The question of whether or not voters actually verify ballots produced by ballot marking devices (BMDs) is presently the subject of some controversy. Recent studies (e.g., Bernhard, et al. 2020) suggest the verification rate is low. What is not clear from previous research is whether this is more a result of voters being unable to do so accurately or whether this is because voters simply choose not to attempt verification in the first place. In order to understand this problem, we conducted an experiment in which 108 participants participated in a mock election where the BMD displayed the voters' true choices, but then changed a subset of those choices on the printed ballot. The design of the printed ballot, the length of the ballot, the number of changes that were made to the ballot, the location of those changes, and the instructions provided to the voters were manipulated as part of the experiment. Results indicated that of those voters who chose to examine the printed ballot, 76% detected anomalies, indicating that voters can reliably detect errors on their ballot if they will simply review it. This suggests that administrative remedies, rather than attempts to alter fundamental human perceptual capabilities, could be employed to encourage voters to check their ballots, which could prove as an effective countermeasure.