We develop and validate a novel experimental design that builds a bridge between experimental research on the theory of spatial voting and the literature on measuring policy positions from text. Our design utilizes established text-scaling techniques and their corresponding coding schemes to communicate candidates' numerical policy positions via verbal policy statements. This design allows researchers to investigate the relationship between candidates' policy stances and voter choice in a purely text-based context. We validate our approach with an online survey experiment. Our results generalize previous findings in the literature and show that proximity considerations are empirically prevalent in purely text-based issue framing scenarios. The design we develop is broad and portable, and we discuss how it adds to current experimental designs, as well as suggest several implications and possible routes for future research.