Abstract
Can messages from leadership affect views of the agencies that carries out policy? I conducted four survey experiments with U.S. samples of about 1,000 likely voters per experiment. Treatments randomized the quality of presidential messages, from reasonable to specious, and I evaluate how perceptions of the policy and the administering agency change from each message. The messages are purposely varied to be specious and ridiculous at times to test how agencies might be penalized by poorly constructed rhetoric. Results show that partisan leaders can increase support for their preferred policy and themselves, but perceptions of administrative agencies show little change. These results suggest that the dynamics of opinion formation for policies and leaders operate differently than opinion formation for the public agencies that might carry out policies.