President George W. Bush initiated several major policy reversals that involve administrative politics. For example, in June 2002, he gave a national address promoting the creation of a Department of Homeland Security after having long opposed a cabinet-level office for domestic security. This speech and the president's subsequent actions helped create the office with features he desired (Canes-Wrone 2006; Maltzman and Adams 2003). Four years later, only a day after the Democrats had gained majorities in the House and Senate, Bush again abruptly initiated a significant administrative change. This time, he fired Secretary of Defense Donald Rumsfeld despite having claimed a week before that Rumsfeld would remain in office for the remainder of the term (Gerstenzang 2006). Political commentators were quick to declare that public opinion drove these policy switches. In the case of the Department for Homeland Security, observers maintained the decision was motivated by polling data that suggested the administration was not adequately addressing the threat of terrorism (Balz 2002; Bettelheim and Barshay 2002). In the case of Rumsfeld, Bush suggested that public opinion was a factor. Speaking to the press the day after the midterm elections, Bush surmised that Americans voted to register their displeasure at the lack of progress we are making [in Iraq] and announced that the Pentagon needed a fresh perspective that would involve new leadership (Bush 2006). One might expect scholarship to shed light on whether these incidents fit general patterns of how public opinion affects presidents' decisions regarding the federal bureaucracy. After all, in recent years, the literature on the administrative presidency has produced a wealth of systematic knowledge about a range of topics, including the politicization of appointments (e.g., Lewis 2008; Moe 1985b), presidential control of the bureaucracy (e.g., Aberbach and Rockman 2000; Hammond and Knott 1996; Wood and Waterman 1991), budgetary politics (e.g., Carpenter 1996; Krause 1996), and White House centralization (e.g., Rudalevige 2002). Yet scant attention has been paid to the ways in which a president's public relations affect his (or her) decisions and behavior with respect to the federal bureaucracy. A couple of studies suggest that lower personal approval ratings increase the likelihood that a president will issue executive orders (e.g., Deering and Maltzman 1999; Mayer 1999). (1) A few other studies, which I will review in the next section, analyze the relationship between personal popularity and the appointments process. Finally, Whitford and Yates (2003) find that presidents' public statements about drug control policy have affected the willingness of U.S. attorneys to prosecute narcotics-related cases. In sum, we know exceedingly little about how presidents' public relations affect administrative politics. When will presidents create new agencies to satisfy public pressure to do something about a problem? How, if at all, are agencies that have been created to satisfy such public pressure likely to differ from others? And what are the consequences of these answers for the size and organization of the federal bureaucracy? Likewise, under what conditions, and for what types of appointments, is public opinion apt to affect the president's choice of nominee or the prospect of confirmation? And when does public opinion provoke presidents to fire officials? The dearth of scholarship on such questions is striking, particularly given the emphasis in the literature on the establishment of a public or plebiscitary presidency (e.g., Lowi 1985; Skowronek 1993). Research suggests that chief executives have increasingly utilized opinion polls for a variety of purposes (e.g., Geer 1996; Jacobs and Shapiro 1995) and taken presidential-legislative debates to the public (Canes-Wrone 2006; Edwards 2003; Kernell 2007; Tulis 1987). Also, many studies indicate that personal popularity increases a president's influence with Congress (e. …