In this paper we study the experience of individual voters on Election Day, November 2006, with two widely used voting technologies, optically scanned paper ballots and electronic voting machines. The focus of our empirical analysis is the relationship between voting technology and several dependent conditions, including voter satisfaction with their voting experience, the time it takes to vote, voter confidence that their ballot will be accurately counted, and self-reported undervotes. We find that a significant portion of the variation between the two voting technologies in voter behavior and affect is a function of election administration. 123 1 Several researchers eschew the term error, preferring to use the aggregate condition of undervotes to characterize the inconsistency between voter intent and the recording of that intent on the ballot. This latter choice is motivated by the view that there are several causes of this condition that are beyond the voter's control. Our use of the term, however, is not intended to suggest that the sole cause of error is the voter. We simply prefer to characterize the inconsistency between voter intent and the recording of that intent as an error (i.e., breakdown) in the electoral process.