A considerable body of data suggests that men know more about politics than do women. Although gender gaps exist in other aspects of political behavior, the unusual magnitude of the gender gap makes it particularly perplexing. In this paper, we advance and test the hypothesis that the knowledge gap is partly an artifact of how knowledge is measured. If men are disproportionately more likely to guess than are women, then observed gender disparities in knowledge will be artificially inflated. To test this hypothesis, we reexamine data used in two recent inquiries concerning the gender gap in knowledge, along with experimental data from the 1998 NES Pilot Study. All analyses point to a common conclusion: approximately 50% of the gender gap is illusory, reflecting response patterns that work to the collective advantage of male respondents.