Research on the public presidency has focused on the direct and often unmediated appeals that modern presidents make to the general public. Presidents have been portrayed as first among the talking heads--"going public" over the heads of members of Congress and other political actors to communicate directly with the general public. The result is that the "public presidency" has been defined as facing outward--the president talks and travels to the country. For instance, Jeffery Tulis (1987) suggests that changing norms of governance unshackled presidents from cooperating with members of Congress and freed them to personally deliver speeches (such as the State of the Union address) to public audiences in order to stir Americans against recalcitrant Washington elites. Samuel Kernell argues that changing institutional incentives motivated presidents to widen their external exposure by increasing their travel and speeches (cf. Edwards 1983; Kendell 1997).
What makes the "public presidency" public, however, is not only its outward oriented activities but also its systematic monitoring of the attitudes of the mass public. The public presidency is two-sided: Presidents take themselves and their policies to the public and they bring the public's opinions and perceptions into the inner sanctum of the presidency. Recent research has unearthed the second face of the "public presidency"--its sophisticated and routinized "public opinion apparatus" for conveying the public's sentiments to the president and his senior aides. (1)
The significance of the presidency's "public opinion apparatus" is that it links the public-talking and the public-listening dimensions of the presidency. What presidents say, how they say it, and where they make their comments is a function of what the White House learns from the public. Too much research on the public presidency has focused on the outward activities of presidents without adequately explaining the content of those activities and examining the reciprocal influence of the public on presidential activities. Put simply, institutionalized White House monitoring of public opinion is the flip side of presidential outreach to the mass public: Routine tracking of public opinion calibrates presidential appeals to Americans.
Connecting the White House's institutionalized apparatus for tracking public opinion with presidential appeals to the public has two implications for continuing research on the public presidency. First, it helps explain a significant empirical puzzle: What explains the content, tone, and place of public presidential talking? This line of analysis broadens the ambit of public presidency research: The emergence of an institutionalized apparatus for tracking public opinion challenges a fundamental tenet of theories about politics--namely, that politicians are crippled by imperfect information and uncertainty about the preferences of voters (Morton 1993; Calvert 1985). Examining the impact of White House polling on presidential behavior repositions the general debate over information uncertainty toward specifying the impact of information that politicians do possess. Second, the theoretical and normative significance of the public presidency lies in the two-way relationship between the president and the mass public. Analyzing the impact of presidential polling on presidential public statements raises fundamental questions about strategic theory and democratic norms that cut across particular institutions and systems of representative government. Put simply, the interaction of presidential polling and promotion strengthens the link between the study of the public presidency and broader theoretical and empirical debates.
Addressing the empirical and theoretical implications of presidential polling is a substantial research project. Recent research has provided a comprehensive historical overview of the origins of presidential polling (Eisinger 2003; Towle 2004) and its use among White House officials (Heith 1998) as well as the impact of private survey results on the policy positions of John Kennedy during his presidential campaign (Jacobs 1992b, 1993; Jacobs and Shapiro 1994) and on efforts to shape the image of President Richard Nixon (Druckman, Jacobs, and Ostermeier 2004; also cf. Geer 1991, 1996; Herbst 1993; Heith 2000).
This article examines a critical part of this broader story. It traces the changes over time in the presidential polling of Kennedy, Johnson, Nixon, and Reagan based on extensive research in presidential libraries. In particular, it traces the evolution in the amount of presidential polling and in the changing functions of these polls as reflected in the types of survey questions that the White House has fielded. The result is a mapping of the extent of presidential polling and a systematic content analysis of the information on public opinion that the White House received. This is, as far as we know, the first systematic inventory of the number and content of the surveys, questions, and types of information reported to the White House across four presidencies spanning nearly three decades.
Our analysis identifies two changes in the extent and purpose of presidential polling since Kennedy: an increase in the amount of presidential polling, and a shift from polling the public's policy preferences to polling its non-policy evaluations related to personal image and appeal--a potentially potent basis for appealing to voters. We suggest that the changing extent and purpose of presidential polling reflects a shift in the weight that politicians place on their two principal strategic goals--securing their reelection (or that of their party) and pursuing the policy goals of themselves and their supporters.
We begin by briefly discussing electoral and policy goals as competing sets of incentives that have motivated presidents to change the scope and nature of their polling apparatus. The second section maps the growth in the number of surveys and survey questions since Kennedy. The third section examines the evolution in the type of survey questions that presidents fielded based on our content analysis of presidential polling and our analysis of strategy discussions within the White House. In general, presidential polling on the public's policy attitudes has decreased or stagnated since Johnson, while polling on the president's personal image and appeal has increased.
Political Motivations and the Allure of Personal Image
Polling by politicians is often assumed to be driven by electoral fear and to be geared toward "pandering" or, more neutrally, responding to the policy preferences of voters. Indeed, a long line of theorizing predicts that competition in two-party elections motivates candidates to converge toward the electorate's median policy preference (Downs 1957; Black 1958), and empirical research has found substantive policy responsiveness--namely, that the policy positions of candidates (and elected officials) follow or respond to the public's policy preferences (e.g., Erikson, MacKuen, and Stimson 2003; Erickson 1978; Kuklinski 1978; Page and Shapiro 1983). In truth, the pioneers of public opinion research such as Dr. George Gallup envisioned scientific polling as providing an impartial means for translating the policy preferences of most citizens into government policy (Gallup and Rae 1940).
Research on median voting does not capture, however, the full spectrum of political motivations; politicians are also motivated by the policy goals to enact the policies favored by themselves and their supporters (Fenno 1978; Bianco 1994, 36-37). A large body of research suggests that politicians make decisions based on what they consider "good public policy" and on the intense preferences of core supporters. In particular, candidates and office holders follow the policy preferences of partisans, who include committed activists and leaders within the political parties as well as voters whose sense of attachment with one party prompts them to routinely choose its candidates (Aldrich 1995; Kiewiet and McCubbins 1991 ; Cox and McCubbins 1993; Kingdon 1989).
Although policy as opposed to electoral goals can pose stark choices to politicians, the reality of policy making in representative democracies is that motivations are rarely one-sided, with one set of goals singularly monopolizing the behavior of elected officials. Politicians who are primarily motivated to maximize their vote by consistently responding to the preferences of majorities are constrained by the political need to hold their narrow but core set of supporters who are responsible for volunteering time, money, and other resources (Page 1978). On the other hand, politicians intent on pursuing policy goals search for strategies and tactics to avoid electoral removal.
Politicians can simultaneously pursue electoral and policy goals by crafting a personal image that appeals to voters on non-policy grounds while leaving themselves leeway for acting on the policies most preferred by themselves and their strongest supporters. Indeed, research confirms the strong incentives for politicians to construct and maintain an appealing image (Fenno 1978; Page 1978; Popkin 1994) and point to the impact of candidate image on the public's attitudes, evaluations, and vote choice (Funk 1999; Mendelsohn 1996).
The strategic motivation for presidents to construct an appealing image helps to explain the allure of "going public." What presidents talk about, who they address, and how they conduct themselves are all tools for calibrating the public's perceptions of them (Jacobs et al. 2003). As William Riker insisted, honing "how policies are presented [and] discussed" is one of the most potent weapons at the …