modestly improve data quality, none provides a reliable solution and all slightly increase overestimation of Democratic support in the 2024 election.