Published online by Cambridge University Press: 14 January 2011
In close, disputed elections, outcomes can depend on determinations of voter intent for ballots that have been filled out improperly. We surveyed 899 adult Minnesotans during a time when the state's U.S. Senate election was still disputed and presented them with ambiguous ballots similar to ballots under dispute in the same election. We randomized participants to three experimental groups, across which we varied the names on the ballot. We found that participants' judgments of voter intent were strongly biased by their voting preferences (p < .002 in all four ballots).