Experiments demonstrate that framing can influence public opinion under certain conditions. Yet outside the laboratory or survey setting, additional constraints might further limit elites' ability to reshape public opinion through framing. The 2009-12 health care debate provides an unparalleled opportunity to observe the interplay of elite rhetoric and public opinion in real-world conditions. This paper couples automated content analyses with survey data from 30,370 Americans to better measure elite frames, public opinion, and their relationship. Multiple empirical tests uncover only limited evidence of framing effects during the debate. While the frames employed by political elites are punctuated, mass attitudes are not. The very language Americans use to explain their opinions proves stable, although there is evidence that the public adopts the language of both parties' elites in a roughly symmetric fashion. Methodologically, the automated analysis of elite rhetoric and open-ended survey questions shows considerable promise in illuminating elite-mass interactions.