In this study we looked to assess the error recognition rate of voter-verified paper audit trails in electronic voting systems, and see if by applying principles of human-computer interaction and usable security we could improve them. We ran a test comparing two prototypes with participants, in which we collected data on their ability to recognize errors in voter verification slips, and conducted a post-test interview to perform a thematic analysis. The first prototype was based on real world paper audit machine designs, which involved a timer-based system; the second was based on our application of HCI principles, which involved the addition of a confirmation and back button instead of a timer. We found that the error recognition rate increased with the confirmation based system. We also found that the confirmation and back button improved the usability of the system, and reduced users' stress and anxiety with the voting process.