Current measures of political knowledge have limited construct validity, severely restricting our ability to draw from them either empirical or normative conclusions about the public’s level of political knowledge. Using a unique survey, I show that respondents’ level of political knowledge relative to their knowledge of other subjects is very sensitive to question choice. Indeed, an individual researcher’s selection of questions will change the normative implications of the results. The lack of construct validity for measures of political knowledge—one of the foundational pillars of research on political behavior—suggests a desperate need for new, reliable measures of knowledge that fairly assess voter knowledge and, by extension, competence. © 2016 Critical Review Foundation.