Presidential Leadership of Public Opinion : An Embedded Survey Experiment

Presidents increasingly go public to influence public opinion on issues, but research suggests that such efforts are ineffective. This presents us with a puzzle: Why do presidents go public so much if doing so is ineffective in influencing public opinion? I present a model that hypothesizes that different types of presidential communications provide voters with different types of information about the president. This information, in turn, interacts with, or primes, the voter's political predispositions, so that different types of voters will react differently to each presidential communication. I test this hypothesis with an embedded survey experiment. The results confirm the hypothesis in general. In the conclusion, I discuss limitations of the experiment and directions for future research